How ‘Banjo-Kazooie’ Became a Bridge Between Marios

Art may largely be a matter of taste, but one conclusion is close to inarguable: 1998 was the best year ever for video games, producing an unparalleled lineup of revolutionary releases that left indelible legacies and spawned series and subcultures that persist today. Throughout the year, The Ringer’s gaming enthusiasts will be paying tribute to the legendary titles turning 20 in 2018 by replaying them for the umpteenth time or playing them for the first time, talking to the people who made them, and analyzing both what made them great and how they made later games greater. Our series continues today with Banjo-Kazooie, the beloved platformer from renowned game developer Rare that built on Mario’s model for 3-D design and, in retrospect, helped set the stage for the long-neglected genre’s recent resurgence.

“The first thing you have to know about this game is that it behaves a lot like Super Mario 64,” Nintendo Power told its hungry readers in August 1997, in the magazine’s first hands-on preview of Banjo-Kazooie. It was simultaneously one of the most encouraging comparisons that the publication could have made, and one of the most intimidating. Super Mario 64, which had debuted in North America less than a year earlier, was one of the most acclaimed and momentous games of all time, a title that the same magazine — granted, not an unbiased source — had proclaimed “nothing less than a revolution” and “the new gold standard in video games.” Labeling Banjo as a platformer that performed like Mario 64 was akin to comparing a Ball brother to Steph Curry before the kid’s first professional field goal.

Some comparisons can’t be escaped. In light of its look, timing, and gameplay, Banjo was destined to be scrutinized as a successor to the genre’s reigning king, with its most memorable moments measured against Mario’s and all its flaws laid bare. For the most part, though, the hype proved to be a blessing. Banjo lived up the impossible billing, or at least came as close as any game could.

After many months of build-up, Banjo-Kazooie came out on Nintendo 64 in June 1998, 20 years ago this month. It garnered rave reviews, won awards for graphics, music, and overall quality, sold more than 3.5 million copies worldwide, and spawned four sequels or spinoffs. Inevitably, contemporary critics continued to compare it to Mario after its release, and some of them even praised it as a more polished product, if a less innovative one. But Banjo-Kazooie was more than a Mario clone: While it took inspiration from Mario, it also anticipated where Mario was going to go. In forming a bridge between watershed Mario moments, the series made an enduring name for itself.

For most of the 1990s, Gregg Mayles made beloved video games. But during a lengthy lull in the middle of that decade, Mayles couldn’t decide what to design next.

Mayles, who’s now 47 and has served since 2007 as the creative director of U.K. developer Rare, joined the then-fledgling company as a quality tester in 1989. During his first several years at the legendary Leicestershire studio, he designed some of the company’s classic titles, including punishing Nintendo Entertainment System shoot ’em up Battletoads (1991) and Super NES side-scrolling platformers Donkey Kong Country (1994) and Donkey Kong Country 2: Diddy’s Kong Quest (1995). After Diddy’s Kong Quest, Rare started prepping a 1996 sequel that would complete the Donkey Kong Country trilogy, but Mayles wanted to work on something new.

For game-makers, the mid-’90s was a time of daunting technological transition. The 16-bit Super Nintendo had been out since 1990 in Japan and 1991 in North America, and with both the SNES and Sega’s Genesis system nearing the end of their life cycles, the adolescent industry was approaching its awkward, exciting transition to 3-D, which required a complete reimagining of many game mechanics.

For Donkey Kong Country, Rare had developed a graphical technique called advanced computer modeling that eked extra detail out of the aging SNES. After finishing DKC’s sequel, Mayles says via Skype, “We took the ACM technology we’d developed and thought, ‘What other genre could we apply it to?’” The company decided to design an adventure-RPG, which was tentatively titled Dream: Land of Giants. It featured a boy named Edison who had a sword, a dog, a girlfriend, and a grudge against a gang of pirates. Mayles designed a demo that “looked like the next step up from Donkey Kong Country,” but the hardware hamstrung his ambitions. “The technology was never going to be able to do what we wanted it to,” Mayles says. Fortunately, new hardware was on the way.

From 1994 to 2002, Rare was part-owned by Nintendo, and it designed games exclusively for Nintendo systems. Thanks to that relationship, Rare was among the first crop of developers to gain access to “Ultra 64,” the development hardware for Nintendo’s next system, the souped-up N64. Rare soon scrubbed its SNES plans and ported its project to the more powerful forthcoming platform, where it ran on a “pseudo-3-D” engine. But as the studio struggled to adapt to the new technology, the game’s story stagnated. The boy began to seem generic, and Rare replaced him. “We briefly dabbled with a rabbit,” Mayles says. “I have no idea where the rabbit came from. It only lasted about two or three days.”

Next, the rabbit became a bear. The bear wore a backpack, the better to carry adventure accessories. He had sneakers, a cap, and a name that went with an instrument that he’d later “learn” to play: Banjo. “It was a bit of a strange character,” Mayles says.

Gregg Mayles

Banjo was born, but Dream died there: After more than a year of fumbling Rare’s first attempt at 3-D, Mayles says, “It just felt like the end wasn’t in sight.” To salvage something from that effort, the studio decided to return to charted territory and build a DKC-esque platformer fronted by Banjo. Mayles dubbed it “2.5-D Banjo”: Like its 2-D antecedents, the game was a side-scroller, but it offered additional depth and range of movement.

