Development of The Legend of Zelda

This page should not be edited lightly

This page is meant to serve as the most complete, insightful, and historically accurate account of how The Legend of Zelda series began, the changes it went through over the years, and why those changes were made. It highlights how market conditions and the minds at Nintendo influenced the development of each Zelda game, both from a creative and business standpoint. It is based on hundreds of developer interviews and behind-the-scenes materials, as well as a sound understanding of the games industry. Please do not edit this page if you are not confident in what you are doing and express your concerns on its talk page instead.

The Legend of Zelda - A Complete Development History Read the eBook version here (v1.4).

The Dawn of Mario

Shigeru Miyamoto Joins Nintendo

Sheriff (1979)

In 1977, a young artist and toy designer named Shigeru Miyamoto joined Japanese toy maker Nintendo. Having impressed company president Hiroshi Yamauchi with his work, Miyamoto was assigned to create art work for an arcade game named Sheriff, where the player found themselves surrounded by bandits and needed to fend them off, saving a young woman they had captured in the process.

Sheriff was Nintendo's first "damsel-in-distress" game—a concept Miyamoto would subsequently re-use across many of his own products. The first of his original titles was an arcade unit named Donkey Kong, which materialized when plans for Nintendo to create a game based on the Popeye license ran into complications. Drawing inspiration from the trio of Bluto, Popeye, and Olive Oyl, Miyamoto conceptualized three characters of his own: Donkey Kong (a large ape), Jumpman (the protagonist of the game), and a female character that would later be given the name "Pauline" (the damsel in distress that Jumpman would need to rescue). Nintendo repurposed the circuit board of Radar Scope, one of their earlier arcade units, to produce Donkey Kong, and the game was released in 1981 to critical acclaim—although not without a slight change. By the time Donkey Kong shipped, the Jumpman character had been renamed "Mario" by Nintendo's American branch.

The game went on to be popular in the North American market and Miyamoto got started on his next project: a sequel titled Donkey Kong Jr., where the characters' roles were reversed. Donkey Kong, the ape from the first game, had been imprisoned by Mario, who was standing guard over his cage. The player would assume the role of Donkey Kong Jr. and attempt to rescue his father from his captor. Eventually, the Mario character proved popular enough to receive his own game, co-starring a brother, Luigi. Miyamoto titled the new game Mario Bros., and when it was released in 1983 it allowed two players to team up and play together.

The Famicom and Takashi Tezuka

By 1980, Nintendo wanted to branch out of the arcade business and had begun researching the possibility of developing a device on which users could play games at home. Inspired by the Atari 2600 and ColecoVision (to which the company later ported their Donkey Kong game), Nintendo's goal was to create a piece of hardware that would be cheaper than the competition, less intimidating to those unfamiliar with technology, and more appealing to children. They dubbed their invention "Family Computer".

The "Famicom" was released in July 1983, and launched with ports of Donkey Kong, its sequel Donkey Kong Jr., and Popeye. The following year, Nintendo found itself another employee that would prove to be a talented designer and make immense contributions to the company's Famicom library: an Osaka University of Arts graduate, Takashi Tezuka.

Tezuka would go on to be a valued collaborator of Shigeru Miyamoto's, and their first collaborative project was Devil World, a Pac-Man-styled maze game where the player controlled a green dragon. It was Miyamoto's first game that was designed specifically for the Famicom and Tezuka served as co-designer on the project.

Super Mario Bros.

Collaborating with one another through 1984, Miyamoto and Tezuka had established a comfortable working relationship and were keen to continue pushing the boundaries of what Nintendo called the "athletic game" genre (later dubbed "platform games") established by its prior titles. Noting that Mario Bros. continued to be popular, Tezuka suggested their next project make use of Mario and Luigi as well. Alongside Miyamoto, he began design work on the game, and the two began to rethink a lot of the logic that had been established in the original Mario Bros. project, such as the consequences of making contact with various enemy types, the game's setting, and how to work around the Famicom's hardware limitations.

The final product, titled Super Mario Bros., was designed to be a culmination of everything Miyamoto and his team had learnt about game design and the possibilities afforded by the Famicom technology. The player would take on the role of Mario and Luigi, running and jumping past an invasion of turtle-like creatures known as the Koopa Troopas, led by their king, Bowser. The game was set in the "Mushroom Kingdom" and commenced after Bowser kidnapped its princess, Toadstool. Mario set out to rescue Princess Toadstool and defeat Bowser and his army in the process, marking a return to the familiar damsel-in-distress concept.

Super Mario Bros. was released in 1985 in Japan. When the Famicom was launched in North America the following year as the Nintendo Entertainment System, it debuted alongside Super Mario Bros., making it one of the system's most popular launch games.

Creating Mario's Opposite

The Legend of Zelda

In 1984, while Super Mario Bros. was in development, Shigeru Miyamoto began conceptualizing a second game in parallel. Nintendo was about to release the Famicom Disk System, a peripheral for the Famicom that was capable of writing data to floppy disks, and needed to develop a flagship game for the device. Miyamoto felt that it would be interesting to create a game that allowed two players to create virtual labyrinths and then explore each other's creations. A prototype was created, but the overall sentiment was that exploring these labyrinths was more fun than actually creating them, and so the team began rethinking its approach.

In initial designs, the player would enter a dungeon right from the title screen

Whatever form this exploration game ended up taking, Miyamoto felt it needed to be the polar opposite of Mario. Where Super Mario Bros. was linear and made it clear what the player needed to do next, Miyamoto wanted this new game to make players think about where to go and what to do. The same team that worked on Super Mario Bros. was working on this game, and began brainstorming ideas for it.

"As with the Mario series, I came up with the concept for the Zelda series from my adventures as a child exploring the wide variety of places around my home," Miyamoto would recall in an interview with Superplay magazine years later. "There were plenty of caves and mountains. We didn't have that many toys to play with, so I would make slingshots or use sticks and twigs to make puppets and keep myself amused."

The initial design for Nintendo's new game—which they had codenamed "Adventure Title" in their design documents—called for the player to enter labyrinthine levels straight from the title screen, similar to The Black Onyx, which Miyamoto would later cite as a source of inspiration along with Ultima. Since the game also originated from Miyamoto's experiences exploring underground caverns as a young boy, it was originally meant to focus on the exploration of cave-like stages. Namco's maze-like RPG, The Tower of Druaga, was popular in Japan at the time, and it's possible that it served as a point of reference as well. Over time, the idea of exploring caves evolved into exploring a number of labyrinthine areas all connected by a large, open field. The player would be able to traverse this open environment, and would be required to think on their feet about where to go and what to do next.

Early design specs

An early overworld map

"We basically decided to do a real-time adventure game," Tezuka would recall in an interview many years later. "No one wants to do physical things like pushing and pulling by selecting them from a menu [like in command-based RPGs]. If they’re going to push something, they want to put some force behind it."

During development, Miyamoto and his team also began forming an image of just what their adventure game's story was about. They named the main character Link, as he was to be emblematic of the game's setting—a combination of the past and the future, with the player being able to travel between the two settings and serve as a "link" between them.

Link was initially designed to be right-handed, but in order to aid in the creation of the game's pixel art and the way he would appear in in-game screens, he was altered to be left-handed instead.

At some point, Nintendo enlisted the aid of Keiji Terui, a screenwriter that had worked on animated shows such as Dr. Slump and Dragon Ball, to create the game's full backstory. This story would be included in the game's manual and give the player better context as to just what their character's motivations were. Sometime during this process, the idea of time-travel was dropped, and Terui instead penned a much more straightforward setup inspired by medieval conflicts in Europe. Miyamoto's fondness for the damsel-in-distress setup also worked its way into the tale, and the game's damsel was named "Zelda," after the wife of famous novelist F. Scott Fitzgerald, whose name Miyamoto took a liking to. It was her name that would also help solidify the game's title: The Legend of Zelda.

A pond hiding the entrance to Level 7

Emptying the pond reveals the entrance

Miyamoto wanted The Legend of Zelda to evoke a sense of mystery. "I remember he had me make a lot of different sounds for when you use the flute (when you warp)," composer, Koji Kondo, would recall. "He was very particular about that one sound. 'It shouldn’t just be ‘pretty’. I want it to evoke something more mysterious', he told me."

As they created the world of Zelda, Miyamoto and Tezuka began encountering the limitations of the Famicom hardware. Like with Super Mario Bros., the development team attempted to work around these in creative ways, but found that they would need to scrap certain ideas entirely.

"Back then, there were a lot of things we intended to do but weren’t able to because of hardware constraints," Miyamoto would reveal. "For example, for the Level 7 dungeon entrance, we just changed the colour of the ground when the water drained, but we intended to have the water actually disappear. And you can burn small trees, but we intended for you to be able to burn down big ones."

The first ever dungeon map (left) and its equivalent Second Quest map (right)

Mistakes were made during the development process as well. Labyrinths (later called dungeons) in The Legend of Zelda were mapped out on graph paper first. Each square on the graph represented a single room, and the pieces were laid out like a jigsaw puzzle. Tezuka, having created the entire map for the game, handed it off to programmer Toshihiko Nakago, who put the map data together exactly as it had been provided to him. Unfortunately, due to an error on his part, Tezuka only used half the data Nakago had coded, and the game ended up being half its original size. As luck would have it, Miyamoto felt the reduced map size made for a better game, and suggested that the other half of the data be used to create an unlockable "Second Quest" for the player to discover.

Difficulties were encountered with the game's soundtrack, too. Composer Koji Kondo had composed a total of five musical tracks for the game, and had intended to use the classical piece "Bolero" by Maurice Ravel for the game's opening title screen. Unfortunately, just as development was wrapping, Kondo was informed that the copyright to Bolero hadn't expired yet, which meant Nintendo couldn't use it and he was instead forced to re-arrange the game's overworld theme for its title screen.

Despite difficulties during development, The Legend of Zelda was an immense success following its release. It went on to sell well over 6.5 million units on the Famicom Disk System and Nintendo Entertainment System, and served as the template for a new brand of Nintendo games going forward. It birthed the acction-adventure genre and still serves as the inspiration for a number of modern videogames today.

The Adventure of Link

File:TAoL Link Concept Art.png

Link's arms and legs were extended to better suit The Adventure of Link

The Adventure of Link was developed by a different team than the first game, partly because it wasn't meant to be a Zelda game at first. Development began with Shigeru Miyamoto contemplating how a side-scrolling action game that used "up and down movements" for attacking and defending could be fun to play. Miyamoto wanted to include the kinds of actions that couldn't be incorporated in the original The Legend of Zelda into a new game, and Tadashi Sugiyama, a graphic designer at Nintendo that had contributed to games such as Ice Climber and Baseball was attached to direct.

Joining Sugiyama as co-director was Yasuhisa Yamamura, while Takashi Tezuka came up with the concept for the game's story. Tezuka did not contribute map designs as he had on the first Zelda game, and was instead working on Super Mario Bros. 2 and Super Mario Bros. 3. Meanwhile, replacing Koji Kondo as composer was Akito Nakatsuka, who had composed music for Ice Climbers two years prior.

Games that were difficult were popular at the time. In the late '80s, the thinking was that a higher difficulty often helped games last longer, and that this appealed to videogame enthusiasts. Sugiyama and his team approached The Adventure of Link the same way—by designing it to be something they would personally find challenging.

In an interview several years later, Sugiyama would recall: "One thing I remember is a call from a client at the time saying 'I just can't seem to beat the last boss.' When I asked more about his progress in the game, he was already fully equipped. Meaning that there was nothing to do but to beat the boss with his own skill, which was rather hard to say straight out. It seemed that he was playing the game for his son... so I felt bad for him."

The Adventure of Link borrowed a number of elements from role-playing games, which was something that subsequent Zelda titles would do as well. In The Adventure of Link, this took the shape of a stat growth system, meant to encourage players to fight monsters over and over. Link could gain experience points and upgrade his life, attack, and magic stats, and each of these attributes could be raised to a maximum of 8 levels. The game also made use of symbol encounters, since routes on the overworld map were fairly narrow and this would add an element of luck to encountering enemies.

The Fairy Spell could be used to access tight spaces

This wasn't the only idea The Adventure of Link borrowed from role-playing games, though. Miyamoto was already planning a third Zelda, which would be produced by the team responsible for the original. In Zelda 3, Miyamoto wanted to introduce a party system, where the player's party would consist of a fighter-like elf character, a magic user, and a fairy. This fairy character was designed early in the planning process, and while she wasn't used in the third Zelda, her design was utilized for a "Fairy Spell" in The Adventure of Link, which would cause Link to turn into a miniature fairy and access small spaces.

The Adventure of Link was played from two different viewpoints: a top-down view like the original The Legend of Zelda, which was used in the overworld, and a side-scrolling perspective, which was how most of the game was played and how combat encounters would take place. Miyamoto would later state that the hardware limitations of the Famicom had a hand in influencing what the team could achieve with The Adventure of Link, and express regret that the game hadn't been more surprising or interesting.

The Adventure of Link went on to sell over 4.3 million units and introduced a number of new features that would later be used throughout the series, including magic, the Triforce of Courage, the towns of Rauru, Ruto, Saria, Nabooru, and Darunia (which would later serve as the names of the Sages in Ocarina of Time), as well as Dark Link—a shadowy reflection of Link that served as the game's final boss.

Years later, Nintendo would investigate the possibility of remaking The Adventure of Link using polygons on the Super Famicom, but this idea would be shelved and indirectly lead to the creation of the first 3D Zelda game instead.

A Link to the Past

Two games in, Nintendo had determined that interactivity—being able to influence different things in the game world—was of paramount importance to The Legend of Zelda games. What defined a Zelda, Shigeru Miyamoto felt, was the freedom to try different things and fail, until the solution clicked inside your head. The game's logic, above all, needed to make sense.

As the Famicom was reaching the end of its tenure, Nintendo released the Super Famicom. Work on a new Zelda game for the system had begun while The Adventure of Link was still in development and was being undertaken by the team responsible for the original The Legend of Zelda. Unlike The Adventure of Link, Miyamoto wanted a return to the familiar top-down perspective for this third game. The primary concern, however, was that games had come a long way since the original Zelda and features such as fantasy settings and puzzle-solving were no longer surprising or particularly original.

Instead, Miyamoto wanted to place an even greater emphasis on players interacting with the game world and performing different kinds of actions. For instance, the player would be able to stand in front of a switch and push or pull it, by holding down the A button in tandem with a direction. Pushing and pulling would have different effects and would give the player incentive to try these actions out on other objects. It would require thinking and encourage players to take initiative. Most importantly, it would create a greater sense of satisfaction and immediacy than simply selecting commands that would perform the necessary action for you.

File:ALttP Sci-Fi Zelda Concept Art.png

Concept art from Hyrule Historia hints at a scrapped sci-fi element

A small team began work on the new game, experimenting with different ideas along these lines for the Super Famicom hardware, with the goal being to add more staff once the basic systems had been worked out. The idea was to utilize development staff as efficiently as possible, and not end up with a team that was larger than it needed to be. The team dedicated its first year entirely to planning, while the second year of development was spent on experimentation. At one point, there were plans for the player to be able to perform actions such as eating or dancing, but the team ultimately opted to limit player actions to essentials like Talk, Push/Pull, Lift/Throw, and Run.

Another concept Miyamoto wanted to explore early on was the idea of diagonal sword swings—being able to attack enemies at an angle. While experimenting with this concept, the team concluded that it was too difficult an action to ask of the player and circumvented the issue by redesigning Link's sword attack to swing in a wider arc, thus being able to hit enemies that weren't directly in front of him.

The team also wanted to revisit ideas they couldn't use in the original The Legend of Zelda. That game was originally conceived as a time-travel adventure where Link would be able to move between the past and the future, and the same idea was discussed once again for Zelda 3. Takashi Tezuka, who was brought on to serve as director halfway through development, revealed that the team initially experimented with a multi-world structure, where events in the hub world would have an effect on the other overlapping worlds. Three worlds were initially conceived, and concept art suggests that one of them might have involved a sci-fi setting. Unfortunately, the team feared that three worlds might get confusing for players and the idea was shelved. The number of worlds was cut down to just two, and Kensuke Tanabe, the director behind Super Mario Bros. 2, thought up the idea of players switching between a "Light World" and a "Dark World".

"Kensuke Tanabe already had an idea for a truly memorable hero-awakening scene when we started this project," Tezuka would recall. "In the midst of a forest, with light filtering down through the leaves, the [Master Sword] stood waiting for someone worthy of wielding it to arrive. Link draws the sword out as the light trickles through the leaves."

File:ALttP Link Concept Art.png

Link was designed to appear more mature

The team wanted A Link to the Past to be a rousing adventure, one that would appeal to Nintendo's foreign fans. To aid this, they redesigned Link to appear more mature than he did in The Adventure of Link. However, Miyamoto had never wanted Link to appear overly cool and serious, so the character retained elements of playfulness.

The team was also careful not to have Link begin the game with his sword. The way Tezuka and Tanabe envisioned the game, players would feel an emotional connection with Link as they guided him through his adventure, culminating in the aforementioned hero-awakening scene where Link would find and draw the Master Sword from its resting place. To accommodate this goal, weapons and items were carefully bestowed such that players would sense Link's growth throughout, no matter their play style.

File:ALttP Fairy Concept Art.png

Concept art for a fairy character that was initially to be playable

The team even wanted to let players use weapons other than a sword and shield, and even use items like arrows and bombs in tandem, to create bomb arrows. In contrast, Miyamoto felt that Link should always have his sword equipped, and so this idea was shelved until the next game in the series. Miyamoto had also wanted to include an RPG-like party setup where Link would be accompanied by companion characters. Since before the game went into development, he had been telling his team that he wanted the third Zelda to include a mix of characters—a fighter-like elf, a magic user, and a fairy whose role would consist of reconnaissance. This idea was was never used either, and the design for the fairy was instead utilized in The Adventure of Link.

Another idea that was scrapped during development included multiple paths through the world, so that the player's experience would be more open-ended. This idea was abandoned due to memory constraints (with Miyamoto hypothesizing that it would have required 150% more memory than the Super Famicom possessed) and the complications it would have caused in terms of game structure. Time and memory constraints also called for the team to scale back some of the ambitions they had for the game's level of interactivity. For instance, they had wanted to make it so using the lantern on a grassy area would cause an "endlessly expanding fire" or bombing the swamp breakwater would cause water to rush into the hole. Such ideas would later be utilized in games such as Four Swords Adventures and Breath of the Wild.

"Our game designers had a pretty good idea of what could be done on the hardware back then, so I don’t believe we had any unexpected implementations," Tezuka would say in an interview with RetroGamer magazine years later. "Having said that, though, we had a long battle with the memory size, and I remember very clearly that the engineering team worked extremely hard to optimize [the game]."

While Zelda 3 was in development, a new member of staff, Yoshiaki Koizumi, was assigned to work on its manual. A graduate from the Osaka University of Arts—the same university Takashi Tezuka had graduated from—Koizumi was interested in telling stories. Games, he felt, afforded an opportunity to create the kind of dramatic storytelling one couldn't find in film, and had applied to Nintendo in the hopes of creating them. As the years would go by, Koizumi would take every opportunity to sneak as much story as he could into Nintendo's games, and that habit would begin with this new Zelda. Using his art skills, Koizumi would provide a visual identity for some of the game's characters and motifs, establishing them through its manual.

"What was funny was that at the time, it didn’t seem like they’d really figured out what most of the game elements meant," Koizumi would reveal in an interview. "So it was up to me to come up with story and things while I was working on the manual. So, for example, the design of the goddesses as well as the star sign associated with them."

The new Zelda logo by Girvin, a Seattle-based design firm

Once all the pieces were in place and development was nearing its end, the team needed to decide on a title for their new game. In Japan, the game was given the title "Triforce of the Gods," but Nintendo of America changed this to A Link to the Past for the game's western release. The American division wanted to avoid any overt religious references and also had the development team make changes to some of the game's characters and text to avoid controversy. The Hylian script was originally introduced in A Link to the Past, and initially contained alphabets that looked like an ankh and other Egyptian hieroglyphs. These were eventually removed.

A Link to the Past's Japanese logo, still using the original typeface

Whilst localizing A Link to the Past, Nintendo of America also contracted a Seattle-based design firm, Girvin, to create a new logo for the game. Girvin had already worked with NOA to create packaging for the Nintendo Entertainment System and designed the logo that would be used for all subsequent The Legend of Zelda games in the West. The Japanese releases would eventually adopt this new logo as well, but not until the next console game in the series.

A Link to the Past was released in Japan in November 1991, and a western release followed in 1992. The game would go on to sell 4.6 million copies worldwide on the Super Nintendo Entertainment System. It is often considered one of the best 2D The Legend of Zelda games, and introduced a number of elements and ideas that would appear in later Zeldas.

Two Steps Forward

Eiji Aonuma Joins Nintendo

All throughout their success with the Famicom and Super Famicom, Nintendo had remained on the lookout for promising new talent, and a number of capable designers had joined their ranks since. These included stalwarts such as Kensuke Tanabe and Yoshiaki Koizumi—designers that had helped shape the company's games in new and interesting ways, and proved the importance of nurturing young, passionate talent.

A number of these talented developers had also coincidentally graduated from the same university—the Osaka University of Arts—which was situated close to Nintendo's headquarters at the time and made for a convenient hiring ground. However, Nintendo had made a habit of hiring interesting talent regardless of background, and one such hire was neither a fan of videogames nor an Osaka University of Arts graduate.

In fact, he had never really played a videogame prior to being hired by Nintendo. What got him hired was his love for creating mechanical puppets.

Eiji Aonuma

Eiji Aonuma graduated from the Tokyo National University of Fine Arts and Music in 1988, where he had completed a masters in composition design, working on Japanese Karakuri puppets. Aonuma's grandfather and uncle were carpenters, and he had grown up watching them craft things, which had inspired him to do the same as a young boy. Whenever there was drafting or craft homework from school, Aonuma would take the opportunity to use a hammer and nail to make creations of his own. Since his parents weren't in the habit of buying him toys as a child, Aounuma would create those himself as well.

Aonuma's talent for wood cutting guided him through university, and would eventually lead to him being hired by Nintendo. During an exhibition where he was showing off his mechanical puppets, Aonuma encountered members of the videogame industry and took an interest in the kind of work they did. He learnt that Yōichi Kotabe, an accomplished animator that had worked on Heidi: Girls of the Alps, had also graduated from the same university, and had gone on to work with Nintendo, where he had created packaging for Super Mario Bros.. Upon learning this, Aonuma procured Kotabe's business card from his university and the two established contact. Kotabe subsequently recommended Aonuma to Shigeru Miyamoto, who interviewed him and was impressed by his work, leading to his hiring.

Marvelous: Another Treasure Island

After he began working at Nintendo, Aonuma was assigned to the department that made games and served as a graphic designer on NES Open Tournament Golf. Since he had never been interested in videogames prior to joining the company, he turned to his girlfriend at the time and asked her to provide him with an introduction to games. Through her, he was exposed to the first Dragon Quest and the PC version of The Portopia Serial Murder Case, both of which were designed by a rising star at Enix named Yuji Horii.

"I stayed up all night to play it and she kept by my side the whole time, coaching me like, 'You need to go south five steps' and 'Now go to the east four steps,'" Aonuma would recall in an interview years later.

Over time, Aonuma began to appreciate the fun of playing videogames and found himself particularly fond of A Link to the Past. After working on a number of projects in collaboration with external developers (including an unreleased game with one Satoru Iwata of HAL Laboratory), he eventually developed a game of his own: Marvelous: Another Treasure Island. The game was released in 1996 and drew inspiration from the A Link to the Past. Upon playing it, Miyamoto invited Aonuma to work with him and the Zelda team on a new project they were developing for the Nintendo 64.

Link's Awakening

While development on The Adventure of Link was ongoing, a different team at Nintendo released the Game Boy, a portable system capable of playing black-and-white games using cartridges players could slot into the device.

The Game Boy was released in 1989, and launched alongside Super Mario Land—the first Mario game to be developed without the involvement of the Mario and Zelda team, who were occupied working on the Famicom and Super Famicom. Eventually, after development wrapped on A Link to the Past in 1991, the game's chief programmer, Kazuaki Morita, managed to acquire a development kit for the Game Boy and began experimenting with it as a hobby project of sorts. At the time, the Zelda department only had access to a single Game Boy development kit and Morita used to it to create a prototype for a Zelda-like game.

"We weren't particularly planning to make a Zelda game for Game Boy, but we thought we'd try it out to see how it [would] work," Takashi Tezuka would recall in an Iwata Asks interview. "So at first there was no official project. We'd do our regular work during normal work hours, and then work on it sort of like an afterschool club activity."

The opening storyboard

The final intro

As the team discovered what the Game Boy was capable of, Tezuka suggested they attempt porting A Link to the Past to the device, and requested a second Game Boy development kit for their work. Because Tezuka had only joined development of the Super Nintendo Entertainment System game halfway through, he wanted to add new features that couldn't be implemented in the original. As these ideas piled up, the Game Boy game eventually began to morph into a completely original title instead of a simple port. Kensuke Tanabe, who had written the story for A Link to the Past, joined the team early on and began working on characters and sub-events for the game. Meanwhile, Yoshiaki Koizumi, who had worked on the manual for A Link to the Past, was invited to join as well. Koizumi was put in charge of the game's main story, as well as its opening cinematic.

At the time, Tezuka was fond of an American television show named Twin Peaks by director David Lynch, and wanted to design a Zelda game that was similar in feel and scope—particularly how the show depicted a drama involving suspicious characters in a small town. To help achieve this, Tezuka laid out a list of requirements for Koizumi and Tanabe to follow: the game would have no Princess Zelda, no Triforce, would not take place in Hyrule, and be set in a relatively small playing field.

Tezuka would recall: "At the time, Twin Peaks was rather popular. The drama was all about a small number of characters in a small town. So when it came to The Legend of Zelda: Link's Awakening, I wanted to make something that, while it would be small enough in scope to easily understand, it would have deep and distinctive characteristics."

Shigeru Miyamoto had mentored his developers to prioritize fun and intuitiveness above all else, which was why Mario and Zelda games were light on storytelling. However, Miyamoto was focused on other projects at the time, leaving the team free to do as it pleased. Koizumi, who viewed videogames as a medium that was well-suited to telling engaging stories, saw the opportunity to give this new Zelda a sense of drama. Without Miyamoto's oversight, Morita, Koizumi, Tanabe, and Tezuka began to turn their Zelda into a character-driven adventure filled with engaging side events and subplots.

A guest from Mario

Link's Awakening introduced a number of Zelda staples

The team spared no effort in surprising the player, even breaking established boundaries by featuring NPCs that looked suspiciously like Mario and Kirby characters, sometimes without permission from the other development teams. This more casual approach to development eventually led to ideas like a fishing minigame designed by Morita, who was fond of fishing as a hobby and programmed it into the game without being asked. Fishing would eventually go on to be present in nearly every The Legend of Zelda game going forward. Without realizing it, Morita, Koizumi, Tanabe, and Tezuka were establishing the template for all The Legend of Zelda games that would follow, particularly the tendency to draw inspiration from Twin Peaks and the idea of characters with an air of mystery about them.

"I'm certain it was an important element in the series making a breakthrough," Eiji Aonuma would say years later. "If we had proceeded from The Legend of Zelda: A Link to the Past straight to The Legend of Zelda: Ocarina of Time without The Legend of Zelda: Link's Awakening in between, The Legend of Zelda: Ocarina of Time would have been different."

"The first real game work that I did was on Link’s Awakening," Koizumi would recall in an interview. "But at the same time, I came in to write the manual, as I did on the previous game. But they had nothing in place. So I ended up making an entire story to go along with the game. The dream, the island, that was all mine. And so that was my first experience doing the kind of work that we would now call 'event design'. But there were not too many people at the time with expertise in that area, so I really had free reign to do what I wanted, so long as I didn’t make Miyamoto angry."

Link's Awakening took a year-and-a-half to complete and was released in 1993 for the Game Boy. The game went on to sell over 3.83 million units on the Game Boy and is credited as being the first The Legend of Zelda game to tell a proper story. It is directly responsible for the stronger story focus in future Zelda games, beginning with the game developed immediately after it.

Establishing the Conventions of Zelda

Ocarina of Time

Following the release of Link's Awakening, Shigeru Miyamoto and Yoshiaki Koizumi began working on a remake of The Adventure of Link for the Super Famicom designed using polygons. The two had been experimenting with a thin, polygonal Link viewed from a side-scrolling perspective similar to the original game. Plans to turn this concept into a full game eventually fell through, and both developers moved on to other projects. However, the team still wanted to create another Zelda based around swordfighting, the way The Adventure of Link was.

In the interim, two important events transpired that would help shape the future of Zelda. The first was the career path of Eiji Aonuma, an artist that had joined Nintendo in 1988. Aonuma had worked on games such as NES Open Tournament Golf and eventually began working on his own game—one heavily inspired by A Link to the Past—on which he served as director. The second was the development of Link's Awakening, a game developed by a smaller team that would set the tone and aspirations for Nintendo's next major Zelda title.

1995 was when the very first signs of this title would manifest. Nintendo had revealed its next game videogame system, the Nintendo 64, to the public. Alongside the new console, the company announced an add-on peripheral for the device, similar to the Famicom Disk System. Dubbed the "64DD," this was a disk drive that would provide the Nintendo 64 with additional RAM, as well as rewritable memory, that would allow for user-created content to be saved to the disk. Nintendo's first game for the Nintendo 64 was going to be Super Mario 64—the first fully 3D Mario game the company had worked on—but following its release, the company intended to release a new 3D The Legend of Zelda as well.

A tech demo for Zelda 64

A tech demo presenting Link rendered using 3D polygons on the Nintendo 64 hardware was shown to the public at Nintendo's Space World event in 1995, prior to the release of the system the following year. It was programmed by Giles Goddard, one of the programmers on the original Star Fox and the programmer behind the interactive Mario face in Super Mario 64. Yoshiaki Koizumi provided the demo's character models and animation work. Finally, Takao Shimizu, who had co-directed the Game Boy Donkey Kong directed the short reel. After completing work on the tech demo, Shimizu went on to direct the next Star Fox game, Star Fox 64. Shortly thereafter, another Nintendo employee, Toru Osawa, was asked by Miyamoto if he would like to direct the company's next Zelda project for the Nintendo 64. Osawa agreed, and picked up where Shimizu had left off. The desire to create a Zelda game based around swordfighting persisted, and Osawa began penning a script for the project around this idea.

The following year, the Nintendo 64 was released alongside Super Mario 64. Shigeru Miyamoto, Takashi Tezuka, and Yoshiaki Koizumi had been spearheading that game's development, and successfully released it to critical acclaim in 1996. As development on Super Mario 64 wrapped, Koizumi and a programmer named Jin Ikeda joined Osawa, and began to experiment with how they could use their learnings to develop a new Zelda game in 3D. Similar to the way the original The Legend of Zelda was conceived for the Famicom Disk System, the initial goal for this game was to make full use of the 64DD, utilizing its hardware to create a persistent world with lasting effects such as trees remaining cut once the player had chopped them down, or Link leaving permanent footprints behind him wherever he walked. The small team began experimenting with the hardware, using the game engine created for Super Mario 64 and modifying it as needed to help build their new Zelda. Koizumi would be in charge of creating the character model and animations for Link, owing to his experience a character animator and assistant director on Super Mario 64.

An early character model of Link by Yoshiaki Koizumi

The first half of development would be spent experimenting with different concepts. One of the early ideas Miyamoto had was that the game would take place entirely within the confines of a castle. Similar to the castle in Super Mario 64, each room would lead to a different kind of environment such as a meadow or an ocean, and Link wouldn't actually be able to venture outside the castle at all. This idea was shortlived. Another suggestion from Miyamoto involved having the game play out from a first-person perspective, and when you'd encounter an enemy, the camera would shift to a side-scrolling view, similar to The Adventure of Link and other role-playing games. This suggestion was rejected by Koizumi, who was putting a great deal of effort into the Link's animations, and wished for his character model to remain visible at all times. In parallel, the team also decided against using the 64DD peripheral. Like a hard disk, the 64DD consisted of mechanical moving parts, and depending on where on the disk the data was stored, it could take longer to retrieve. This would potentially limit the number of animations that could be programmed for Link, which led to the team's decision. However, without the 64DD the development team would now have access to far less storage space, which created a challenge in itself.

One of the other major problems the team faced was how combat in a 3D space would work. The flow of combat and choosing an appropriate camera angle to depict it was one of the earliest issues the developers found they needed to solve. Since the game was to be based around the idea of swordfighting, Osawa, Koizumi, and Ikeda decided to visit Toei Kyoto Studio park to seek out inspiration and find a solution to their problem.

"As we went along looking at everything, it was so hot that we ducked into a playhouse to cool off," Osawa would recall in an Iwata Asks interview years later. "They were doing a ninja show. A number of ninja were surrounding the main samurai and one lashed out with a kusarigama (sickle-and-chain). The lead samurai caught it with his left arm, the chain stretched tight, and the ninja moved in a circle around him."

• 

  Z-Targeting, inspired by Japanese theatre

This show led the team to conceive of Z-targeting, or what is now known as a "lock-on"—an invisible line connecting the player character and the enemy, that allows one to circle around the opponent and keep them in your sights. Since Z-targeting would ensure that the enemy remained within the player's line of sight, it provided a practical alternative to manual camera controls. (A similar concept had already appeared in Super Mario 64 where, if the player tried to read a signpost, they would sometimes end up going around it in circles.) The ninja show at Toei Kyoto Studio Park also demonstrated to the team that staged swordfights followed a very specific pattern. Enemies would attack the protagonist one at a time, so that he could engage them one-on-one, instead of being overwhelmed. This added tension to every encounter. The combination of this knowledge and the idea for Z-targeting gave the developers a foundation upon which they could begin building the game's swordfighting system. Koizumi designed a lock-on marker that could be used to indicate what object the player was locked on to, and created it in the shape of a simple fairy, which Osawa dubbed "Navi"—referring to someone that would serve as the player's navigator. Without meaning to, Osawa, Koizumi, and Ikeda had provided Link with his first companion character—something Miyamoto had originally wanted to do in A Link to the Past.

The conception of Navi also allowed the developers to deal with some of the memory challenges they were facing. The team decided that in the world of this new Zelda, every character in Link's village would have their own personal Navi-like fairy. This made it so you could get away with just displaying someone's fairy (which consisted of a simple round model) if the player was standing far away from them, and have the actual character model appear as the player got closer. This system, in turn, led to a story idea whereby Link wouldn't have his own fairy companion at the start of the game, meet one during his adventure, and have to part with her at the end of the story. It was a trademark example of Nintendo using game mechanics to dictate story ideas—something that would frequently occur during the development of future Zelda games as well.

An early version of Link on horseback

As ideas for the game's story began to expand, so did the team and its ambitions. During the development of the original Super Mario Bros. and The Legend of Zelda, the team working on the games would brainstorm ideas for both, and categorize them into "Mario ideas" and "Zelda ideas" as they deemed appropriate. During the development of Super Mario 64, Yoshiaki Koizumi had done the same, reserving certain ideas for a Zelda game, and would often refer to these notes. One of the ideas that had been discussed for Mario 64 was the inclusion of a horse that the player would be able to ride. While that idea hadn't made it into Mario, Miyamoto wanted to include it in Zelda, and so the team began designing it. They started by photographing horses for reference and even discussed bringing an actual horse into the studio to help with design. (Eventually, they had to settle for balancing a plank across two footstools) Once the team decided to include a horse, it meant the game would need larger fields that the player could ride across, and so designer Makoto Miyanaga began creating Hyrule Field—a large, central area that would serve as a hub to the game's other locales.

At first, Hyrule Field was densely populated with trees but the team discovered that this would slow the player down too often on horseback, and so the amount of trees was reduced. Routes and roads were laid across the vast landscape of Hyrule Field to help steer players towards places of consequence, and to prevent them from getting lost. The team also began implementing weather effects and a day-and-night cycle, in addition to littering the field with hidden secrets and items for discerning players to find.

"It created quite a fuss when I first made [Hyrule Field]," Miyanaga would recall. "People were like, 'You can't make it that big!' Even riding a horse, it was so big that you would get bored riding around it, so we had to add something. Then lots of people took a hand in it, having enemies appear and putting holes here and there. We'd be like, 'This area's a bit empty, so I'll make a hole and put something in it.'"

Link was made to look "cooler" than in past games

In parallel, attention was lavished on the character model for Link, and no expense was being spared to animate the character convincingly. This Zelda, like The Adventure of Link, featured a more grown-up Link and Koizumi had designed him to look more "handsome" than in prior games at the request of his wife. The character's sideburns were reduced and he was given a sharper nose. Additionally, Koizumi pierced Link's ears to give him a more striking appearance. (Although, he noted it wouldn't have suited Nintendo's style to have Link appear "too cool", and so he also gave Link his familiar long johns.)

"We’d been fussing over how Link should open a treasure chest for three years," Miyamoto would recall in an interview in 1998. "When we got the idea of using motion capture, there were some in the staff who were against it. We ended up deciding that just a little would be okay. My company is sometimes worried about losing money, so when motion capture was suggested we were met with a 'Do you really need that much equipment? Isn’t what you’re doing now okay?' sort of reaction. We started out using wireframe motion capture, but soon we made our own method which actually cost twice as much. But what’s the point of doing something that’s already been done before? When we were photographing horses, we even went as far as discussing how to bring a real horse into the studio. In the end we got two footstools and a plank and making our own horse like that."

"On the day I went to the studio, there was a fantastic iron-frame treasure chest with a sword and shield inside. It clearly had cost a lot of money. When I asked 'What is all this for?' the triumphant reply was 'We figured out how to open a treasure chest!' Their conclusion was that before you opened the chest you needed to kick the hinge first or there’s no way the action looked realistic. I wonder if the motion capture team made that, too… It was really good stuff."

In addition to greatly expanding the world of their first 3D Zelda, the team also recognized the importance of presenting this world in manner that felt both cinematic and easy to control from a gameplay standpoint. At the time, Miyamoto felt that this new Zelda game was already complex and challenging enough, and so the team refrained from giving players the ability to jump. Instead, the game was designed so Link would jump automatically when approaching a gap or chasm. The kind of jump Link would perform was coded into the terrain's data, and this approach allowed the team to dictate other elements of the scene as well, such as how the camera should behave. For instance, when Link approached a cliff's edge or crossed a bridge on horseback, it would be possible to code context-sensitive camera movements into the terrain data without taking control out of the player's hands. This decision would carry forward to every subsequent third-person Zelda game for the next several years.

The costs and workload required to create a game using 3D graphics spilled over to concept design as well. Up until that point, Nintendo had often outsourced illustrations and promotional art to companies that specialized in creating artwork. However, 3D graphics had started to become mainstream and digital art featuring 3D-rendered models needed to be produced for newer games like Super Mario 64. Finding that it wasn't technologically or financially feasible to outsource these renders to other companies, Nintendo had purchased an SGI supercomputer and its artists had begun learning the PowerAnimator tool to create key artwork for Nintendo 64 games. As this workflow was still in place after Super Mario 64 was completed, the team carried it over to the new Zelda as well.

Early character designs

Partway through development, it occurred to Miyamoto that he would like to see a younger version of Link appear in the game. He was against the idea of Link turning into "just another cool hero" and preferred that the character remain playful and childish as he had in most prior Zelda games. At Miyamoto's insistence, the team began exploring the idea of a younger Link and Koizumi began testing the idea of applying the same animations to two separate character models—one older and one younger. The presence of two Links was justified by adjusting the game's story. At some point in the game, the younger Link would draw the Master Sword from its pedestal and would be transported several years into the future, where he would be an adult.

"Link's archenemy is Ganon, so I thought they should meet once when he's a child," Miyamoto would recall. "The innocent eyes of a child are able to see through to the truth, so Young Link knows instinctively that Ganon is a bad guy. When Adult Link meets him again, and Ganon says he's that boy from years before, it really hits you. You think to yourself, 'That's right. I'm that child from before.' Putting in that scene was really fun for me."

As development progressed, Miyamoto began adding more members to the team. The year prior, Nintendo had released Marvelous: Another Treasure Island for the Super Famicom. The game had been directed by Eiji Aonuma, an artist that Miyamoto had personally had a hand in hiring, and was heavily influenced by A Link to the Past. Aonuma had also worked on a number of games in collaboration with external developers such as HAL Laboratory, but wanted to create more in-house titles with teams at Nintendo. Noting his interest in games that were similar to Zelda, Miyamoto invited him to join his team.

Blueprints for the Water Temple

At the time, the Zelda team lacked someone that was capable of designing the game's dungeons in three dimensions. Despite having no prior experience with work of that nature, Aonuma was assigned the task and began brainstorming ideas on paper, trying to maintain a sense of logic within his dungeons while striving to constantly surprise the player. The first dungeon he worked on was the Forest Temple, where the path would twist and warp as Link traversed it. Aonuma would later state that the Forest Temple was his favourite dungeon in the game.

"While playing the previous [The Legend of Zelda game], I tried to put in elements to solve the questions that I had for those games," Aonuma would recall in an interview years later. "For example, it is a terrible rule to restart from the entrance if the player fell inside the a^[sic] dungeon. I made it clear for the player to see the entrance of the room where the boss is as he walks into the dungeon. You can say I put baits."

Aonuma's talent as a puppet designer helped him visualize the logic and moving parts for complex, three-dimensional dungeons, including the game's infamous Water Temple. He collaborated with other members of the development team, routinely tweaking his dungeons to be able to accommodate different items such as the Hookshot, ensuring that none of them would disrupt progress within any of the dungeons. Eventually, Aonuma grew into the role of a systems director and was made one of several sub-directors that would lead different aspects of development.

One of the programmers assisting Aonuma with the creation of his dungeons was Kazuaki Morita, who had spearheaded development of Link's Awakening when he had begun experimenting with a development kit for the original Game Boy. Morita was fond of fishing and had included a fishing minigame in Link's Awakening, among other complex side events. Whilst working on the boss for Aonuma's Water Temple dungeon, Morita noted the presence of a pool-like water body. He happened to have the model of a fish on hand, and began experimenting by having the fish swim around inside the pool. This prompted Morita to develop his second fishing game for a Zelda title and he began fleshing it out, handling elements such as the sound and the music of the minigame by himself.

Meanwhile, other parts of the game had begun to grow in scope and detail as well. Miyamoto was of the opinion that if Link could ride a horse, the team should include mounted archery and one-on-one duels. The team was able to include the former, but not the latter. (One-on-one duels would later be implemented in Twilight Princess, which would serve as a spiritual successor to this game.) Miyamoto's fondness for making Zelda games highly interactive had persisted as well, and the team was working on adding minute details to objects such as wooden signboards. At Miyamoto's behest, the team had designed it so that swinging your sword vertically would cut wooden signboards in half, while swinging it diagonally would cut them a different way. If the cut pieces of wood happened to fall into a body of water, they would float. Just as with A Link to the Past, Miyamoto pushed for a strong sense of interactivity, and other members of the team such as Aonuma and Morita shared his obsession with detail, leading to constant iteration and polish.

Talon from Ocarina of Time

This attention to detail spilled over into the game's story and characters as well. Yoshiaki Koizumi and Kazuaki Morita had been in charge of the story and events for Link's Awakening. Inspired by the characters and atmosphere of an American TV named show Twin Peaks, Link's Awakening had featured a cast of peculiar and memorable characters. Since the team for this new Zelda consisted of members of the Link's Awakening team, the staff pushed further in that direction. An owl-like character similar to the one in Link's Awakening was included, as were others inspired by those in the prior game—most notably Talon and Malon, a father-daughter pair that bore a striking resemblance to Tarin and Marin from the Game Boy game.

"When we decided to handle Link growing up from a 9-year-old child to a more mature 16-year-old, I wanted lots of characters to fulfill various roles," Miyamoto would recall. "For example, [the owl] Kaepora Gaebora is a grandfather figure who gives Link all kinds of advice and looks out for him. And since Link is a boy, I wanted girls besides Princess Zelda to show up."

The latter half of development on the new Zelda was spent almost exclusively on adding content to the game in the form of dungeons, events involving the ocarina songs, and other sub-events. At its peak size, the development team grew to 120 people—forty or fifty of which were in-house Nintendo employees, with the remaining staff composed of people from external companies. The end result was a game that had a large number of staff working under five directors, one of whom was Yoichi Yamada—the co-director of A Link to the Past.

File:OoT JP Logo.gif

Ocarina of Time's Japanese logo

Following multiple days meant to allow time to polish the game, the team was under pressure to release "Zelda 64" by fall of 1998. The Nintendo 64 was facing stiff competition from the PlayStation, and Nintendo needed a major release to keep their platform competitive. After close to three years of development, the game was released under its final title, Ocarina of Time, in November 1998. Starting with this game, Nintendo would use the new logo designed by Girvin for The Legend of Zelda brand in Japan as well, and it would remain in use for the next several years.

While it couldn't turn the fortunes of the Nintendo 64 around, Ocarina of Time is generally considered one of the best and most memorable games in the series. It created the template for a number of 3D action-adventure games and for a number of 3D Zelda games after it. Selling over 7.6 million units, Ocarina of Time would become the benchmark by which all future Zelda games would be measured for years to come. Its immense success and cultural importance would cast a long shadow over the franchise for years thereafter, and would lead to the retooling of The Legend of Zelda a number of times, as Nintendo would attempt to create a game that could surpass their first 3D Zelda.

Majora's Mask

By 1998, Nintendo found itself faced with a new challenge. Over the last few years, developing videogames had become an expensive prospect with the advent of 3D technology. Budgets had ballooned, team sizes had increased, and the amount of time, money, and effort required to create games was on the rise. Developers needed to learn to manage 3D cameras, physics, lighting, and a number of other technical minutiae that had been introduced with games like Super Mario 64, Ocarina of Time, and Eidos Interactive's Tomb Raider, and player expectations were at an all-time high.

Nintendo in particular had felt the full brunt of creating a big-budget 3D blockbuster with Ocarina of Time, and it had happened at a time the company was facing stiff competition from Sony's PlayStation. The PlayStation had deprived the Nintendo 64 of the third-party support previous Nintendo platforms such as the Famicom and Super Famicom had enjoyed, and the console had been getting by primarily off the back of Nintendo's own games.

The writing was on the wall—Nintendo needed to be able to release games faster if they wanted to continue supporting the Nintendo 64 in any reasonable capacity until their next platform was ready. With this in mind, Shigeru Miyamoto decreed that the company needed to do more with less—to create high-selling games with lower budgets than Ocarina of Time and Super Mario 64, and much quicker turnarounds.

"I feel there is a bad atmosphere that you can't do something new at Nintendo these days," Miyamoto would say to Japanese magazine 64 Dream. "I never thought things like this before. So now we are changing ourselves to an organization that allows people to do new things and energize ourselves. I'm saying to my people that from now on let's go for the game that can be developed within six months and sell a million copies. If you want to finish a game within six months, you have to make it within two months because you need to polish it for another four months. If someone asks me who can make such a thing, I'd tell them that I used to do it. It isn't a great thing to take three years. [Ocarina of Time] would have been finished in a much shorter period if we had cut some parts."

Ocarina of Time was originally meant to be compatible with the Nintendo 64DD, a Japan-only peripheral for the Nintendo 64. While the DD wasn't ready in time for that game's release, it was on track for release in the year 2000. By plugging in this Disc Drive underneath the console, it allowed the system to expand and rewrite a large amount of data. Nintendo had already released an updated version of Link's Awakening for the Game Boy Color, and Miyamoto asked his team to do the same with Ocarina of Time for the 64DD, in order to give the Nintendo 64 a second Zelda game in a short amount of time. This project, titled "Ura Zelda," was meant to use remixed dungeons from Ocarina and add other enhancements such as fleshing out unresolved plot threads. Eiji Aonuma, who had designed the dungeons for Ocarina of Time single-handed, was put in charge of the project.

However, Aonuma quickly grew bored at the prospect of remixing his older designs from Ocarina of Time and began working on new dungeons instead. Eventually, he mustered up the courage to request that he be allowed to create an entirely new Zelda game, and was granted the permission to do so—provided he could manage it within one year.

"It’s a shame when a game takes 3 years to make. So, I figured, why not do it in 1?" Miyamoto would say to Japanese publication Hobo Nikkan Itoi Shinbun. "I wanted to be able to say 'We can do it too!' I thought that if we just used the engine for Ocarina of Time and layered a new scenario on top of that, we’d be able to create a reasonably large game in 12 months."

• 

  Zelda Gaiden was to be a darker, more mature game

Development of this new Zelda game began using just half the team behind Ocarina of Time, with a few new members bringing the total count to 30 - 50 developers at the outset. At the time, Yoshiaki Koizumi had been working on a board game that loosely involved the concept of playing with time, inspired by the movie Run Lola Run. In in-game time, this game took place over the course of one week, but could actually be completed by the player in an hour through the manipulation of time. The idea was to develop a compact game that could be replayed over and over. Eventually, the game was cancelled and Koizumi recalled by Miyamoto to work on the new Zelda alongside Aonuma.

Having directed a game entirely by himself before, Eiji Aonuma was put in charge of the overall project as the supervising director and the director in charge of the entire overworld. Meanwhile, Yoshiaki Koizumi was put in charge of sub-events and characters. Mitsuhiro Takano was in charge of the game's script. Kenta Usui involved with dungeon design. Yoichi Yamada was head of system management. Finally, Takumi Kawagoe was the director in charge of cutscenes. With this team in place, development began in earnest and the new game was tentatively titled "Zelda Gaiden". Aonuma and Koizumi used the latter's ideas for the time travel game, refined them, and began putting them into Zelda Gaiden, creating the Groundhog Day-like three-day system in the process. The idea was to create a game that took place in a single location with fewer dungeons, but provide a greater sense of depth and replayability. Rather than focusing on a story that was grand in scope, this new Zelda would involve Link being intimately involved with the inhabitants of a single town.

Majora's Mask promotional art by Yusuke Nakano

The Twin Peaks influence returned for a third time, and in a much more obvious manner than in Link's Awakening and Ocarina of Time. The team had decided from the outset that it wanted to create a darker, more mature Zelda game and this was reflected in the game's characters, as well as the artwork. Nintendo's Yusuke Nakano, who had served as an illustrator on Ocarina of Time, was a fan of an overseas comic book at the time, and drew Link with dark shadows across his face and body to reflect the new game's tone. Meanwhile, Koizumi's time-travel concept called for the game to play out over the course of a week, but the team felt that a week might be too long and the townsfolk's schedules might become too hard for players to remember. The Groundhog Day idea was trimmed to three days instead, which better fit the idea of Zelda Gaiden being a more compact experience. In parallel, the team also began developing sub-systems that would tie into the time-travel mechanic.

"The development of Ocarina of Time was so long, we were able to put in a whole lot of different elements into that game," Aonuma would reveal in an Iwata Asks interview several years later. "Out of those, there were ideas that weren't fully utilized, and ones that weren't used to their full potential. One of those was the mask salesman. So in Majora's Mask we felt it would be fun if Link himself transforms whenever he puts on those masks. As a basis of Zelda games, you're able to use items to do all sorts of different things, and we felt it would be a lot of fun if Link would acquire all these abilities by putting on these different masks. We felt that would expand the gameplay. So we made the game so Link could transform into Deku Link to fly in the air, Goron Link to roll across land, and Zora Link so that he could swim underwater. We also gave each of them a storyline. Once we decided we were going with masks, everything just came into place."

Zelda Gaiden, which would later be titled Majora's Mask, began to take on a sense of challenge. While Ocarina of Time was designed to be a "hospitable" experience for the player, Majora's Mask would challenge them to see if they had what it took to complete it. The game was being designed for players that had already played Ocarina of Time, and so the thinking was that a more challenging affair would be appropriate. However, it quickly became evident that a team this size would not be able to produce a new Zelda in a year. As a result, Miyamoto and Aonuma began to pull other members of the Ocarina of Time team into the project. By the time the complete team was assembled, there were six directors working on the game, similar to Ocarina of Time's multi-director system.

Meanwhile, the notion of using the Nintendo 64DD had been shelved a second time, owing to the device being delayed, and the team instead opted to make use of another peripheral: the Expansion Pak. This smaller, more compact device added an additional 4MB of RAM to the Nintendo 64, bringing its total memory to 8MB. The plan was to release the Expansion Pak as part of the Nintendo 64DD package, but due to the delay of the latter device Nintendo opted to release the Expanion Pak sooner. A demo for Zelda Gaiden that was 50% complete was shown off at Nintendo's Spaceworld event in 1999 and by this point, the central theme involving masks had been implemented, and attendees had a chance to sample the game's story, set in the land of Termina—a parallel world that was not connected to Hyrule in any manner. The 4MB Expansion Pak had been implemented as well, allowing for higher resolution textures and fully 3D interiors, as opposed to the pre-rendered interiors found in Ocarina of Time. It came off as a different kind of Zelda game, created by a confident team with no inhibitions.

At the same event, Shigeru Miyamoto informed the media that "Ura Zelda" was still in production, and was still meant to be be compatible with the 64DD upon its release. The game, he said, had been put on hold so that the team could concentrate on Zelda Gaiden, but would be finished and released as a remixed version of Ocarina of Time with randomized elements at some point in the future.

"This time there's a fear in Zelda"

As development on Majora's Mask progressed, members of the team began inserting more of themselves and their families into the game. Jason Leung, the screen writer for the English version of the game, revealed in an interview: "Normally, we wrap things up around 10 p.m., but tonight we finished up early since Mr. Miyamoto was taking the Zelda team out to dinner. There, game system director Eiji Aonuma and supervisor Takashi Tezuka told me how they've incorporated things from their everyday lives into the game. Development began in August, 1999 (though ideas for a sequel began right after Ocarina of Time was finished), and the team rarely got to go home. As a result, many of the characters—like the Deku Scrubs, who are involved in a cross-country trading sequence—talk about not being able to spend time with their wives. During the development process, the programmers would often say, 'Let's not bring my wife into this,' which was their way of saying that they didn't want to be reminded of their home life. They already felt bad that they were spending so much time at the office to work on perfecting the game. As a little in-joke, Mr. Takano scripted that the mayor in the game says "Let's not bring my wife into this," during his exhausting, overlong council meeting."

The end result of these efforts was a sequel to Ocarina of Time that, while built on the same underlying technology, was incredibly unique. In fact, Majora's Mask's controversial Groundhog Day mechanic even ruffled feathers among Nintendo's in-house debugging team. In an interview, Miyamoto would state: "Even though it's a forbidden thing for a Zelda game, we still decided to have the clock tick in the dungeons. When we first sent it to the Mario Club, we had loads of angry feedback saying ‘It doesn't fit Zelda!’ But after a while that feedback would change to ‘It's actually a good thing’."

Following a year of strenuous overtime and crunch, Majora's Mask was released on April 2000 in Japan. At the suggestion of Nintendo's then-president, Hiroshi Yamauchi, the game used the marketing slogan "This Time There's a Fear In Zelda," indicative of its darker tone. A U.S. and European release followed later that fall, and the game went on to sell just 3.36 million units worldwide—less than half the sales of Ocarina of Time. While this was largely down to low sales of the Nintendo 64 platform itself, it would emerge in the years that followed that the development philosophy behind Majora's Mask was the beginning of an identity crisis that would afflict The Legend of Zelda brand for the next decade. As Nintendo would struggle to determine its priorities, Zelda would go through a tumultuous few years.

The Capcom Partnership

The Formation of Flagship

In 1997, prior to the release of Ocarina of Time, Yoshiki Okamoto, a veteran designer at Capcom, had just established a new studio dedicated to creating games for multiple platforms. During his time at Capcom, Okamoto had directed a number of well-respected games such as Final Fight and the incredibly popular Street Fighter II, and was also the supervisor overseeing the company's new Resident Evil series of games, the first of which had just been released to immense success on the PlayStation.

Okamoto's new studio, named Flagship, was set up as a subsidiary under Capcom and was jointly funded by Capcom, Sega, and Nintendo. The idea was that Flagship would specialize in developing story scenarios for Sega Saturn and Nintendo 64 games in collaboration with the three companies. Among these projects was a game meant to serve as a prequel to Resident Evil, titled Biohazard Zero (eventually better known as Resident Evil Zero). The idea for the project initially came about due to the Nintendo 64 hardware itself, when Capcom's designers discovered that Nintendo 64 cartridges would allow for quick switching between two separate playable characters without the need for a loading screen like with the CD-ROM format used by the PlayStation. Despite starting production on the Nintendo 64, Resident Evil Zero would eventually be rebooted as a Nintendo GameCube project and released in 2002.

In 1999, while Flagship was involved in the scenario creation for Resident Evil Zero, Okamoto approached Nintendo about the possibility of the studio contributing to first-party games, including a new The Legend of Zelda. Nintendo's Shigeru Miyamoto agreed to his proposal, and the staff at Flagship that were free began drafting a story for a remake of the Nintendo Entertainment System's The Legend of Zelda on the Game Boy Color. Okamoto would supervise the project, with the intent being to introduce a new generation of players to the appeal of that first Zelda game.

While Flagship was starting with a remake of the first The Legend of Zelda, the team planned to use that project as a launching-off point for entirely new Zelda games afterward. Okamoto estimated that porting the first Zelda over to the Game Boy Color would take three or four months. Following this, the team would use the same infrastructure used for the remake to create two entirely new games, with the stories of all three being connected in some fashion. The three-game trilogy would be titled Legend of Zelda: The Mysterious Acorn, with the three chapters titled Chapter of Courage (the remake), Chapter of Power, and Chapter of Wisdom—named after the three pieces of the series' Tri Force artifact. Unfortunately, his vision wouldn't come to pass the way he had imagined.

Oracle of Seasons and Oracle of Ages

The Mysterious Acorn

The first screenshots for The Legend of Zelda: The Mysterious Acorn appeared in Famitsu magazine in August of 1999. Nintendo announced that they would demo the new Game Boy Color trilogy at their upcoming Spaceworld event, along with Zelda Gaiden (which hadn't yet been named). The Game Boy Color titles, Nintendo said, comprised a trilogy that could be completed in any order, with each game having the ability to affect the stories of the other two. The following year, all three games were given tentative English titles: The Mystical Seed of Power, The Mystical Seed of Courage, and The Mystical Seed of Wisdom. Nintendo intended to release one of the three games in late summer, with the second following in early Fall, and the final game in time for Christmas.

Capcom designer Hidemaro Fujibayashi was in charge of development. Yoshiki Okamoto, founder of Flagship and supervisor of these new Zeldas, had originally assigned Fujibayashi to serve as his assistant of sorts on the trilogy. Fujibayashi would compile the ideas Okamoto and his team had, and eventually used them to write the original proposal Flagship had presented to Nintendo. Eventually, Fujibayashi began taking an active part in development himself and was promoted to director by Okamoto. Fujibayashi then began approaching Capcom artists and programmers to put a full-fledged development team together—one capable of developing the entire game instead of just its story.

Link from Oracle of Seasons

Link from Oracle of Ages

"The core of the games was pretty much decided," Fujibayashi would recall in an interview. "That is to say, the fact that [the games] would be on the Game Boy Color, the use of the four seasons, and the decision to retain the feel of the 2D Zelda games. It was also decided that it would be a series, so I thought the link system up as a way to make use of that idea. I wanted, for example, that if you missed an enemy in the first game, you would encounter it in the next one. That’s the kind of game I wanted to make it. Zelda is a game with a solid world, so I thought we could express the characters’ 'existence' like in the N64 games on the Game Boy, too."

"We wanted to go in a different direction from the big serious story games like Final Fantasy," Okamoto would say. "This is an action-oriented RPG. It's a 'lighter' style, kind of like a weekly TV drama (as opposed to an epic film). We knew that we could use the same basic style as the existing Zelda games and make two really fun games. We also liked the possibility of having multiple endings and the replay value that you get from two linking games."

By mid-2000, the project had begun to run into complications. One of the three games in the trilogy had been shelved. The original intent was to remake the Nintendo Entertainment System's The Legend of Zelda for the Game Boy Color, followed by two new games. The idea was that Capcom and Flagship would use a remake as a test bed of sorts and eventually develop two entirely new games using the same underlying technology. All three games would serve as a connected trilogy and allow the player to complete them in any order, using a password system to keep track of progress. Unfortunately, the Flagship team, eager to jump right into their new games, had failed to account for unforeseen complications on the remake.

Prior to development kicking off, Flagship hadn't accounted for the fact that the Game Boy Color used a narrower screen than a television. As a result, the GBC wasn't able to display rooms built for the Nintendo Entertainment System's Zelda in their entirety. The player needed to move around to be able to view the full extent of the room they were in, which meant it was easy to miss details such as stairways, cracks in the wall, and other similar clues meant to steer progress. Additionally, because Flagship primarily specialized in writing story scenarios for videogames, the team had trouble reconciling its story ambitions with how the game would actually play. As a result, the developers would constantly need to rework the story and environments to fit one another.

Eventually, it was decided that Capcom and Nintendo would only release the two new games, both taking place in a new setting: the land of Holodrum. When the team was about 60% through development, Nintendo's Yoichi Yamada joined the project in a supervisory capacity, and Fujibayashi began consulting directly with Yamada and Miyamoto, strengthening his relationship with Nintendo in the process.

Since the two games—finally dubbed Oracle of Ages and Oracle of Seasons—were being developed alongside Majora's Mask, references to that game, as well as its predecessor, were included. Oracle of Ages contained characters that appeared in Majora's Mask, while in Oracle of Seasons the player would meet characters from Ocarina of Time. Flagship and Capcom's development staff would also insert their own personalities into the game, with Fujibayashi later stating that the Oracle games were indicative of the differences between the people of Kyoto (where Nintendo was headquartered) and the people of Osaka (where Capcom's office was located). For promotional art, Nintendo's Yusuke Nakano collaborated with Flagship to design the game's characters. Nakano would listen to how the development team envisioned the in-game sprites and create illustrations based on these notes.

Despite going from three games to two, the development team still had its work cut out for it, with team members crunching to have the project completed on time. "There’s a 'Black Tower' in Oracle of Ages, with people made to work there," Fujibayashi would reveal in an interview. "Their dialogue is along the lines of 'There’s no end to this work' or 'I can’t go home'. There were also team members that couldn’t go home much during development, so we put those characters in as a parody. But our team feels really cozy, so the general atmosphere was great. People who’d just come by with a message would end up in a meeting and chat with us for two hours before leaving again."

Oracle of Seasons

"After we started to produce a three-title concept, where players would reach the same goals no matter in which order they chose to play the games, it was difficult for us to see all of the problems in making three linking games," Okamoto would reveal in an interview. "When Mr. Miyamoto said, 'Wouldn't it be simpler to create two titles, instead of three?' we said, 'Yes, of course!' He really saved us. Then, we moved in the direction of the two-title concept. To be honest, I think that it would've been impossible to develop three titles like that. Even now (with two titles releasing simultaneously) we are working very hard to prevent program bugs."

Oracle of Seasons and Oracle of Ages were eventually released on the same day, first in Japan and then in North America and Europe. As intended, the games were designed to be played in any order, with the second serving as a sequel to the first. Due to the fact that they'd been delayed, both were published just a month prior to the release of Nintendo's next portable platform, the Game Boy Advance. They went on to sell a combined total of 3.96 million units worldwide and would lead to director Hidemaro Fujibayashi joining Nintendo as a full-time employee of the company years later.

The Onset of Gamer Drift

Spaceworld 2000

The Zelda tech demo for the Nintendo GameCube

In May 1999, prior to the release of Majora's Mask and Oracle of Seasons and Oracle of Ages, Nintendo revealed the codename for its next home console: "Project Dolphin". The console's GPU was being developed in collaboration with a company named ArtX led by Dr. Wei Yen, who had been responsible for the Nintendo 64 graphics chip. ArtX was later acquired by ATI, which would remain the graphics provider for Nintendo consoles for the next several years. Meanwhile, for the Dolphin's CPU, Nintendo partnered with the technology firm IBM, which produced a custom chip named "Gekko" to power the device.

As development on Majora's Mask was wrapping up, the development team began working on plans for the next Zelda title, which would be released for Project Dolphin. Coming off the failure of the Nintendo 64, there was pressure for the system and its games to perform well, and the team experimented with different art styles and rendering techniques to determine what the next The Legend of Zelda game should look like. By this point, Eiji Aonuma, the director of Majora's Mask, had been tasked with directing the next home console Zelda game as well, and was slowly gaining the confidence to take the reins of the Zelda franchise.

In August 2000, Nintendo officially revealed Project Dolphin, dubbing it the "Nintendo GameCube". At E3 the following year, they unveiled fifteen launch titles for the platform, including Luigi's Mansion and Star Wars: Rogue Squadron II. Alongside these, Nintendo also presented a tech demo for what a The Legend of Zelda game could look like on the Nintendo GameCube hardware, similar to their Nintendo 64 tech demo years ago. The demonstration featured an adult-looking Link, modeled after his Ocarina of Time design, engaging Ganon in a one-on-one duel. The short reel was extremely positively received, with most assuming it represented how the next The Legend of Zelda game would look.

The Wind Waker

The Zelda tech demo for the Nintendo GameCube

By late 2000, work on Majora's Mask had been completed and the game released. While it had been well-received, there wasn't much Majora's Mask—or any other game—could do to bolster sales of the Nintendo 64, which had ceded both the Japanese and Western videogame markets to Sony's PlayStation. Knowing the Nintendo 64 was on its way out, Nintendo had revealed its next videogame console, the Nintendo GameCube, to the public. Members of the Zelda team had already begun experimenting with this new system, and Nintendo intended for them to get a game out as quickly as possible.

Under the direction of Eiji Aonuma and Yoshiaki Koizumi, Majora's Mask had set itself well apart from Ocarina of Time. The game had been a happy accident of sorts, with the development team co-opting ideas that were originally meant for other projects and fleshing them out to give Majora's Mask its unique three-day system. Since the team had also re-used a number of art assets originally created for Ocarina of Time, they were able to complete development of the game in about a year.

This time, they wouldn't have the same luxury. For starters, the team had lost Koizumi, who had been tasked with directing the next Mario game. Furthermore, the team felt it had exhausted its ideas on the Nintendo 64 Zeldas and was unable to come up with new ones using those games as a template. Earlier in the year, a few members of staff had even produced a tech demo for the Nintendo GameCube's public reveal that depicted a semi-realistic Link and Ganondorf locked in a duel, but while the short reel had fans excited the team felt it was uninspired and looked too similar to Ocarina of Time.

Character designers Satoru Takizawa and Yoshiki Haruhana—both of whom had worked on the Spaceworld 2000 tech demo—felt the need to create something entirely new instead. Takizawa had been in charge of creating enemies for both Ocarina of Time and Majora's Mask, while Haruhana had been in charge of NPCs. The two wanted to give the next Zelda a unique visual identity.

• File:Toon link.jpg

  Toon Link by Yoshiki Haruhana

• File:Moblin TWW Concept.jpg

  A Moblin by Satoru Takizawa

Eventually, it was Haruhana that found a way forward, when he presented the team with a cartoon-ish illustration of Link. According to Haruhana, he had been browsing through a videogame magazine and felt that all the games within looked too similar to one another. He grew concerned that the next Zelda would end up feeling the same way, and began thinking about what needed to be done with its art to make it stand out. The new Link design he presented to the team accomplished just that, and sparked the imaginations of the other artists. In turn, Takizawa illustrated a Moblin character fashioned after the new Link's stylized look, believing that this new style would allow for more striking animation. Animation tests soon followed and the team began experimenting with how this new Link and the Moblin would fight. The resulting demo reel convinced the designers and director Eiji Aonuma that this was the way to go, and that cel-shading (or toon-shading) techniques would help achieve the visuals they envisioned.

"At the time, when the GameCube came out, within the computer graphics world, within the industry, 'toon shading' was kind of a buzzword," Aonuma would recall in an interview several years later. "There was a lot of chatter about it, but no one had really explored it in games yet, at the time. The staff that I work with was curious, so we challenged it. We tried it."

File:WW Link concept.jpg

Unused concept art for adult Toon Link

Link on the King of Red Lions

Inspired by their striking new visual style, the development team began brainstorming ideas around Zelda's core foundations: interactivity and an in-game logic that would make sense to the player. One of the advantages of the stylized new visuals was that they made it easier to highlight in-game objects and puzzle mechanics in a way that would stand out.

The belief among the team was that photorealistic visuals made it more difficult to convey things to the player. For example; if you needed the player to bomb a breakable wall, it was harder to make the wall stand out from non-bombable walls if the artwork was overly realistic. And so, the team leaned into the idea of a highly-stylized, animated world filled with moving parts and visual flourishes. Moblins would have strands of rope hanging from their spears, while suspension bridges would feature them heavily as well. In order to make these strands animate convincingly, a programmer that had been in charge of Majora's Wrath, an enemy in Majora's Mask that used whips, would be assigned to them.

Animator Yōichi Kotabe, who served Nintendo in an advisory capacity, would later reveal that the game's visual style was heavily inspired by Toei Studio's Wanpaku Ouji no Orochi Taiji anime, which he had worked on in 1963.

The development team decided to set this new Zelda among the seas early in the development process. The team envisioned what sort of characters would reside in such a world and used the open sea to design the game world’s mechanics. They began working out just how large they could make the sea without overburdening the Nintendo GameCube hardware. The Nintendo GameCube couldn't load environments in quickly enough if islands were too large, placed too close together, or if Link approached them too quickly. In order to make transitions seamless between sailing and docking at an island, the team experimented with the size of the sea as well as the placement and sizes of the islands themselves.

Similar to Ocarina of Time and Majora's Mask, gameplay elements helped define this new Zelda's story. Once the team had decided to set the game on the sea, they came up with the idea that Hyrule was now at the bottom of the ocean. This led to the question of how exactly Hyrule had been flooded to begin with, which Aonuma used as an opportunity to position the game's story as a sequel to Ocarina of Time, set hundreds of years later. Aonuma was fond of storytelling in his games, and would later confirm that Ocarina of Time had two endings, one of which connected to this new game.

By August 2001, the development team had made enough headway on the game for Nintendo to show a clip off at Space World 2001, followed by a playable demo at E3 the next year. Reactions to the game were mixed, due to the fact that it looked so different from the Space World 2000 tech demo, which had employed a more realistic, mature visual style. To many, the dramatic change in the graphics indicated that Nintendo was trying to appeal to children, instead of expanding upon the edgier style of Ocarina of Time, which had cemented itself as the standard for Zelda games on home consoles. This perception was something the company would spend the next several months combating, up until release. Meanwhile, the split opinions also made their way back to the development team, who acknowledged the controversy but pushed ahead with development, hoping to win the public over with the finished product.

Several years later, late Nintendo president Satoru Iwata would recall: "If I think back, people were cleanly split into two groups. With one happy and saying 'The characters are so expressive that it’s like I’m controlling an anime,' and another resisting it, saying 'It’s like a game for small kids with the characters this cute.'"

"We never hesitated in our desire to make a completely new Zelda game," Aonuma would reveal. "But we did notice the negative reaction when we announced it, so we were uneasy. But developing the game timidly would have been the worst thing, so we plunged ahead, determined to go all out hoping to gain acceptance."

"Ura Zelda" resurfaces as a pre-order bonus

Development on the game—dubbed The Wind Waker—was completed in late 2002. In order to complete it on time, the development team had to cut two entire dungeons from the game and replace them with a quest to recover Triforce pieces scattered around the sea instead. The shortened development cycle would later lead to criticism of the game's overworld, which players felt was empty and lacking in things to do. Nintendo marketed The Wind Waker heavily as its release drew close, advertising it as "animation you could touch" in Japan. To spur pre-orders of the game, Nintendo also offered a bonus disc to fans, containing a Nintendo GameCube port of Ocarina of Time. The port contained a bonus mode titled "Master Quest," which was based on Ocarina of Time's previously unreleased expansion, Ura Zelda, which had been in development for the Nintendo 64. As a result, The Wind Waker quickly saw Nintendo's most successful pre-order campaign in history, with Nintendo of America announcing that the game had seen 560,000 pre-orders in North America alone.

While pre-orders and initial sales for The Wind Waker were satisfactory, they began to lose steam quickly afterwards, growing to just 4.43 million units worldwide—far lower than Ocarina of Time's 7.6 million. In Japan, this was largely attributed to the fact that the videogame market had started to decline. Meanwhile, in the west, sales were low primarily due to The Wind Waker's cartoon-ish visual style, which had proven unpopular with The Legend of Zelda's audience in North America.

By this point, Aonuma had a sense of what Zelda's audience expected from each subsequent game. Aonuma felt that for a game to succeed, two things were necessary:

1. To introduce fun, new elements that would surprise the player

2. To retain the elements that players found appealing in prior games

The problem with The Wind Waker was that it hadn't struck the balance of new-versus-existing appeal well enough. Nintendo's unintentional misdirect with their Space World 2000 tech demo had led players to believe that they would expand upon the moody visual style introduced in Ocarina of Time. When the actual game ended up going in an entirely different direction, it alienated a large portion of the audience that expected future games to improve upon Ocarina of Time.

However, Shigeru Miyamoto felt that there was another reason that The Wind Waker hadn't been able to find a large audience. From his perspective, the Zelda team hadn't been able to add any truly new ideas to the core gameplay since the series had gone 3D. This, he felt, had resulted in seasoned gamers growing tired of the formula, while those that weren't fond of videogames found Zelda too complicated—especially in Japan, where the market had begun to decline. Miyamoto's style of developing games hinged on reinvention rather than refinement, and he felt it was important for Nintendo to constantly seek out new audiences with new ideas.

Whether his theory about The Wind Waker was applicable outside Japan or not, Miyamoto's outlook trickled down to the development team and for the next several years the Zelda team would struggle to balance the needs of its western and Japanese audiences. Miyamoto would encourage the staff to think up new gimmicks and ideas to combat Japan's declining videogame market and this would often lead to games that felt misguided or at odds with the western audience.

Four Swords

File:ALttP&FS Sewer Passageway Artwork.png

A Link to the Past

Following the release of Oracle of Seasons and Oracle of Ages in 2001, work began on a new portable Zelda game, this time for the Game Boy Advance, which had been released the same year. The new game would be developed by Capcom in collaboration with Nintendo once more, and Hidemaro Fujibayashi would direct. Shortly after it began development, however, the project was put on hold in favour of getting another game out first: a remake of A Link to the Past.

A remake of an older Zelda for one of Nintendo's portable platforms had been a long time coming. Nintendo had already experimented with the idea of remaking A Link to the Past for the Game Boy, but that project had eventually transformed into Link's Awakening. Later, Flagship had attempted to remake the original The Legend of Zelda for the Game Boy Color, but that project resulted in Oracle of Ages and Oracle of Seasons.

This time, there was a simpler, clearer vision in place: A Link to the Past would be re-released with minimal changes for the Game Boy Advance, and an entirely separate multiplayer game dubbed Four Swords would be included as bonus material.

The four Links

Nintendo had grown increasingly wary of a phenomenon dubbed "Gamer Drift" in Japan. It referred to the ongoing decline of the videogame market, and the company had been experimenting with different ways to combat it across their various franchises. In the case of Zelda, one of the experiments the main development team had attempted was allowing the Nintendo GameCube game, The Wind Waker, to connect to the Game Boy Advance, similar to the way Pokémon games were able to connect across home and portable consoles.

Nintendo wanted to explore the concept of connectivity differently with a separate Zelda game, which led to the creation of the Four Swords mode. Daiki Iwamoto, who had created cinematic sequences for Ocarina of Time was one of the programmers in charge of porting A Link to the Past to the Game Boy Advance. Meanwhile, Hidemaro Fujibayashi served as the director for Four Swords, which required at least two players to play. Playing the Four Swords quest would in turn unlock bonus material in A Link to the Past.

Four Swords allowed up to four players with Game Boy Advance devices and four copies of the game to play together. Each player would be represented by a Link of a different colour, and Link's overall design was modeled after the Toon Link concept created for The Wind Waker.

A Link to the Past & Four Swords was released in 2002 for the Game Boy Advance and sold 1.89 million units on the device. Nintendo of America used the game's re-release as an opportunity to redo A Link to the Past's English localization, unifying its terminology with that of future Zelda games.

Gamer Drift & Aonuma's Conundrum

The Wind Waker

By 2003, the Japanese videogame market had declined visibly, both in terms of hardware and software sales. This phenomenon was being referred to as "Gamer Drift"—where existing customers were losing interest in videogames and not enough new customers were being created to take their place. Gamer drift was something Nintendo in particular took very seriously. They had attributed low sales of The Wind Waker in part to the phenomenon, and the general sense was that the next Zelda game was going to face an uphill battle in Japan. While the North American videogame market was much healthier, The Wind Waker hadn't sold to expectations in the Americas either, owing to the game's divisive visual style. As a result, there was a sense of uncertainty about where The Legend of Zelda needed to go next, and what exactly the future would hold.

Meanwhile, director Eiji Aonuma faced a personal conundrum. Aonuma had served as one of many directors on Ocarina of Time and as the sole director of the two console Zelda games succeeding it. By the time development on The Wind Waker wrapped, Aonuma found himself exhausted by the experience. While on tour in Europe to promote the game, he also realized he was uncomfortable conducting interviews alongside his mentor, Shigeru Miyamoto. Miyamoto had invented the Zelda franchise and had strong opinions about just what constituted the virtues that made the games appealing. While Aonuma had directed the past few Zelda games, Miyamoto still had the final say in what was allowed and what wasn't, and his outlook on what qualified as "Zelda-ness" was often unclear to the director.

"Mr. Miyamoto points out every mistake that I made in front of the reporters!" Aonuma would jokingly reveal in an interview several years later. "For example, the most frequently asked question from the reporters is about that 'Zelda-ness (What makes a game a Zelda game)'. It’s a hard question to answer, even for us. Even Mr. Miyamoto is inconsistent with his answers. In one interview he answered, 'Zelda games are unique,' and then in another he suggested, 'Zelda games demonstrate growth'. I’m like, 'which one is it?' But in an interview, I must give an answer to every question. So I would talk about that 'Zelda-ness' just as Mr. Miyamoto would describe, only to be interrupted by Mr. Miyamoto himself disagreeing with me saying, 'No, that’s different,' in front of all the reporters!"

Fatigued by The Wind Waker's development and the subsequent promotional tour, Aonuma eventually informed Miyamoto that he wanted to step down as director of Zelda and work on something else instead. In turn, the latter requested that Aonuma stay and promoted him to producer, suggesting he take on a supervisory role and focus on making the series better from a distance. Inspired by the opportunity, Aonuma agreed and began brainstorming a variety of different The Legend of Zelda games—one of which would ultimately save the franchise from an untimely demise.

Four Swords Adventures

Four Swords Adventures

During the development of Four Swords and The Wind Waker, Nintendo had begun to experiment with the idea of connecting the Game Boy Advance to the Nintendo GameCube, similar to how the Nintendo 64 could import player data into Pokemon Stadium from the Game Boy. Connectivity between the Game Boy Advance and Nintendo GameCube was used in games like Pokemon Colosseum, and was also a feature Nintendo made use of in The Wind Waker, through an item called the Tingle Tuner. To use it, one would connect the Game Boy Advance to a Nintendo GameCube via a Link Cable and activate the item, which would allow the player to take control of Tingle. The Tuner would then enable a number of smaller features and abilities within the game, which would assist Link in his quest.

After development wrapped on The Wind Waker, director Eiji Aonuma was promoted to producer of the Zelda series. Aonuma had grown weary of directing Zelda games due to the heavy workload involved, and doing press interviews alongside his mentor, Shigeru Miyamoto, was adding to his stress. Following a press tour in Europe, Aonuma asked to be let off the Zelda team, but Miyamoto requested that he stay and take on a more supervisory role as producer instead, from where he could help improve the Zelda series as a whole. As the creator of the series, Miyamoto's word would still be ultimate but Aonuma would now have the opportunity to help steer the franchise going forward.

The first project would Aonuma take on as producer was a successor to Four Swords, the multiplayer Zelda game that had been included with A Link to the Past on the Game Boy Advance. Up until that point, the more recent portable Zelda games had been developed by Capcom, with development support from Nintendo's own Zelda team. Aonuma hadn't been involved with these titles, owing to his responsibilities on the home console Zelda games, but as producer he would now be in charge of every Zelda that was in development. At the behest of Miyamoto, Aonuma and his team began working on the game, tentatively titled Four Swords for Nintendo GameCube. Miyamoto felt such a game would promote Nintendo GameCube-to-Game Boy Advance connectivity possibly help capture the interest of Japanese consumers.

At this point, Aonuma was overseeing three separate Zelda games: Four Swords Adventures, The Minish Cap, and a third game tentatively titled The Wind Waker 2.

Four Swords Adventures differed from Four Swords in that it now required both a Nintendo GameCube and a Game Boy Advance to play. The game would be played from a top-down view and four players would be able to explore together on the TV screen. When players entered a dungeon, the view would shift to their individual Game Boy Advance screens. The team felt that maintaining a top-down view all throughout would make it easier for all four players to tell where they were.

With Aonuma promoted to producer, a new director needed to be appointed for Four Swords Adventures. Since Hidemaro Fujibayashi, the Capcom director behind the Oracle games, was working on The Minish Cap, the job went to Nintendo's Toshiaki Suzuki instead. Suzuki had previously directed Super Mario Advance for the Game Boy Advance and helped design the Tingle Tuner feature for The Wind Waker. He was a fan of the Zelda series and would go through a number of prior Zelda titles, handpicking elements from nearly every game to pay homage to in Four Swords Adventures, whether it was in the form of puzzles or other in-game elements.

Four Swords Adventures was originally scheduled for release in Japan in February 2004, but was delayed by a month to accommodate a polished single-player mode. The game's main campaign, Hyrulean Adventure, was initially meant to require at least two players the way Four Swords did, but the team changed plans partway through development to allow solo play as well. Two months prior to release, Shigeru Miyamoto advised Suzuki that the single-player campaign needed more polish, and so the team spent an extra month redoing the way the game played with a stronger focus on the single-player aspect.

Four Swords Adventures was eventually released in March 2004 in Japan and a few months later in the west. While the game was well-received at E3 the year prior, it ultimately sold poorly owing to the fact that every player needed their own Game Boy Advance and a Link Cable with which to connect it to the Nintendo GameCube. This, Nintendo would later admit, made it too difficult to convince customers that they needed to play the game.

West vs. East

The Minish Cap

Concept art by Capcom. The text reads: "We want this to be a little more ominous than A Link to the Past"

After Oracle of Ages and Oracle of Seasons, Nintendo had continued to collaborate with Capcom to create more The Legend of Zelda games. Capcom had initially begun planning a game for Nintendo's next portable platform, the Game Boy Advance, as far back as 2001 but this project had to be put on hold to allow the development of Four Swords, which would be included with the Game Boy Advance remake of A Link to the Past. Once development on Four Swords wrapped in 2002, Capcom's Zelda team returned to its other project, now under the series' newly-appointed producer, Eiji Aonuma.

Director Hidemaro Fujibayashi had two broad goals for the project: to do something that nobody had done before, and to make a game that was representative of Capcom's talent for 2D artwork. To give the game a unique gimmick and help showcase Capcom's artistry, Fujibayashi decided to explore the disparity between big and small, inspired by the Gnat Hat item from Four Swords, which allowed Link to shrink in size.

While Nintendo's own Zelda team was accustomed to creating in-game assets early on and experimenting with them to determine the look and feel of a game, Capcom chose to create a large number of concept drawings first to convey this new Zelda's unique theme.

Concept art for a barrel house

"Our first approach was to take our early image sketches and try to convert them to 2D graphics on the Game Boy Advance," Eiji Aonuma would recall. "One picture showed a tiny creature in a barrel, and a tiny Link. Another picture showed a corridor made out of the gaps between normal sized furniture, and an unknown world opening up beyond... stuff like that. Once we saw those images, we could feel 'this is going to be an interesting game'."

Playing with perspective

One of the development team's goals was to include elements that were more geared toward 3D, but create them using 2D artwork instead. This visual trickery was accomplished by distorting the environments and Link's sprite to depict scale and perspective. In some cases, the world around the player would become larger as Link would shrink down in size. Other times, the world would appear the same as he shrank and Link himself would appear as a tiny dot on the screen. During development, Fujibayashi would state that the team was aiming to create a game that could compete with the 3D Zeldas, and be considered the pinnacle of 2D gaming.

While Fujibayashi served as the game's director on the Capcom side, Eiji Aonuma took on the role of supervisor-cum-director at Nintendo. By early 2004, Aonuma was already involved with another Zelda project—a "realistic" Zelda for the GamCube that would eventually become Twilight Princess. However, since the Nintendo GameCube wasn't selling well, Nintendo's business was largely being sustained by the Game Boy Advance and their first priority was to keep sales of the device stable.

Aonuma would divert his attention from the console Zelda to see development of The Minish Cap through, ensuring that the Game Boy Advance had a marquee title for the holiday season that year. Working on The Minish Cap would also allow Aonuma a brief respite from the development of Twilight Princess, which was having a difficult time getting off the ground.

The Minish Cap marked the last Zelda game Nintendo would collaborate with Capcom on. Following its release, director Hidemaro Fujibayashi became a fulltime employee of Nintendo and eventually went on to help create one of the most lauded games in the series. The Minish Cap itself would only sell 1.76 million units, owing to the fact that the Game Boy Advance was in the process of being succeeded by the Nintendo DS.

Twilight Princess

Early concept art

The Wind Waker had not performed to expectations. The game had sold relatively poorly in Japan, owing to the fact that the country's videogame market had begun to decline. Meanwhile, despite a successful pre-order campaign, sales in the west were slowing faster than usual. Series producer Eiji Aonuma would discover that this was because The Wind Waker's cartoon-ish visuals had alienated the upper-teen audience that represented the typical Zelda player in North America—the series' largest market.

By the time The Wind Waker was released, game development costs on home consoles had risen significantly. This was something that Nintendo had been wary of for some time, as the company believed that creating games with constantly escalating budgets was an unsustainable business. Meanwhile, Four Swords Adventures, a concept pitched by Shigeru Miyamoto, had failed to set a new direction for either Zelda or the Nintendo GameCube, and Nintendo was uncertain as to what the Zelda franchise needed to do next. A declining Japanese market, rising development costs, and apathy from the western audience were putting pressure on Eiji Aonuma and the core Zelda development team to achieve some sort of breakthrough, failing which the franchise was under threat of being shelved permanently.

At the time, Nintendo was already planning the next Zelda game for the Nintendo GameCube. Tentatively titled The Wind Waker 2, the game would use the same cel-shaded visual style as The Wind Waker but would take place on land instead, with Link riding on horseback like in Ocarina of Time. During the initial stages of planning, the development team would discover that Toon Link's proportions didn't lend themselves well to horseback riding, and while an adult version of Toon Link had already been contemplated for the original The Wind Waker, the team felt this wasn't the solution they were looking for.

By this point, Eiji Aonuma, who had directed The Wind Waker and was now in charge of overseeing all Zelda games, had determined that three things were necessary for the next Zelda to sell well:

1. A cooler, more realistically proportioned Link

2. The ability to explore on horseback

3. A convincing world similar to those seen in fantasy movies such as Lord of the Rings

All three elements had been present in Ocarina of Time and Aonuma sensed that fans wanted Nintendo to build upon that style of game, rather than do something completely different. At the end of 2003, he discussed the matter with Shigeru Miyamoto, informing him that he wanted to make a realistic Zelda game, which would expand upon Ocarina of Time's appeal. Miyamoto was initially skeptical; his style of developing games called for a constant stream of new ideas rather than refinement of old ones—a trait that was commonly seen across Mario games. Eventually, though, he gave Aonuma permission to attempt a more realistic Zelda that would follow in the footsteps of Ocarina of Time, and advised he use the opportunity to accomplish things Ocarina of Time couldn't. The news gave The Wind Waker 2 project, on which progress had slowed, a much-needed jumpstart, and the team changed course to start planning a more realistic, mature Zelda game.

Initial concept design for Link

Nakano's finalized concept image

Work on the game began with Yusuke Nakano creating a more grown-up look for Link. The goal wasn't to make the game photorealistic, but to capture the same sense of atmosphere as in Ocarina of Time. A number of concept images were drafted, with the initial designs having Link appear more rugged than usual. These images were shared with Nintendo of America, who informed the development team that the Link players liked most was the adult Link from Ocarina of Time. Based on their feedback, Nakano reworked reworked his Link design, coming up with a new illustration that could be used to set the tone for how the entire game would look.

Aonuma and the team were planning for this new Zelda to be larger in scope than any game prior, and would need to collaborate with external artists for the extensive amount of artwork it would require. Nakano's concept image of Link was meant to help convey the look and feel of the game to everyone working on it and ensure a stylistic consistency across the board.

Alongside artwork, the team began developing a system for horseback combat—something they originally wanted to do in Ocarina of Time. Over the next four months, the developers created a prototype where Link could fight enemies while riding a horse in a realistic looking world. Using this prototype, Aonuma and Nintendo of America began editing a teaser trailer to show off at E3 2004. The trailer was designed so that the audience wouldn't realize they were looking at a Zelda game at first, until the footage would gradually reveal Link and the series' title to elicit excitement.

The team's plan worked. At E3 2004, the teaser for the new Zelda received a standing ovation from the audience in attendance. To Aonuma and his team, this reinforced that they were on the right track.

"When it was announced with a surprise trailer at the 2004 E3, it received a standing ovation from the media audience," Aonuma would recall during a post-mortem. "This was a very exciting moment for us, but we were still in the very early stages of converting the game into something more realistic. We knew that we had to create a Zelda game that would live up to the expectations of fans in North America, and that if we didn’t, it could mean the end of the franchise."

During these early stages of development, one of the concerns the development team had was that they weren't able to formulate new gameplay ideas. Similar to The Wind Waker, a number of the ideas in this new Zelda were shaping up to be similar to prior games. At the same time, the team didn't want to make too many radical changes for fear that it might alienate part of their audience once more—especially in Japan, where sales were already shrinking. One radical idea that was rejected early on was a first-person perspective, mirroring the development of Ocarina of Time. Aonuma felt that being able to see the player character was an important aspect of Zelda, and while the team had performed first-person trials, the idea was scrapped relatively quickly.

Once he returned from E3, Aonuma found himself having to divide his time between a number of different Zelda projects. There was The Wind Waker 2, which was being developed by his team, The Minish Cap which was being developed in collaboration with Capcom, and a third project that Aonuma had assembled a small team to prototype: a The Legend of Zelda game for the upcoming Nintendo DS handheld, scheduled for release later that year. Meanwhile, the Wind Waker 2 team was brainstorming ideas around how the more realistic visuals of their new Zelda might tie into gameplay. As 2005 approached, the staff suggested to Aonuma that they experiment with a return to the idea of Light and Dark worlds introduced in A Link to the Past, and how switching between environments might lead to interesting gameplay. This time around, the two environments would intersect and have an impact on one another—an idea that had been contemplated for A Link to the Past, but ultimately shelved. The dark world, dubbed the "Twilight Realm," would be designed to look like a highly technologically-advanced society, marking the first time that Nintendo would actually add sci-fi elements to a Zelda game, after having contemplated the idea twice before for the original The Legend of Zelda and A Link to the Past.

Early concepts for Midna, possibly from The Wind Waker 2

Concept for Midna on Wolf Link

Following a bizarre dream while he was travelling overseas, Aonuma asked the team to have Link turn into a wolf while in the Twilight Realm. This would allow the development team to give Link a completely different skillset while in wolf form, thus creating a different style of gameplay. At Miyamoto's behest, the team also created a character that would ride on Wolf Link's back, to make for a more visually interesting sight. Initially, this companion character was designed to look relatively unassuming, but as development progressed she began to play a central role in the game's story and her design evolved. Eventually, Link's new partner inherited the traits of a goblin/devilkin character that was originally being designed for The Wind Waker 2. Design notes called for the character to look like a mix between a monster and a child, and she was named Midna (from the word "midnight").

The development team initially planned for Link to be transformed into a wolf at the very start of the game. Aonuma liked the idea of disorienting the player at the very beginning and forcing them to learn their way around as a wolf, but the idea was vetoed by Miyamoto.

Once the basic concepts of the regular world, the Twilight Realm, and Wolf Link were in place, Aonuma left his team to flesh them out and diverted his attention to another project that required it: The Minish Cap, which was scheduled for release later that year on the Game Boy Advance. Nintendo was uncertain as to how the Nintendo DS would be received and sales of the Game Boy Advance proved to be stable, even in Japan's declining game market. It was imperative that the handheld be sustained with marquee software, and so Aonuma would immerse himself in seeing The Minish Cap through to completion. He would later admit that the game served as an escape from the stresses of working on the more realistic Zelda project.

When he returned, he found that development on the project—which had been named "Twilight Princess"—was struggling. While the team had come up with disparate gameplay ideas and events, the actual structure of the game hadn't been created yet. Additionally, while Wolf Link was fun to play as, there was nothing new or surprising about playing as human Link. To help course-correct, it was decided that Aonuma would need to take on a more hands-on role as the game's director, while Miyamoto would step in as producer.

• 

Response to a playable demo at E3 2005 was positive, but Aonuma felt the game lacked a uniqueness—especially in comparison to his Zelda project on the Nintendo DS, which now featured intuitive touch controls, setting it apart from any game prior. When they returned from E3, Miyamoto suggested to Aonuma that he consider making Twilight Princess compatible with Nintendo's upcoming console, the Wii, which featured a unique motion-sensing controller. Miyamoto felt the Wii Remote was well-suited to the Zelda's bow and arrow item, and while he was apprehensive about how this would impact the game's development schedule, Aonuma agreed that pointer and motion controls would help it stand out. A month later, it was decided that Twilight Princess would be released for both the Nintendo GameCube and the Wii, effectively doubling the amount of work the development team would be required to put in to a project that was already planned to be the most ambitious Zelda game to date. In August 2005, Nintendo announced that Twilight Princess would be delayed from its Fall 2005 release window into 2006. It was Nintendo president Satoru Iwata that had suggested the delay, asking that the development team make the game "120% Zelda".

Twilight Princess

In addition to incorporating motion controls, the delay was also intended to help Twilight Princess add both content and polish. This had an unintended side-effect: As development progressed, the project ballooned out of control, with Aonuma later recalling that even he had been unable to rein it in once it had grown to a certain size. For starters, the game's fields were larger than any Zelda game prior so players could experience the thrill of riding on horseback, and this meant that the team would need to populate the overworld with enough activities for players to partake in. Meanwhile, dungeons were designed to be more elaborate, exploratory, and less straightforward than The Wind Waker's linear, guided spaces. In particular, the development team wanted to include open-air dungeons that felt like they were connected to the outside world, which added to the team's workload when designing the Forest Temple and the City in the Sky. In certain areas parts of the overworld, the team even experimented with creating dungeon-like gameplay, which would eventually result in Link's quest to collect Tears of Light across Hyrule. Additionally, key locations in the game were lined with a host of activities and minigames to participate in. Twilight Princess's introductory sequence alone—set in Link's home town of Ordon Village—featured so many different ideas that Miyamoto asked the team to design a lengthy tutorial spanning three in-game days of performing village chores.

Attention was also lavished on the design and personality of Link himself. The Link of Twilight Princess was modeled to be 16 years old and Aonuma liked the idea of him being a strong character with a slightly rebellious personality. To convey this, the development team gave Link a hidden move where he would sheathe his sword stylishly after defeating an enemy—inspired by the likes of George Chakiris' character in the 1961 musical West Side Story. Link also began the game as a farmhand, so it made sense that he would be capable of feats of great physical strength. This was conveyed in the way that Link would force dungeon doors open with his hands, as opposed to them opening automatically when the player pressed the A button. Partway through the development process, Nintendo of America would note that Link's voice, provided by his voice actor from Ocarina of Time, felt out of place. Subsequently, the development team would sample a number of different voices and eventually hire Akira Sasanuma, the voice actor for Dearka Elsman in Gundam SEED, for the role. Aonuma felt Sasanuma's mischievous tone was well-suited to the new Link, almost making him sound a little bit like a bad guy.

A great deal of attention was also given to Link's horse. While Nintendo hadn't been able to hire a horse for reference during the development of Ocarina of Time, character designer Keisuke Nishimori did have the opportunity to go horse-riding for reference during Twilight Princess. Nishimori's key takeaway from the experience was how large a horse actually felt next to a human being, and this helped bring a greater sense of realism to the horses in the game.

Nintendo had never worked on a project of this scope prior, and the development team eventually buckled under the pressure. Part of the problem was that there was a large number of designers working on the project, and each had their own idea of what constituted a Zelda game, which would often lead to heated debates among the team. This would ultimately lead to decisions being made collectively, rather than by a single authority, which would in turn mean that nobody was taking sole responsibility for various parts of the game. Team leaders would lose track of their teams and members of staff would be unable to properly execute their tasks. To solve these challenges, Aonuma would help bring different aspects of the game together while Miyamoto would provide the team with instructions pertaining to polish and the sense of interactivity and realism that he felt was important for a Zelda game. Features such as being able to toss cut planks of wood into a river and watch them float were added at Miyamoto's behest. Such features would often involve lengthy stretches of development time, with Miyamoto later revealing that the wooden plank feature took nearly as long to develop as Link's horse.

Link fishing

The game's motion controls on Wii proved challenging to implement as well. The development team demoed Twilight Princess at E3 2006, just five months before its release, and realized that it would need to make a number of adjustments to how the Wii Remote had been implemented. Notably, players that had the opportunity to play the game at E3 feedbacked that they wanted to be able to use the Wii Remote like a sword. Aonuma and his team had already experimented with this idea early on, but had chosen not to pursue it as they felt it would grow tiresome for the player. Following E3, the feature was polished and implemented once more, alongside motion controls for the game's fishing minigame. Program director Kazuaki Morita, who had designed the fishing minigame in Link's Awakening and Ocarina of Time did the same for Twilight Princess, programming the game's Wii Remote controls and making it so the weather would affect the player's ability to fish. The team would continue to make adjustments to Twilight Princess's motion controls and other elements right up until release.

Twilight Princess was, at the time, the largest game Nintendo had ever worked on, and the company's very first "AAA" project in scope and team size. Because of its scope and the addition of motion controls, Aonuma would later admit that the development team hadn't been able to incorporate all the ideas they had into the game, leading to players feeling that its overworld could feel empty—the same criticism that The Wind Waker had faced.

Regardless of its shortcomings, Twilight Princess did what it was meant to. It served as an important piece of launch software for the company's Wii console, with three copies of the game being sold for every four Wii consoles in its first week. With its atmospheric visual style, sense of scale, and engaging story, the game successfully revived The Legend of Zelda brand in North America and Europe, going on to sell over 8.85 million units worldwide—the highest sales in the series until 2017's Breath of the Wild. Twilight Princess would represent the Zelda brand for the next several years, with the game's version of Link being the series' main representative in promotional material as well as in Super Smash Bros.—even after the release of 2011's Skyward Sword. Twilight Princess would also give Nintendo a concrete idea of what players wanted from The Legend of Zelda and eventually influence the development of Breath of the Wild, with Eiji Aonuma stating that the latter wouldn't have been possible without Twilight Princess serving as a foundation.

Monolith Soft Joins Nintendo

In 2003, after Satoru Iwata had taken over as Nintendo's president, the company purchased a 2.6% share in Japanese toy maker Bandai, prompting rumors of a buy-out. Fuel was later added to the fire by a member of Bandai's founding family who claimed that there had been a conspiracy within the company's board to facilitate such a takeover by Nintendo.

Nintendo and Bandai both denied the rumor, with Iwata reiterating that Nintendo had no interest in taking over Bandai the following year.

While a takeover by Nintendo never occurred, Bandai would go on to acquire Namco, the videogame publisher behind games such as Pac-Man, Ridge Racer, and Tekken, in 2005. Namco also owned Monolith Soft, the development studio that was producing its Xenosaga trilogy of role-playing games. The studio had originated at Squaresoft in 1996, following which the team had gone independent and found a buyer in Namco, which had agreed to fund its games. Unfortunately, development on Xenosaga had run into complications and the games weren't selling to expectations. Following the merger, it became evident that Namco Bandai, the newly formed entity, wasn't entirely keen on investing in the kinds of games Monolith Soft liked to create. When Monolith Soft was approached by Nintendo, who offered to acquire the company and allow them to continue creating unique and original games, they agreed.

Concept art for Earthbound on the Nintendo GameCube

Nintendo had already expressed an interest in Monolith Soft's talent prior to Namco's acquisition by Bandai. In 2003, Nintendo president Satoru Iwata had pitched a new Earthbound game to series creator Shigesato Itoi for the Nintendo GameCube. Iwata had borrowed the talents of Monolith Soft co-founder, Yasuyuki Honne, who had served as art director on Chrono Trigger and Xenogears at Squaresoft, for the pitch. While the project never saw the light of day, Nintendo would strike a deal with Namco to have Monolith Soft develop other games for its platforms, including Baten Kaitos, Soma Bringer, and Disaster: Day of Crisis. It was during the development of Soma Bringer and Disaster that Monolith Soft was given the choice to change hands.

On April 27th, 2007, Namco Bandai and Nintendo announced that Nintendo had acquired Monolith Soft. Namco, which owned Monolith Soft, held 96% of the company's stock, and sold 80% of its shares to Nintendo, making the studio a first-party developer. Over the next few years, Monolith Soft would go on to develop the critically-acclaimed Xenoblade Chronicles and Xenoblade Chronicles X, amassing an expertise in creating large-scale open-world games and stating that they wanted to grow into the equivalent of Bethesda Softworks. The company would also open a second studio in Kyoto, closer to Nintendo's own headquarters, and begin contributing to the publisher's first-party projects, including future The Legend of Zelda games. Following the release of 2015's Xenoblade Chronicles X, Monolith Soft's main production team would take an active role in the development of Breath of the Wild.

Phantom Hourglass

The Nintendo DS, launched in 2004

Phantom Hourglass first began development in 2004. At the time, the brand was in decline, owing to poor sales of the last several Zelda games. The last console game, The Wind Waker, had alienated Zelda's North American audience with its cartoon-like visuals, while sales in Japan had been weak owing to the ongoing decline of the Japanese videogame market. Meanwhile, Four Swords Adventures on the Game Boy Advance had sold poorly due to requiring that players own both a Nintendo GameCube and Game Boy Advance, and later that year, The Minish Cap would also see poor sales, having been released at the very end of the Game Boy Advance's life.

Nintendo as a whole had been growing less relevant as well. Between poor sales of the Nintendo 64 and Nintendo GameCube, the declining Japanese market, and the fact that it had lost most of its third-party development partners to PlayStation, the company felt it had hit a dead end. To work around this problem, Nintendo had been experimenting with hardware innovations that could support entirely new kinds of games. The company's upcoming Nintendo DS platform had been designed with this goal in mind, featuring two screens and a touch screen for easy input. This enabled Nintendo's developers to create unique experiences for the Nintendo DS, such as Brain Age, a puzzle-solving game designed to keep the brain active.

While Nintendo worked out how to stay relevant in an increasingly competitive games industry, producer Eiji Aonuma was doing the same with The Legend of Zelda. Aonuma understood that Zelda's largest, most dedicated audience was in North America, where the videogame market was still healthy, unlike Japan. The two markets represented different challenges, and would need to be tackled separately. To help the brand back on its feet in the west, Aonuma began work on Twilight Princess, a more mature Zelda that was meant to serve as a successor to Ocarina of Time. Aonuma had pinpointed exactly what the western audience's expectations of Zelda were, and Twilight Princess was meant to deliver on those expectations.

Japan, however, was proving trickier to find solutions for. Nintendo had determined that Japanese consumers were losing interest in videogames as a whole, and felt that this was largely because games had grown too complicated and difficult to play. Solving this problem, Aonuma felt, could be the key to Zelda's troubles as well. At a GDC post-mortem, he would state: "Working on Majora’s Mask for the Nintendo 64 and then The Wind Waker for the Nintendo GameCube in succession, I began to worry that, due perhaps to the growing number of buttons necessary to control the game, or the 3D environment, new players might find these games intimidating and avoid them. I felt that there were sure to be many players who thought ‘Zelda looks fun, but there’s no way I can play it’, and give up before even giving it a try. For that reason, ever since then I have been thinking of ways to square this circle: how to make the controls easier without losing any of the unique fun-factor of a Zelda title."

A prototype for Phantom Hourglass

Around May 2004, after Four Swords Adventures had wrapped, the game's developers began experimenting with the Nintendo DS. Daiki Iwamoto, a programmer that had worked on Four Swords Adventures, had begun to prototype a multiplayer Zelda along the same lines for the Nintendo DS, when Aonuma stepped in and advised that the team devise a new style of Zelda gameplay using the Nintendo DS's touch controls instead. At the time, Aonuma was dividing his time between Twilight Princess and The Minish Cap, and left the five-man Nintendo DS team to continue with its experiments. When he returned from E3 the following month, the team informed him that the Nintendo DS hardware was capable of supporting cel-shaded graphics like the ones used in The Wind Waker. Aonuma had been disappointed by The Wind Waker's poor sales and asked the team to implement cel-shading in their prototype, hoping to give the style a second chance at success. Soon afterwards, the team presented Aonuma with a tech demo of Toon Link moving around in a 3D environment on the Nintendo DS's top screen. On the bottom screen was a map with an icon that the player would drag to move Link around. Feeling that these controls weren't intuitive, Aonuma asked the team to flip the two screens, moving the game world to the bottom screen instead, which would allow the player to control Link by touching him directly.

Early concept art for Link's boat

Partway through the project, Hidemaro Fujibayashi—who directed The Minish Cap and both Oracle of Seasons and Oracle of Ages—joined the team. Following The Minish Cap, Fujibayashi left Capcom to join Nintendo and was now part of the company's internal Zelda development team. While Daiki Iwamoto was directing the Nintendo DS Zelda, Fujibayashi was appointed sub-director and began working on the game's story. The initial goal was to develop a sequel to The Wind Waker that was smaller in scope and could be completed quickly, but as development progressed, the team found itself adding more and more content to the game, particularly to the ocean. Like The Wind Waker, the Nintendo DS Zelda took place on the open sea, and the development team used lessons learnt from the development of that game. The team put thought into how large the ocean would be, the size of individual islands, the distance between them, and the speed at which players would be able to sail. The developers also made it a point to include more sidequests, customization, and things to do in the overworld than in The Wind Waker, as they felt that game had been relatively lean on content. In parallel, they continued to refine the game's touch controls. Initially, the game used a combination of touch and button controls, but after the team had designed the boomerang item—which players would control by drawing its path on the touch screen—Aonuma asked that the game use touch input exclusively.

File:Phantom Hourglass Title Screen.png

The game's title screen

Taking notes on the Nintendo DS

At GDC 2006, Nintendo president Satoru Iwata revealed the Nintendo DS Zelda to the public under its final title: Phantom Hourglass. The name came from an in-game item—the Phantom Hourglass—which served as a timer while Link was in dungeons, and was chosen because Nintendo of America liked the word "hourglass".

Dungeons in Phantom Hourglass were designed such that each room served as a miniature puzzle in itself, often requiring the player to interact with several different items, switches, and enemies to solve. The player would be allowed to take notes on the Nintendo DS touch screen for reference, and this would aid in other parts of the game as well, including secrets hidden in the overworld. To challenge players, the team imposed a time limit on the game's main dungeon, the Temple of the Ocean King, requiring that players revisit the dungeon six times throughout the course of the game, and race against the clock to complete it as quickly as possible, using the notes they had taken during prior runs—as though they were collaborating with a past version of themselves. This, they felt, was a new twist that would throw players off balance.

Aonuma wanted Phantom Hourglass to play differently from any Zelda prior, and to appeal to a broad range of players, including children and adults. Once development wrapped on Twilight Princess, he began focusing his attention on the game, instructing the team in how to polish its pacing, environments, and controls, resulting in its release date being delayed from 2006 to the following year. Special care was also taken to make the game accessible to Japanese children by allowing the player to tap a kanji character to view its furigana reading. Aonuma's personal goal was to create a game that could compete with Brain Age in terms of simplicity and wide appeal.

"At first, we had the idea of creating a good game in a short time," he would recall in an interview. "We thought Brain Age was our rival. Brain Age’s like that smart transfer student. The Zelda Team’s not in the top places, but it studies hard. And then comes this transfer student and easily gets the first place without studying. That’s very frustrating. After three long years, we finally finished Twilight Princess and the transfer student’s the one that’s smart and cool and gets the first place? Damn it!"

The team also wanted to continue experimenting with multiplayer gameplay in Zelda and included a Battle Mode feature that could be played with two Nintendo DS systems connected wirelessly to one another. Aonuma would reveal that the idea of Link not being able to see his opponents until they made a move was something the team had "been sitting on for a number of years".

After its release in 2007, Phantom Hourglass went on to sell over 900,000 units in Japan—the highest of any Zelda game since Ocarina of Time. Following poor sales of both The Minish Cap and Twilight Princess in Japan's declining market, Phantom Hourglass successfully appealed to a new audience on the Nintendo DS through its visual style and intuitive touch controls. Notably, a number of women and children purchased the game. Globally, Phantom Hourglass would sell over 4.76 million units, beating out sales of The Wind Waker and becoming the highest-selling Zelda game on a portable platform for the next several years.

Spirit Tracks

Development on Spirit Tracks began in 2007, shortly following the release of Phantom Hourglass. Series producer Eiji Aonuma asked the development team to begin working on a sequel to the game re-using the same basic mechanics, similar to how Majora's Mask followed Ocarina of Time. For this second game, Aonuma wanted to do away with the concept of sea exploration and return to a land-based game, where players would discover new places while traversing across a vast landscape.

As a child, Aonuma's son enjoyed a picture book titled The Tracks Go On, which was about a group of children laying down train tracks to explore the world. Aonuma would often read the book to his son as a bedtime story, and liked the idea of laying tracks across Hyrule in a Zelda game, feeling that it was a good fit for the series' pioneering spirit.

• 

  A world map drawn early in development

• 

  The railway network, with the Tower of Spirits at its center

After Aonuma suggested the idea to the development team—led once again by Phantom Hourglass director Daiki Iwamoto—experiments began to see just how much freedom they could afford the player in laying down their own tracks. One of the problems the team encountered early on was that players wouldn't know where to lay the tracks and allowing too much freedom would let players go places they weren't supposed to visit before the story allowed it. These experiments carried on for an entire year. The solution the team eventually settled on was to make it so this version of Hyrule had tracks to begin with, but they had been erased by a curse and the player would be required to put them back the way they were. There was debate over taking agency away from the player, but the team ultimately felt that this system offered an acceptable middle-ground between freedom and player guidance.

Ideas other than trains were considered

Zelda possessing a Phantom

There was also initially debate about whether or not a train fit the image of The Legend of Zelda series. The lead designers on the team were Koji Takahashi (who would go on to direct the art of Animal Crossing: New Leaf) and Seita Inoue (who would later direct the art of Splatoon). Inoue was originally in charge of designing the game's menus but ended up involved in the designs of the trains and areas around the game's main dungeons as well. After consulting the designers and other parties, the team ultimately decided to proceed with the train idea, not wanting to be too bogged down by what felt "Zelda-like" and what didn't. Like in Phantom Hourglass, players would be allowed to collect parts to customize their trains, with the team making it a point to add interesting variations and themes for players to experiment with.

The other core feature of this Zelda was the idea of controlling two characters instead of one. In Phantom Hourglass, players had been able to switch to playing as a Goron, and the team had been experimenting with the idea of a sub-character along those lines for a while.

Iwamoto liked the idea of Link being accompanied by Princess Zelda for a change, and this led to the idea of Zelda being able to possess and control Phantoms. To set her apart from prior versions of Zelda, she was written to be more comedic and act more like a girl her age than a member of royalty.

Spirit Tracks received its English title before the development team had settled on a name in Japanese. After Nintendo of America had titled the North American version, the team initially tried variations around the word "soul" (inspired by "Spirit"), but ultimately decided it felt too heavy and haunting. After soliciting suggestions from the development staff, the team settled upon "Train Whistle of the Wide World" as the game's Japanese subtitle.

Spirit Tracks was in development for two years. When it was released in Japan, Nintendo and retailers overestimated demand for the game, resulting in too many copies being shipped to stores. Spirit Tracks sold through just 47% of its initial shipment at launch, with stores quickly slashing the game's prices in order to move more copies. It would go on to sell close to 700,000 units in Japan and 2.96 million units worldwide.

Misreading the Market

A Fear of Getting Lost

The decline of the Japanese videogame market

By 2007, the Japanese videogame market's decline had sped up significantly, with home console sales at an all time low. Sales of the PlayStation 3 in Japan were extremely poor in comparison to the PlayStation 2, and while the Wii was doing better it was due to casual software like Wii Sports, which was selling to a "blue ocean" audience that wasn't interested in traditional videogames. The videogame industry was largely being supported by portable platforms such as the Nintendo DS, which were better suited the Japanese lifestyle. In 2007 the Nintendo DS accounted for over 50% of hardware sales and 40% of software sales across Japan.

Mirroring Japan's shift to portable gaming, Phantom Hourglass had been a moderate success, attracting a new audience to the series with its intuitive touch controls and welcoming visual style. In comparison, Twilight Princess had not sold quite as well. While the game was on its way to becoming the best-selling Zelda title globally, Japanese sales had been underwhelming, owing to low interest in home console games and the fact that the Japanese lifestyle largely centered around portable devices by this point.

Nintendo had seen the shift coming. In 2004, prior to the launch of the Nintendo DS, Nintendo's president, Satoru Iwata, stated: "Games have come to a dead end. Creating complicated games with advanced graphics used to be the golden principle that led to success, but it is no longer working. The biggest problem is that [developers] need to satisfy the core gamers, who want games with more volume and complexity, while they also need to satisfy average users, who don't have as much knowledge about games. The situation right now is that even if the developers work a hundred times harder, they can forget about selling a hundred times more units, since it's difficult for them to even reach the status quo. It's obvious that there's no future to gaming if we continue to run on this principle that wastes time and energy [in development]."

Nintendo had created the Nintendo DS and Wii in response to a business model they felt was unsustainable. Both devices were built on relatively low-spec hardware with unique control mechanisms at the forefront instead. The company felt they would help reset the status quo and enable developers to reach broader audiences while spending less money on development.

A few months after Twilight Princess's release, Nintendo appeared to have doubled down on this line of thinking—at least as far as the Japanese market was concerned. Shigeru Miyamoto stated: "I think a lot of people who bought the Wii are not necessarily the types of people who are interested in playing that kind of game. And a lot of the people who would want to play it [due to chronic shortages of the console] can’t find a Wii! But mostly, I think it’s that there are fewer and fewer people who are interested in playing a big role-playing game like Zelda [in Japan]."

Nintendo's blue ocean audience on the Wii exhibited very different habits from the enthusiast gamers the company had catered to for so many years. In particular, they were less predictable than enthusiasts and it could be difficult to pinpoint just what would click with this audience and what wouldn't. While the blue ocean represented a real opportunity to create new kinds of games, Nintendo staples like Mario and Zelda weren't always able to appeal to them. Unless the company found a solution to this problem, its new audience would never connect with Nintendo's full breadth of properties.

The answer, Nintendo felt, lay in simplicity. The blue ocean audience appreciated simplicity and straightforwardness, and did not appreciate complex, exploratory games. The company pointed to past Mario titles in support of its theory. While 2D Mario games such as Super Mario World had often sold to an extremely broad audience, 3D Mario titles such as Super Mario 64 and Super Mario Sunshine never managed to reach those same heights, particularly in Japan. This, Nintendo felt, was because the 2D Marios were simpler and more straightforward, whereas the 3D Marios turned casual gamers off because they found the idea of getting lost in a 3D environment too intimidating.

Super Mario 3D Land

The company designed Super Mario Galaxy and Super Mario Galaxy 2 around this theory. Super Mario Galaxy's spherical planetoids were designed so players would always return to where they had started, making it easier for them to keep track of where they were. Meanwhile, Super Mario Galaxy 2 would feature more 2D elements and camera angles that gave the impression of playing a 2D Mario game, even though the game itself was in 3D. Additionally, Nintendo would include a DVD with the Japanese version of Super Mario Galaxy 2, titled "Super Mario Galaxy 2 for Beginners," that was meant to ease new players into the game.

This belief that consumers in Japan were turned off by the prospect of getting lost would ultimately lead to the development of Super Mario 3D Land on the Nintendo 3DS. Super Mario 3D Land was designed to be the "missing link"' between 2D and 3D Mario games. It was designed to play like a fusion between the two schools of design, with 3D environments but a fixed camera angle and course design that steered the player towards the end. This development philosophy would eventually spill over into The Legend of Zelda, resulting in what is generally considered one of the weakest games in the franchise and cause Nintendo to thoroughly re-examine its priorities while developing future Zelda titles.

Skyward Sword

Skyward Sword began development in 2007, when Nintendo's Hidemaro Fujibayashi presented the development team with a planning document for a new Zelda game on the Wii. Fujibayashi had previously directed The Legend of Zelda games developed by Capcom, then served as sub-director on Phantom Hourglass. Series producer Eiji Aonuma appointed Fujibayashi to direct the new Wii Zelda and discussions began around the idea of the game using the Wii MotionPlus, a peripheral that could enhance the Wii console's motion-sensing capabilities to make them more accurate.

The team initially spent half a year experimenting with the Wii MotionPlus, hoping to learn the intricacies of the device and how it could be used to enhance sword combat in a Zelda game. Twilight Princess had allowed for motion-controlled swordfighting, but due to the limitations of the original Wii Remote the player didn't have full control over the swing and direction of their sword. This time, the developers wanted to use the MotionPlus accessory to allow for more subtle control over Link's sword. As a variety of challenges arose, the staff found itself struggling to tame the MotionPlus, and experiments continued for another year-and-a-half.

The primary challenge was that the MotionPlus was extremely sensitive, and designing sword combat to look convincing was proving difficult. One of the challenges the team faced was how to make Link look both realistic and charismatic while swinging his sword around, and studied the human skeletal structure for reference. After an extensive period of trial-and-error using Wii Sports Resort—which used the MotionPlus for a rudimentary swordfighting minigame—the development team eventually managed to design a swordfighting system it felt could serve as the foundation for its new game.

Assigning sword swings to motion controls freed the A button on the Wii Remote up for other functions, and the team assigned a new "dash" feature to the button. In addition to allowing Link to run, the dash would serve as a context-sensitive action capable of letting him vault over enemies or run up walls, rather than bumping into them. This led to the creation of a stamina meter that would prevent Link from being able to dash infinitely.

Twilight Princess's radial item menu

Skyward Sword's gesture-based radial menu

Once the team was comfortable with how the MotionPlus accessory worked, Fujibayashi brainstormed ways to use its motion-sensing capabilities for other actions, such as item selection. Fujibayashi wanted players to be able to select items by making gestures without having to look at the screen. To accommodate this, a radial menu similar to the one in Twilight Princess was designed and key items like the bow and bombs placed at angles that would be easy to remember. While intuitive, this mechanism for selecting items would lead to an unattractive user-interface—a trait that would extend across the entirety of Skyward Sword's UI, owing to the fact that its motion controls were new and required visual cues explaining how they worked.

By 2007, the year Skyward Sword began development, the Japanese videogame market's decline was speeding up. Twilight Princess sold well worldwide, breaking the series' global sales record, but sales in Japan were weak. At the time, Nintendo's Shigeru Miyamoto had stated: "I think a lot of people who bought the Wii are not necessarily the types of people who are interested in playing that kind of game. And a lot of the people who would want to play it [due to chronic shortages of the console] can’t find a Wii! But mostly, I think it’s that there are fewer and fewer people who are interested in playing a big role-playing game like Zelda [in Japan]."

While the Wii had popularized simpler games like Wii Sports and Wii Play in Japan, Nintendo felt that the current Japanese market was intimidated by more complex games, especially those involving exploration and sophisticated 3D environments. The company felt that this was most evident in the sales of its Mario games. While 2D Mario titles such as Super Mario World would often sell well in Japan, 3D Mario games such as Super Mario 64 and Super Mario Sunshine wouldn't perform nearly as well. This, Nintendo believed, was because the 2D Marios were simpler and more straightforward, whereas the 3D Marios turned casual gamers off because they found the idea of getting lost in a 3D environment intimidating. In response, the company developed Super Mario Galaxy and later Super Mario Galaxy 2, both designed to prevent players from getting lost and appeal to Nintendo's new audience of casual gamers on the Wii.

The idea of games growing too complicated was something the Zelda team had been wary of as well. Nintendo had felt for years that creating big-budget and increasingly complex games was unsustainable, and Shigeru Miyamoto would routinely encourage producer Eiji Aonuma and his team to create more compact, experimental games such as Majora's Mask and the portable Zelda titles instead. Miyamoto was skeptical of developing more games like Ocarina of Time and Twilight Princess owing to the vast scope of those projects, and after Twilight Princess's exacting development cycle, Nintendo would go to great lengths to make Skyward Sword a more compact and less exploratory game. The development team would design Skyward Sword to contain fewer and smaller environments than Twilight Princess, hoping to entice players with a breadth of activities instead of a vast, connected world to explore. This would ultimately result in the team opting not to create an overworld for the game at all, inspired by the simple manner in which Mario games let users choose levels from a menu.

In an Iwata Asks interview, the following exchange between producer Eiji Aonuma and late Nintendo president Satoru Iwata would explain the team's decision:

Diving onto the Surface

Aonuma: Usually, when we make a The Legend of Zelda game with a continuous body of land, we need an overlapping part to join one game field to the next. This time, we made all kinds of gameplay for the forest, volcano and desert areas, and needed to create roads for going back and forth among those places. Every time, it was quite a struggle to figure out how to handle those roads.

Iwata: Roads are particularly essential to a game like The Legend of Zelda.

Aonuma: That’s right. But the first thing we thought of this time was that perhaps we didn’t need those roads.

Iwata: What do you mean?

Aonuma: Well, [director] Fujibayashi-san and I talked for a long time about how, if we could make the gameplay in each area dense, then we wouldn’t need to physically join them. Then the question was “How do we design it?"

Iwata: And what did you think of?

Aonuma: Course selection in Super Mario games.

Iwata: Course selection?

Aonuma: Yes. In Super Mario games, there’s a course selection screen, and you waltz on over to it and hop in.

Given that course selection in Mario games was a straightforward process, Nintendo felt that the same design could be applied to Zelda and would achieve similar results, neglecting to consider that the appeal of Zelda had always been exploring an expansive world. The end result was a skydiving system by which Link would drop down into each area from the sky, rather than traversing an explorable overworld. Aonuma and his team had already experimented with blurring the line between dungeons and the overworld in Twilight Princess, and decided to take that idea even further in Skyward Sword. They designed each area so that it felt like a dungeon in itself, doing away with the idea of an overworld entirely to prevent players from getting lost, and to save on development time. In order to further prevent the player from losing their way in these areas, the team conceptualized the Dowsing feature. Fujibayashi would later reveal that Dowsing was created so that the team wouldn't need to create landmarks or other environmental cues to guide the player along. Nintendo would promote this aspect of the game in an Iwata Asks chapter titled "Making the 'Not' Lost Woods".

Aonuma would reveal: "I think we were able to keep such a big project together because the game world this time is structurally simple. We talk about all these 'dense' places, but structure-wise there are only four—forest, volcano, desert and sky."

Zelda designed to be more expressive

Technology makes a return

Among the other concessions made for Japanese players, Skyward Sword was also modeled after Japanese tastes in how its hub town was designed. Link's time at Skyloft, his home town, would see him and his peers attending high school together, a popular setting for Japanese games. The plot would start out akin to a high school drama, with Zelda and Link being childhood friends—a decision that was meant to endear players to Zelda. To make it easier to design events, the development team used a special tool that allowed designers to implement character dialogue and events themselves, rather than the programming team having to do it for them. Owing to the fact that the rest of the team was busy with its responsibilities, Aonuma used this tool and personally tweaked the game's opening sections and dialogue in Skyloft himself. Daiki Iwamoto, the director of Phantom Hourglass and Spirit Tracks was in charge of the sky area surrounding Skyloft, and Aonuma would volunteer to design the game's opening to help ease his workload.

Meanwhile, Xenoblade Chronicles developer Monolith Soft would contribute to the game's field layout designs, conceptualize sub-events, and even write pieces of in-game text, with Disaster: Day of Crisis director Keiichi Ono leading the Monolith Soft team.

As development progressed, Aonuma, Fujibayashi and their team designed Skyward Sword such that each of the three main areas in the game would involve a gimmick of some sort. For the Lanayru Desert, the team considered the notion of time-traveling in realtime, inspired by the contrast between the past and present in Ocarina of Time, and the normal realm and Twilight Realm in Twilight Princess. This resulted in the development of the game's Timeshift Stone item, which would allow Link to turn the present into the past, impacting the environment and enemies around him. The Timeshift Stone, along with the Hook Beetle item, resulted in the creation of the game's ancient civilization, consisting of clay robots powered by electricity. Like in Twilight Princess, the team would once again experiment with introducing technological themes to The Legend of Zelda's world.

Fi, Link's new partner

The idea to make Skyward Sword the first chronological story in the Zelda series came from the MotionPlus accessory and focus on swordplay. In an Iwata Asks roundtable, Aonuma would reveal: "This time, the theme is the sword which makes use of the Wii MotionPlus accessory. When you think of a sword in The Legend of Zelda, you think of the Master Sword. Rather early on, we decided to address the origin of the Master Sword. About that time, we began talking about how that would make this the first story in the series, and we wondered about involving the birth of Hyrule Kingdom. On the other hand, there was the setting of the floating island in the sky, and we thought, 'How did that get there?' We settled on having the sky and surface world, and on top of that, it was going to tell the story of the creation of Hyrule, with the untold story of the origin of the Master Sword.

The story setup would lead to the creation of Fi, Skyward Sword's partner character. The design for Fi and who would serve as lead concept artist on Skyward Sword were both decided through an internal competition within the development team. Takumi Wada was chosen to be the game's artist, while Yusuke Nakano—who was now mentoring younger artists at the company—provided the designs for Fi. Similar to Navi and Midna, Fi was meant to provide the player with hints wherever the team felt they might get stuck. Concerned about their casual audience, the development team would inadvertently make Fi overbearing and intrusive, with the character routinely interrupting the player with hints and tutorial text. This would be one of the most common criticisms of Skyward Sword upon its release.

At one point, the development team considered making Zelda a playable character in the game. The idea was to allow the player to play through a "Second Quest" from Zelda's perspective after she had landed on the Surface world. This idea was ultimately discarded, but the setting of that story was used in Skyward Sword's ending cinematic.

When it was released, Skyward Sword was backed by one of the largest promotional campaigns Nintendo had ever conducted. The game was used to celebrate The Legend of Zelda's 25th anniversary, with the "Symphony of the Goddesses" concert series leading up to—and following—its release. In conjunction with the game, Nintendo and publishers Shogakukan and Dark Horse released Hyrule Historia, an encyclopedia that would solidify the series' long-debated timeline and provide development insights into each game. Commercials starring Robin Williams and his daughter—who he had named after the series—were broadcast around the world to promote Skyward Sword as well.

Skyward Sword received positive reviews from critics upon release in 2011, but the game has since gone on to be less fondly remembered. Despite its many concessions for less experienced gamers in Japan, Skyward Sword would sell just 352,000 units in the country and a total of just 3.67 million units globally, making it one of the lowest-selling Zelda games on a home console. This was partly because the Wii was nearing the end of its tenure by the time Skyward Sword was ready for release, but also because the game stirred controversy among The Legend of Zelda's audience in the west through its linearity, excessive tutorials, and lack of an overworld. Skyward Sword was a game that didn't cater to any single audience particularly well—veteran gamers in the west or more casual gamers in Japan. Negative feedback for the game would make the development team re-examine the Zelda series' identity once more, as they had done following the release of The Wind Waker. Producer Eiji Aonuma would later confirm that criticism of Skyward Sword's linear structure and lack of exploration had directly resulted in the development of Breath of the Wild.

Rethinking the Conventions of Zelda

The Need for Change

Ocarina of Time

Ever since they began creating software for the Nintendo 64, Nintendo had grown wary of the time and money required to develop videogames. It had taken the company three years to develop Ocarina of Time, with producer Shigeru Miyamoto stating after the game's release: "I feel there is a bad atmosphere that you can't do something new at Nintendo these days. I never thought things like this before. So now we are changing ourselves to an organization that allows people to do new things and energize ourselves. I'm saying to my people that from now on let's go for the game that can be developed within six months and sell a million copies. If you want to finish a game within six months, you have to make it within two months because you need to polish it for another four months. If someone asks me who can make such a thing, I'd tell them that I used to do it. It isn't a great thing to take three years. [Ocarina of Time] would have been finished in a much shorter period if we had cut some parts."

Through his years developing games at Nintendo, Miyamoto had grown into an extremely efficient product manager and was running the entirety of Nintendo's Entertainment Analysis & Development (EAD) department, responsible for most of the company's output. Miyamoto's outlook was that the best games were a) unique, and b) developed within moderate budgets and timeframes. He disliked the games industry's tendency towards ever-increasing budgets and game sizes, and wanted Nintendo to create software that wouldn't compete with other developers' games, but sidestep them entirely.

It was an outlook Miyamoto shared with Nintendo's president at the time, Hiroshi Yamauchi. Under their leadership, most of the company operated under a similar philosophy. Nintendo prided itself on its unique style of developing games, preferring to prototype several experimental concepts with small teams, figure out what was working and what wasn't, and get products out the door using whatever stuck.

Nintendo's external development partners would operate in the same manner. In 1999, when the company was working with Angel Studios—one of the companies it had partnered with to create Nintendo 64 games—Nintendo would sign three-month contracts with the developers. Miyamoto would tell them that he didn't want to see design documents or plans, and would instead ask them to develop prototypes for their ideas using smaller teams. If an idea wasn't clicking, Nintendo would shelve the project and have the team focus on the ideas that were, turning those into full-fledged products instead.

Most of Nintendo's games were developed the same way—to be fun, replayable experiences that were quickly prototyped and obsessively polished, all within a relatively short timeframe. Projects like 1080° Snowboarding and PilotWings 64 would routinely be completed in a single year and go on to sell upward of a million units.

The two exceptions to this rule were the company's key properties: Mario and Zelda.

Twilight Princess

Both brands saw a major increase in development time and budget on the Nintendo 64, taking three years apiece to complete. As time and technology would progress, this approach would be more easily justified for Mario, which would frequently sell upward of 10 million units and often act as a showcase for Nintendo's best ideas. In the case of Zelda, however, Miyamoto would caution the development team against letting the games grow too large in scope. Following the release of Ocarina of Time, he would ask director Eiji Aonuma to create a new, more compact Zelda game within a year, leading to the development of Majora's Mask, which was put together using ideas originally intended for other projects. This would continue as Nintendo introduced the Nintendo GameCube and Game Boy Advance, with games like The Wind Waker and Four Swords Adventures seeing relatively short development cycles and sometimes compromising on quality and vision as a result.

What Nintendo would fail to recognize at the time was that The Legend of Zelda was an outlier among its other brands.

While franchises like Mario and Donkey Kong lent themselves to fun, experimental ideas framed within shorter games, Zelda had grown popular for entirely different reasons. Audiences enjoyed Ocarina of Time because of the game's sense of scale and immersion, not just because it was fun to play. Exploration and world-building, not gimmickery, were what defined a Zelda game, and these required considerable budgets and development time to realize. It wasn't until Twilight Princess that Nintendo would acknowledge this and devote the time and budget needed to create a new Zelda of that scale. Inspired by the atmosphere of Ocarina of Time and the grandeur of Lord of the Rings, Twilight Princess would be the first Nintendo game in years that could truly be considered a big-budget, AAA production for its time.

Unfortunately, Twilight Princess would put the Zelda team through a grueling three-year development cycle. While it would sell over 8.8 million units globally, it was also one of the most expensive games Nintendo had ever produced, leading to the company scaling its next Zelda project—Skyward Sword—down considerably, once again in the name of inventiveness and the notion that fun could substitute for scope and ambition. At the same time, Nintendo had also grown apprehensive about developing games they considered too complex or exploratory, for fear that they wouldn't appeal to casual gamers in the company's home market of Japan. This audience, Nintendo felt, was intimidated by 3D games where it was possible to get lost, and would make concessions across both Mario and Zelda games to appeal to them.

The result of this ping-ponging was an entire decade of Zelda games following Ocarina of Time that, while well-received, had turned out rather inconsistent. Majora's Mask was wonderfully inventive, but smaller in scope than Ocarina of Time and didn't provide a reliable template the series could use going forward. The Wind Waker alienated fans with its cartoony visuals and was padded out by lengthy fetch quests, owing to its relatively short development cycle. Twilight Princess had come closest to replicating Ocarina of Time's success, but its enormous open world could feel empty. Finally, Skyward Sword didn't know who it was designed for, neither capturing the exploratory essence of Zelda nor appealing to Nintendo's casual audience. Every game had missed the mark in some way, failing to recapture the pioneering spirit of Ocarina of Time, or be as well-rounded a product as that game had been in 1998. Despite Nintendo's constant promises to surpass Ocarina of Time, it never quite managed to do so.

By 2012, series producer Eiji Aonuma knew that the franchise needed some sort of drastic overhaul, or Zelda was at risk of growing irrelevant amidst a sea of more impressive and focused action-adventure games. Skyward Sword had been released the year prior with one of the largest marketing campaigns in Nintendo's history and only sold 3.6 million units worldwide—far less than Ocarina of Time and Twilight Princess, and also coming up short of The Wind Waker. The series needed to find a new direction for itself, one that could sustain it well into the future.

Fittingly, this reinvention of the franchise would begin in typical Nintendo fashion: with a smaller, more experimental game.

A Link Between Worlds

After the development of Spirit Tracks wrapped in 2009, the majority of its development staff moved on to Skyward Sword, which was to be released for the Wii. Only three members of the team stayed back, among which were designer Hiromasa Shikata (who had served as one of many sub-directors on Twilight Princess) and programmer Shirou Mori (who had served as lead programmer on the Nintendo DS Zelda games). The two, along with a third member, were to brainstorm ideas for a The Legend of Zelda game for Nintendo's upcoming portable platform, the Nintendo 3DS.

Since the Nintendo 3DS was capable of connecting wirelessly to other Nintendo 3DS systems, Shikata and Mouri initially thought up a game around player communication, similar to Four Swords and Four Swords Adventures. When they presented the idea to their managers, it was turned down by Shigeru Miyamoto, who felt the concept sounded stale.

A prototype where Link could blend into walls using Toon Link as a placeholder

Distraught, the trio began rethinking the game from scratch and Shikata chanced upon the idea of Link being able to merge into walls. Based on this idea, Mouri developed a prototype featuring Toon Link, which the team presented to Miyamoto once again in October 2010, this time being granted the approval to flesh out their new concept. Two weeks later, though, the three developers were called upon to help develop launch games for the upcoming Wii U console, and the project was shelved. Shikata would go on to work on Nintendo Land while Mouri would be assigned to New Super Mario Bros. U.

Prior to joining those teams, the developers left their prototype in the hands of Zelda series producer, Eiji Aonuma, hoping that he would help keep their project alive so they could resume work on it in the future. At the time, Aonuma was still working on Skyward Sword, which was to be released the following year. It took another year for the game to be completed, and once it was, Aonuma revived the prototype developed by Shikata and Mouri. In November 2011, he assigned two other developers to it, having them tweak and improve the wall-merging system further so that it could be used to start designing dungeons. Aonuma intended to bring both Shikata and Mouri back once they had completed work on their Wii U projects, and would keep them involved in discussions all throughout the process.

In parallel, Aonuma and Miyamoto had been discussing the idea of porting A Link to the Past to the Nintendo 3DS. In a press interview in early 2011, Miyamoto had stated that he would be interested in porting the game to the Nintendo 3DS with stereoscopic 3D visuals, and would later gauge Aonuma's interest in the idea. Aonuma felt that remaking an older top-down Zelda in stereoscopic 3D didn't sound exciting enough and would elect that the team use Shikata and Mouri's prototype to create a sequel to the game instead. The prototype was originally designed to be played from an isometric perspective, similar to the Nintendo DS Zeldas, but Aonuma felt that the contrast between regular gameplay and Link merging into a wall wasn't striking enough. Using the top-down perspective of a classic Zelda game, he theorized, would contrast better with the side-scrolling perspective of the wall-merging ability.

To convey the idea to the team, Aonuma used a tool to render A Link to the Past's map in stereoscopic 3D, following which Miyamoto greenlit the project. By this point, Miyamoto was no longer taking a hands-on role in Nintendo's major games and was instead focused on mentoring the company's younger employees, which meant that Zelda was now largely in the hands of Aonuma and his team, even more so than in the past few years.

By the time the project had been fully greenlit, it was July 2012 and Aonuma hoped to have this new Zelda released for the Nintendo 3DS the following year. He also had enough time to gauge player feedback on Skyward Sword and learnt that audiences weren't particularly pleased with the linear, restrictive nature of that game. Sales of Skyward Sword had been poor in comparison to other Zelda games as well, and Aonuma understood the need to do something different with the franchise going forward. It was Hiromasa Shikata, who was now back on the project and serving the role of director, that would suggest the change begin with his game.

Shikata had the impression that A Link to the Past allowed the player to tackle its dungeons in any order. When he discovered that this wasn't the case during a replay, he suggested they adopt a non-linear structure for its Nintendo 3DS sequel. While Aonuma and the team liked the idea, it posed a number of problems, chief among which was how items would work. In Zelda games up until that point, the player would clear a dungeon, retrieve an item, and use that item to find the next dungeon. If players were allowed to find and tackle dungeons in any order they liked, the game would need to make every item available to them at once. Aonouma, inspired by a hobby, solved the problem by suggesting an item rental system. Players would be allowed to rent items from a shop using the Rupees they had in their possession. If they died, the items would be returned. However, if they were willing to pay a higher price, they would be able to buy the items permanently and keep them forever. This would give players incentive to avoid dying, make Rupees more useful at the same time, and allow players to explore any dungeon they liked.

File:ALBW Wall Concept Art.jpg

Different visual styles were explored

File:ALBW Zelda Designs.jpg

Concept art for Zelda

Hilda, Zelda's Lorule counterpart

Once the item rental system was in place, it gave birth to a sense of freedom that hadn't existed in prior Zelda games. The team, now having grown to full capacity, designed the game such that players would be able to go back-and-forth across the entire map as they liked. To speed up travel, players could discover new routes or quick-travel points that would warp them to different zones on the map. Like A Link to the Past, the game was broken up into light and dark worlds, Hyrule and Lorule, with the latter being a twisted version of Hyrule. This gave the project its title: A Link Between Worlds.

By this point, the team had also settled on a visual style for how Link would look while he was merging into a wall. The artists experimented with different styles, including graffiti, pop art, and fauvism, eventually settling on a mural-like aesthetic for the wall ability. This gave rise to A Link Between Worlds's villain, Yuga. Once the art team had decided that Link would turn into a mural painting while travelling along walls, the question became why. And so, they decided that the villain of the game would be a crazed painter that could turn people into art using magic.

A Link Between Worlds's artistic bend would carry over to the more technical aspects of the game as well. It was decided early during development that the game would showcase the appeal of stereoscopic 3D on the Nintendo 3DS, and Shiro Mouri—who was now serving as lead programmer—would mandate that the game run at 60 frames-per-second at all times, as it allowed for a smoother 3D experience and would also let the team make use of the Nintendo 3DS's touch screen for quick item selection. Once they learnt what they could do with stereoscopic 3D, the team would design elaborate multi-layered dungeons using verticality to showcase a sense of depth that hadn't been possible with prior top-down Zelda games.

A Link Between Worlds

As development progressed, the goal of A Link Between Worlds became to challenge the conventions of a traditional Zelda game and prove that Nintendo was capable of introducing some much-needed change to the franchise. The item rental and wall-blending systems were both fresh, new features that hadn't been explored before. The game itself was designed to be structurally non-linear and afford the player a great deal of freedom. Puzzles were designed around the concept of height and dungeons even included light platforming elements that weren't typically seen in Zelda. Aonuma would point out that these features were possible because of the younger members of the development team—similar to how Aonuma and Yoshiaki Koizumi had helped reshape Zelda for a new generation on the Nintendo 64.

A Link Between Worlds was completed in October 2013 and released to critical acclaim. Earlier in the year, Aonuma had said to fans that the Zelda development team was rethinking the conventions of the series with its next console game, which was in development for the Wii U. A Link Between Worlds, which was released in the interim, helped give Nintendo's audience confidence that the company was capable of delivering on its promise, and would go on to sell over 4 million units worldwide over the next few years.

Tri Force Heroes

The three Links

Development on Tri Force Heroes began after A Link Between Worlds was completed and lasted less than two years. Since the majority of the Zelda team was already working on Breath of the Wild, Nintendo partnered with Grezzo, the studio behind the Ocarina of Time and Majora's Mask remakes for the Nintendo 3DS, to develop the game. Grezzo was hired to create levels, enemies, and the game's user interface, while Nintendo would program the main game systems and design level layouts.

Prior to developing A Link Between Worlds, director Hiromasa Shikata had served as a designer on Spirit Tracks and felt the co-op mechanics of that game would make for an interesting two-player experience. Shikata felt that the Nintendo 3DS provided an opportunity to allow for easy co-op, unlike prior multiplayer Zeldas on the Game Boy Advance and Nintendo GameCube. The development team settled on a three-player game when system director Shiro Mouri, who had served as a lead programmer on A Link Between Worlds, suggested a mechanic whereby players would be able to climb atop one another and form a totem pole. (The team found that depicting more than three players in a totem pole formation was difficult without zooming the camera out too far or distorting the stereoscopic 3D effect of the Nintendo 3DS.)

Early in development, the team considered including voice chat for when the game was being played online with strangers; however Shikata felt that this would lead to situations where a more experienced player would end up giving orders to the other two, which would go against the spirit of finding solutions to puzzles together. To mitigate the issue, he developed the idea of communication icons, inspired by the Line messenger app. These icons wouldn't convey exactly what the party was meant to do, but would broadly allow players to communicate what they wanted from one another. This, the team felt, would make for a puzzle in itself.

While designing the game's puzzles, the team based them around items that would work best with three-player co-op. The Boomerang could to grab other players, while Gust Jars could be used to blow players over large gaps. In order to make it easier for players to play together, the team designed the game so that it could be shared via the Nintendo 3DS's Download Play feature. In this mode, the game's music would be changed to 8-bit, in order to compress data while transferring the game to another Nintendo 3DS. For players that liked the 8-bit tunes, the team created the Timeless Tunic.

Zelda series producer Eiji Aonuma found the three-player mechanic and totem pole concept interesting because it would mean that the player at the bottom would be in charge of movement, while the players above would be in charge of using items and attacking, which would necessitate cooperation. However, Aonuma requested that the team also implement a singleplayer mode and suggested they allow the player to switch between characters—similar to the very first game he had directed, Marvelous: Another Treasure Island. While playing in singleplayer, the player would be able to take control of two AI-controlled companions dubbed "Doppels". In Japanese, these companions were called "Manebito"—the title of a cancelled piece of software that Nintendo had originally intended to release for the Nintendo GameCube.

Early concept art of Styla's curse

Tri Force Heroes's focus on fashion came from the development team wanting to allow players to collect gear and customize their characters like in Phantom Hourglass and Spirit Tracks. The team eventually settled on the idea of players collecting outfits, which led to the game's story taking place in a kingdom obsessed with fashion. Initially, the story involved a kidnapped princess, but Shikata felt this was unnecessarily complicated and didn't fit the theme or lighthearted nature of the game. The plot was then changed so that the game's princess, Styla, would be cursed to wear an ugly, irremovable jumpsuit and Link would need to find a way to break the curse. At one point, the team even wanted to include an outfit inspired by the Japanese idol industry in the game, but learnt that idols were viewed differently in the west and changed the costume to one of a cheerleader instead, feeling that it broadly conveyed the same image of "giving support".

A much larger change that was made during development was Link's design. The development team initially planned to model Link after his design in A Link Between Worlds, but a while into production Aonuma requested that they use Toon Link instead. Despite the change in the character's appearance, Nintendo would maintain that this Link was the same as the one in A Link Between Worlds, with Tri Force Heroes taking place after that game.

Tri Force Heroes received its western name sometime during E3 2015, just months prior to release. However, Aonuma felt the term "Heroes" was a little too serious and wanted to give this particular Zelda a different, more lighthearted vibe. He was reminded of the French novel, The Three Musketeers, and decided to name the game after the book, with its final Japanese title being The Legend of Zelda: The Three Triforce Musketeers.

Tri Force Heroes was released in October 2015 worldwide, and received mixed reviews, averaging out at a 7/10 on review aggregator Metacritic owing to repetitive gameplay, lack of replayablity, and a weak singleplayer mode. The game would go on to sell over 1.33 million copies worldwide.

Breath of the Wild

Following the release of Skyward Sword, series producer Eiji Aonuma learnt that players weren't happy with the linear, restrictive nature of the game. Aonuma and the Zelda development team had been trying to come up with a new template for the series ever since Ocarina of Time, and had tried a variety of approaches between The Wind Waker, Twilight Princess, and Skyward Sword, each with its own, distinct goals. What Aonuma had determined by 2012 was that the series would need a structural overhaul if it was to remain relevant.

This had already influenced A Link Between Worlds, which was designed to be non-linear in structure and afford the player a great deal of freedom. The goal of A Link Between Worlds was to "rethink the conventions of Zelda" and prove to players that Nintendo was capable of reinventing the series to remain current with modern trends. As development on the Nintendo 3DS game—led by Twilight Princess sub-director Hiromasa Shikata—was ongoing, a separate team led by Skyward Sword director Hidemaro Fujibayashi was also planning the next console Zelda, meant for release on the Wii U.

EARLY CONCEPTS AND DEVELOPMENT

Work on the Wii U Zelda began with two broad goals:

1. To create an open-world game set in a large, seamless environment

2. To rethink the conventions of the Zelda series

Series producer Eiji Aonuma suggested that the team begin by re-imagining Link and the kind of adventures he would go on. The development team had a considerable amount of freedom to brainstorm interesting takes on the character, with over one hundred designs being submitted. One early idea involved a possessed Link with an artificial arm that would be capable of turning into different items. Another re-imagined Link as a modern-day wandering minstrel, riding around on a bike and seeking the power of the sacred "Tri-Caster," which would consist of Din's Drum, Farore's Bass, and Nayru's Keyboard. Younger members of the development staff would even pitch ideas such as "The Legend of Zelda: Invasion," which would see aliens invading Hyrule and stealing cattle, and "Hyrule Wars," a game where Link would charge across a war-torn battlefield with enemies shooting laser beams.

• 

  A more western design

• 

  A more mature tone with a scarred Link

• 

  A possessed Link with an artificial arm

• 

  Link with a samurai haircut

• 

  Link in a space suit with a Metroid

• 

  Biker Link in modern clothing

• 

  With a hat-like hood

• 

  Early version of Link in blue

• 

  Closer to the final look

• 

  The Legend of Zelda: Invasion

In these early concepts, most designs of Link included features such as capes or bags. Because the team intended to create an open-world game, they wanted Link to exude a sense of adventure and imagined him as a travelling hero. The team settled on giving him blue clothing early on, because the color blue stood out against the backgrounds they were designing and because Aonuma felt that his usual appearance in a green tunic and hat had grown predictable. Aonuma also asked the team not to make him look overly "cool," as he didn't want the player to feel that they were controlling an already accomplished hero that had no room to grow.

The Wind Waker with modern lighting and shaders

While the artists were brainstorming ideas for Link's new design, art director Satoru Takizawa and his team were experimenting with different art styles for the game. They began by conducting PBR (physically-based rendering) tests using realistically-lit cityscapes and then brought Link's character models from The Wind Waker, Twilight Princess, and Skyward Sword into their new HD development environment. The team found that The Wind Waker's cel-shaded graphics stood out the most of all three games after they had applied modern shaders and lighting techniques to them, and decided that cel-shading would once again form the basis for their new art style. At the same time, the team understood that The Wind Waker's cartoonish visuals were controversial and too childish in appearance for the kind of game the new Zelda was going to be. Wanting to strike a balance between the cel-shaded look of The Wind Waker and the more realistic games like Twilight Princess, they opted for a visual style that would mix cel-shading with realistically proportioned characters. This, Takizawa reasoned, worked well with the game’s unique physics-based gameplay. It didn’t look overly realistic, but was believable enough that it would remind players of real life and convey how the game’s physics worked at a sub-conscious level.

These experiments were conducted in 2012, and in January 2013 Nintendo would officially announce that development of a new Zelda game for the Wii U was underway. Aonuma would announce his intent to rethink the series' conventions, alongside a remaster of The Wind Waker for the Wii U, based on the team's graphical experiments.

OPEN-AIR EXPLORATION

Development on the new The Legend of Zelda game for the Wii U kicked off in earnest in early 2013. Skyward Sword director Hidemaro Fujibayashi had been put in charge and began pondering how to rethink the conventions of Zelda within the structure of an open-world game. Fujibayashi wanted to create a game where the player would progress through the world, spot interesting terrain, and intuitively want to experiment with what they could do within the game's world. To accomplish this, Fujibayashi felt it was necessary to create an "active" game (where the player would have a large number of options available to them and be encouraged to take initiative and experiment) rather than a "passive" one (where the player would simply play within the confines of pre-scripted mechanics). The development team began experimenting around this idea by creating an area where the player could go anywhere and would not be obstructed by impassible walls. If they came to a wall, they would simply be able to climb to the top, jump off, and glide to their destination. This would transform the wall into another optional path the player could take and allow for "active" decision-making while traversing the world. In particular, Fujibayashi felt that the ability to glide anywhere in the game world was a crucial ingredient that would give the player a newfound sense of freedom, and that it represented a return to the "essence" of Zelda, referring to the open-ended nature of the original Nintendo Entertainment System game. To accommodate this style of gameplay, the team added a jump button to its game—a first for the series.

To convey the idea to general manager Shigeru Miyamoto and Eiji Aonuma, Fujibayashi and his team fleshed out their open-world prototype. They created an early version of the game's starting area, with Link getting out of the Shrine of Resurrection and being able to explore the area around it. The team scattered trees everywhere, making certain that every single tree was climbable, and hid Rupees and other items in places they felt the two might wander off to whilst exploring. When asked what the player would do in his new Zelda game, Fujibayashi replied that players would be able to do "everything".

• 
  Prototype screenshot: Link skydiving into a map of Kyoto
• 
  Prototype screenshot: Link diving toward Himeji Castle

Fujibayashi's open-world idea was greenlit for development and the team began working on a larger prototype using the map from Twilight Princess. The idea was to prototype a large world with the entire map of Twilight Princess embedded within it, and see if the team could turn that game's map into a single, seamless area using the increased processing power of the Wii U. The terrain of Twilight Princess was placed in the upper-left corner of the world, with the remainder being occupied by a map of Nintendo's home town of Kyoto, using map data from Google and Zenrin Co. Ltd., in order to help the team get a sense of scale. The Kyoto section of the map was created using 3D tools, with models of real-life landmarks such as the Kinkaku-ji and Kiyomizu-dera temples placed within it. Although the map was based on Kyoto, the team also placed a model of Himeji Castle from Japan's Hyōgo Prefecture, making a model of Link stand on top of it and look down over the terrain.

During these early experiments, Fujibayashi and the development team were also considering the idea of Link skydiving from a high altitude, similar to The Legend of Zelda: Skyward Sword. While this feature wouldn't be included in the final game, it would subsequently be revisited in the sequel.

Breath of the Wild's map, made up of trianglular shapes meant to draw the player's gaze

After work began on the actual game's world, the development team used the "Field Triangle Rule" whilst designing its fields, scattering a variety of triangular shapes across the map. Triangular shapes would accomplish two things: a) draw the player's gaze, and b) give players a choice as to whether to go over the triangle or around it. These triangles would be of varying sizes, each serving different purposes. Larger triangular shapes such as mountains would serve as landmarks that could be spotted from a considerable distance away and draw the player's attention. Medium-sized triangles were given the role of partially obscuring structures, facilities, and other elements in the world behind them, making the player want to explore. Finally, smaller triangular shapes such as rocks and undulations were included to serve as obstacles across shorter distances and keep the player on their toes. Rectangular shapes were used to similar effect. Unlike triangles, which would gradually reveal what was behind them as you approached, rectangles were useful for obscuring things from sight completely.

One of the team's primary objectives was to ensure that the game world felt "dense" enough. They did this by balancing three keywords: distance, density, and "time spent". Since this was Nintendo's first open-world project, Fujibayashi and the terrain artists would walk around the company's home town of Kyoto with a map of the city, using the experience to determine just how large they wanted their game world to be, and how long it would take players to traverse it. This would also give the team a sense of how to spread events, enemy encounters, and other such elements out across the world, accounting for both distance and density. The third keyword, "time spent," was used to refer to the "amount of time the player might spend on a single element placed in the game world," depending on its size and complexity.

The developers also tried to ensure that players would have sufficient reason to visit every part of the game world. They initially attempted this by erecting Sheikah Towers throughout the map, populating the space between them with various events. However, this made playtesters feel as though they were being herded from tower to tower, and so the team re-examined its approach. They began using the concept of "gravity," strategically placing structures of varying visibility and significance across the map, which were meant to draw the player in different directions, allowing them to get sidetracked and constantly discover new things. Given the number of playtesters involved, Aonuma would request that the team build a PC tool that could track the journeys of up to 100 players, updating with markers of their travels every hour. This would allow the team to identify areas of the map players weren't engaging with and modify them.

Eye-catching landforms, reminiscent of Xenoblade

"Gravity" would also affect players differently depending on their playstyle. The team found that more aggressive players might prioritize enemy encampments in the hope of finding better gear, whereas those that enjoyed uncovering the game's map might work towards activating all the Sheikah Towers first instead. During night-time, the concept of gravity would work differently, with structures that were emitting light drawing the player's attention over everything else.

The new Zelda's world would eventually end up being over 12 times the size of Twilight Princess. To help design the world, the Zelda team enlisted the aid of Monolith Soft, developers of the Xenoblade games. Monolith Soft had amassed a tremendous amount of experience designing an open-world game with Xenoblade Chronicles X, and a large number of designers that had worked on the game helped create environments and dungeons for this new Zelda, sometimes giving the landscape of Hyrule a look reminiscent of the Xenoblade series. During the development process, the team also realized that an open-world game's appeal didn't come solely from an abundance of story events and things to do, but that a large part of the appeal was simply being able to journey and explore, to which pockets of emptiness were actually beneficial.

The Zelda team also looked to western-developed open-world games such as The Elder Scrolls V: Skyrim for reference, to determine what sort of challenges they would encounter and how they could improve upon the template of open-world titles. The team also made it a point to design its game world such that players could play for brief periods of time, without being trapped in any one activity for too long. Dungeons in Skyward Sword could take several hours to complete, and so the team focused on just four main dungeons for its new game, and a number of smaller puzzle rooms scattered around the world, dubbed "Shrines".

Upon experiencing the game's world and wildlife setting, Nintendo of America dubbed this new Zelda an "open-air adventure," and gave it its title: Breath of the Wild.

MULTIPLICATIVE GAMEPLAY

The Zelda development team's goal with Breath of the Wild was to create a vast, non-linear, open-world game. One of the challenges of creating such a game, they realized, would be designing a wide enough variety of puzzles and other bespoke gameplay elements to keep the player suitably entertained all throughout. Director Hidemaro Fujibayashi understood that designing individual puzzle mechanics by hand was going to require a tremendous amount of resources, and began brainstorming if it was possible to mass-produce puzzles around a consistent set of rules instead.

• 

  A prototype meant to demonstrate Breath of the Wild's gameplay systems to the development team

Fujibayashi found inspiration in Breath of the Wild's gliding mechanic, and how it had materialized as a side effect of walls being made climbable. This, he felt, was "multiplicative gameplay," where one interaction between the player and field had led to the rise of a new gameplay possibilities. Fujibayashi and and technical director Takuhiro Dohta looked to prior Zelda games and the interactions that were possible between the player and game world in those titles. In Ocarina of Time and Twilight Princess, players had been able to cut down signboards and watch the individual pieces of wood float down a river. In A Link to the Past, the team had originally intended to allow players to start "endlessly expanding fires" with the lantern, but the idea had to be shelved due to hardware and time constraints. These were all scripted puzzles involving simple science laws, but it was now possible to expand upon them using modern hardware. This, Fujibayashi and Dohta felt, would aid in the development of an "active" game, where the player would constantly feel encouraged to experiment with the game's logic and try out different things.

Fire interacting with arrows

To help convey this to the development team, Dohta developed a prototype that resembled the original The Legend of Zelda for the Nintendo Entertainment System, but allowed Link to interact with the game world in a variety of ways. Link would be able to use the elements, such as fire or wind, to trigger different reactions in the game world. He would be able to set fire to trees and have the fire spread to the surrounding shrubbery. He would be able to cut trees down, roll them into the water, and ride them down a river. The elements would also be capable of interacting with other objects like arrows, to create Fire Arrows, which could then be used to deal with enemies. This would lead to the development team creating an entirely new physics system for Breath of the Wild, where elements could interact with nearly every single object in the game, causing different kinds of reactions, and be used in creative ways to solve puzzles. The rules of the game were designed to be exaggerated versions of real-world laws, which the team felt would make them feel believable and encourage players to test their theories out in-game.

Breath of the Wild was originally meant for release in 2015; however, in March 2015 producer Eiji Aonuma posted a video to Nintendo's social media channels, announcing that the game had been delayed. Aonuma would go on to clarify that this was because the development team had "experienced firsthand the freedom of exploration that hasn't existed in any Zelda game to date" and that they had discovered several new possibilities for the game that they wanted to flesh out. Additional comments from Miyamoto would suggest that the advent of the game's complex physics system had led to its delay.

STORY & CHARACTERS

As part of creating the final designs for Link and Zelda, the team re-examined the way Link's eyes looked at the request of producer Eiji Aonuma. Aonuma felt that, in prior games, Link's eyes always appeared as though they were strained and narrowed, even when the player wasn't doing anything particularly strenuous. As a result, art director Satoru Takizawa opted to make Link's eyes more rounded, in order to give them a more melancholic appearance. This look was finalized prior to the game's debut trailer at E3 2014.

Concept art for Zelda.

For Zelda's new look, Takizawa began designing her outfit based on her blue, casual attire from A Link to the Past. Like Link, Zelda was given more rounded eyes and was also given thicker eyebrows so it would be easier to convey facial expressions. The amount of fan talk around her eyebrows, Takizawa would later say, came as a surprise to him.

Meanwhile, conversations around Ganon's appearance had already come up early in the project, when Takizawa and director Hidemaro Fujibayashi were discussing the game's broader story. The inspiration behind this incarnation of Ganon was Taira no Masakado, a real-life samurai belonging to the Kanmu Taira clan of Japan's Heian period.

Masakado was notable for leading the first recorded uprising against the central government in Kyoto, and was considered something of a demigod by the locals, who felt the need to "appease his malevolent spirit" for fear that neglect would lead to natural disasters and other misfortunes. As a result, it was decided that Ganon would similarly need to be represented as an actual calamity of sorts—something that was overwhelmingly terrible, beyond the control of human hands, and wouldn't require words to explain.

Calamity Ganon was envisaged as one of the three "main scenarios" of Breath of the Wild's story, with the game's end goals being to learn about the past, acquire the Master Sword, and defeat Ganon. However, since telling a linear story was at odds with the goal of Breath of the Wild as an exploratory, non-linear game, the Master Sword quest was turned into a separate optional event and the quest to uncover the past designed as a series of fragmented, optional discoveries.

WORLD AND VISUAL DESIGN:

Jōmon vessels inspired the look of the Guardians and Ancient Shrines

Breath of the Wild was going to allow players to explore a vast Hyrule, and the development team put a great deal of effort into world-building and the lore for the game's different regions, tribes, and cultures. Of particular importance were the Sheikah, whose ancient architecture defined much of this version of Hyrule, including the Shrines, Guardians, and the Sheikah Slate. The Zelda team looked to the Jōmon period in Japanese history for the Sheikah's architecture, with the designs of the Shrines and Guardians both inspired by flame-styled Jōmon pottery. The Sheikah themselves were designed to resemble Japanese shinobi archetypes, with their home town, Kakariko Village, being modeled after a Japanese mountain village.

In addition to the Sheikah, distinct characteristics and back-stories were also designed for the Gorons, Zora, Rito, and Gerudo, as well as a mysterious tribe dubbed the "Zonai", who were said to be an ancient civilization that collapsed long ago. The Zonai Ruins in Faron and Lomei Labyrinth in Akkala were created to give players a sense that Hyrule was once shared by someone other than the Hylians and Sheikah, who also worshipped the Triforce.

To make Hyrule feel like a living world, the team also designed a wide variety of wildlife that would roam the land and be able to interact with the player. Most fauna in the game was designed to be non-aggressive, but would defend itself if provoked. The player would also be able to hunt a large number of the fauna for meat and other resources that could be used in cooking and crafting. Instead of allowing players to cut grass for Hearts, the developers wanted them to survive using natural resources.

"During production, there were a number of things that I focused on expressing visually because we couldn't express them in other ways," art director Satoru Takizawa would say. "Smell is one of them. A gaming console can't simulate a sense of smell, but I talked with the terrain and effect designers about how to create a world that gave you the impression of a smell. For example, in the real world, there is a very specific smell right before it rains. I wanted to create the illusion of that smell with the visuals in this game. At the same time, I focused on creating a world where imitation sounds could be communicated. I wanted the onomatopoetic sounds of humidity, mugginess, or dryness to be felt and conveyed without sound effects. If we could simulate senses like smell and touch through visual information, the world would feel truly alive, and that was one of our biggest goals."

Unlike Skyward Sword, the team conveyed a lot of story and world-building through the environment in Breath of the Wild, with even more facets of the world being expanded upon in the game's Creating a Champion art book.

A return to the original logo used on the Nintendo Entertainment System

Similarly, the user interface (UI) for Breath of the Wild was designed to be the polar opposite of Skyward Sword as well. Where Skyward Sword's UI could often be large and intrusive, the development team's goal with Breath of the Wild was to create a UI that only displayed information when necessary. The key to this was a warm-white color the team dubbed "Zelda White" and used across icons, logos, and even the game's packaging. Icons in the game were designed to be borderless and simple, and UI elements were cut down to the bare minimum to allow players to immerse themselves in the game's world.

Based on feedback from Nintendo of America, the team also developed a "Pro HUD" option for veteran players that would reduce on-screen UI elements even further.

To signify a return to Zelda's exploratory nature, the Japanese logo for Breath of the Wild reverted to the style used in the original The Legend of Zelda for the Nintendo Entertainment System, with the North American and European logo remaining the same but recolored Zelda White.

MISCELLANEOUS DEVELOPMENT NOTES

• Breath of the Wild was originally meant to be released just for the Wii U. In early 2016, after it had already been delayed once, Nintendo decided to make the game a launch title for the Nintendo Switch and the team was asked to make the game cross-platform across both the Wii U and Switch.

• This led to the team having to scrap ideas that would have made use of the Wii U GamePad, as well as making changes to the game's story, which would initially have featured the Sheikah Slate in a more central role.

• Aonuma would reveal that Miiverse was responsible for giving him easy access to the opinions and feedback of overseas fans, which helped decide upon a direction for Breath of the Wild.

• An early idea involved Link adopting a dog that would serve as his navigator. While this idea didn't make it into the game, it did inspire the Wolf Link Amiibo feature.

• Horses in Breath of the Wild had their own dedicated programmer and the feature was constantly worked on and tweaked all throughout development.

• One of the reasons Link doesn't wear a hat in Breath of the Wild was that the team wanted to make the character look cool. However, making a hat look cool had gotten harder and harder with modern graphics. In Twilight Princess, art director Satoru Takizawa dealt with the issue by making Link's hat longer, but this had the side effect of making it look comical when it flapped in the wind.

• Eiji Aonuma would reveal that he had attempted creating ambitious open-world games with both The Wind Waker and Twilight Princess; however, the team wasn't able to accomplish its goals with those games due to time and manpower constraints.

• Composer Hajime Wakai came up with the game's musical style through trial-and-error. He initially tried including music from Twilight Princess, but because this version of Hyrule was so vast, Wakai felt there was no music that would be able to live up to the world's grandeur and sense of inspiration. This led to the decision to create more ambient, understated music instead. Once this decision was made, the team began using the term "environmental BGM" instead of "world BGM" during development.

• Games that members of the team were playing whilst developing Breath of the Wild included The Elder Scrolls V: Skyrim, The Witcher 3, Far Cry 4, and Monster Hunter 4.

• The Fishing Rod wasn't included in Breath of the Wild because the team didn't want to include items that only served a single purpose.

• One of the unused ideas the team considered was giving Link the ability to shrink and enter tiny towns, similar to The Minish Cap.

• Aonuma would compare Breath of the Wild to going from Japanese food to Western style food. Presumably, he was referring to the game's open-world elements and non-linear structure, both of which are rarities in Japanese games.

• Despite Aonuma's love for cooking and stating years prior that he would like to include the feature in a future Zelda game, the cooking feature in Breath of the Wild came about naturally, not because he suggested it.

• While Breath of the Wild was designed to be the polar opposite of Skyward Sword, a handful of its features first appeared in that game. These include the ability to dash, the stamina gauge, and Link's shield taking damage from enemy attacks.

• Towards the end of development, Breath of the Wild's development team grew to over 300 people in size, between Nintendo and other collaborators.

• Aonuma asked the development team to add the Master Cycle Zero to the game as one final reward the player would unlock after playing all the way through to the end of the Champions' Ballad DLC. Following this, the team actually went out and purchased a motocross bike they could ride to understand how one should control.

Breath of the Wild was released globally in March 2017 to critical acclaim. Following Twilight Princess, it was the second time Nintendo had assigned a AAA development budget and schedule to a Zelda title, and the game built upon a number of elements that players had enjoyed in both Ocarina of Time and Twilight Princess, including a more realistically proportioned Link, a vast world to explore, and the ability to ride on horseback. Breath of the Wild would go on to sell over 18 million units across the Nintendo Switch and Wii U, successfully growing the brand and providing it with a template for future games. Producer Eiji Aonuma would state that the game's "open-air" structure would form the basis for future Zeldas, and Nintendo would announce a sequel to the game two years later.

References

Real world

hide ▲

Advertisements for The Legend of Zelda Series • Development of The Legend of Zelda • Heroes of Hyrule • Nintendo Fun Club News • Nintendo Selects • Stylus • The Legend of Zelda - Cancelled Games • The Legend of Zelda - Series Sales

Category:Real world