At best, 2.5-D Banjo would have amounted to an incremental move for Rare. But the extended, scrapped development of Dream that preceded the proposed compromise project inadvertently saved the studio from settling for a forgettable face-saver. About two months into Rare’s efforts on “2.5-D Banjo,” Nintendo gave the designers a glimpse of a “really, really early” development version of Super Mario 64, which would launch alongside two other titles when the Nintendo 64 debuted in Japan in June 1996 and introduce Nintendo’s iconic character — and a generation of gobsmacked gamers — to a fully realized third dimension.

“As soon as we saw that, we thought, ‘Well, this is the direction all 3-D games are going to go,’” Mayles says. Rare soon ceased work on its 2.5-D tweener and built its own 3-D engine, using Mario as a model. That took care of the tech; the new mascot, meanwhile, gained a sidekick, an acerbic bird named Kazooie who traveled in the more affable Banjo’s backpack and augmented his companion’s physical and comedic powers. Finally, the foundation for Rare’s magnum-opus platformer appeared to be in place.

Mayles once described Battletoads as an example of Rare “looking at what was popular and then putting our spin on it.” To a certain extent, the same could be said about Banjo-Kazooie. The magic of Mario 64, Mayles remembers, was the freedom it promised the player. “If you saw somewhere, you could go to it,” he says, adding, “It was such a significant step forward in terms of making it feel like you could go and do anything.” In attempting to top Nintendo, Rare sought not only to capture that feeling, but to expand it so significantly that even Mario might feel restrictive by comparison.

Banjo tasks the player with maneuvering its bear-bird combo through a series of nine themed stages, all linked by Gruntilda’s Lair, a hub world that’s home to Gruntilda, an evil witch who has kidnapped Banjo’s sister Tooty. Each of the game’s early stages offers upgrades to the flying, climbing, swimming, and hopping protagonist tandem’s set of moves, enabling more complex attacks and techniques for traversing the maps and a sense of progression that was missing from Mario. In addition to fending off attacks by Gruntilda’s minions and avoiding the typical hazards and (literal) pitfalls of platformer environments, the player must solve puzzles and explore nooks and crannies to collect hidden puzzle pieces (“Jiggies”) and musical notes that unlock new levels and lower barriers on the quest to rescue Tooty.

Although Banjo and Kazooie visit some terrain that resembles Mario maps from two years before, including a desert, an ice level, and a haunted house, Mayles observes that “all the levels in Banjo are a lot more grounded in reality.” They’re also larger, more detailed, and more richly rendered, a product of Rare’s extra time with the N64 architecture, as well as the studio’s expertise at making games look good. Banjo, which is widely regarded as one of the most visually vibrant titles ever to grace the N64, looked almost too good for its time, straining the pre–Expansion Pak system to such a degree that Rare was forced to implement processor-saving kludges like 2-D background sprites and selective loading. “There was a lot of effort that went into culling the background, so when you couldn’t see parts of the level, they weren’t being drawn,” Mayles says.

Gregg Mayles

Starting with the second stage, Treasure Trove Cove, which wraps around a massive mound of rock, Mayles designed environments with an eye toward limiting the number of objects on screen, although he found that obscuring sections of each stage served a secondary purpose. “It was almost like a happy accident, that hiding stuff gave you a reason to go and explore,” Mayles says. “But originally, it certainly came from a very conscious decision of trying to get the game to run at a decent frame rate.” With a tiny team by today’s standards — only 10 people when Banjo began, eventually expanding to roughly 15 — optimizing the tech took enough time that the game, which was originally slated to arrive in November 1997, swapped release dates with Rare’s Diddy Kong Racing, an intended 1998 title that was ahead of schedule.

Although Nintendo knew roughly what Banjo was supposed to be, the company played no part in its development. Mayles and his team had no need of outside help, because late-’90s Rare was Reaganing. According to data from the Internet Game Database, over an eight-year span starting in 1994 — encompassing the period from Donkey Kong Country to 2001’s Conker’s Bad Fur Day — only Capcom, Square, and Nintendo EAD developed more games with an IGDB user rating of at least 80 percent. Rare, which trailed Nintendo by only one 80-something-percent title during that incredible run, demonstrated a commitment to quality that crossed genres, alternating between platforming touchstones like Banjo and Conker and groundbreaking first-person shooters GoldenEye 007 (1997) and Perfect Dark (2000).

Although that string of releases seems magical — especially since the secretive studio tended not to talk about its methods — it was another product of Rare’s learning and borrowing from the best, as it had in loosely basing Banjo on Mario 64. “We’d seen how Nintendo worked,” Mayles says, continuing, “Nintendo was able to do these very different games. That’s because they kept the teams isolated and almost kept them in competition with each other — just the right level of competition.” At Rare, discrete teams created distinctive games, and designers found the right fits through a Sorting Hat–style process. “If people had a more serious outlook on stuff, then they naturally ended up on the games that had a more serious edge,” Mayles says. “All the people that had the Banjo sense of humor ended up on Banjo.”

Mayles describes that sense of humor as “very dry, very typically British, slightly sarcastic, happy to poke fun at ourselves.” Much of the dialogue was done on the fly, with a lot of lines almost ad-libbed. And while the game’s kid-friendly language matched its cartoony look and inclusive “E” rating from the ESRB, Banjo’s comedy had an off-kilter undercurrent that would reach its fullest expression in the cute-faced but foul-mouthed Conker. “All the characters basically had something wrong with them,” Mayles says. That subtle subverting was one way in which Banjo departed from the plumber-approved template perfected by the more anodyne Nintendo, but the most important precedent that Rare’s platformer set had more to do with its structure than its looks or its script. Although Mario 64 pioneered a new dimension, Banjo was the first to truly explore the space.

The transition to 3-D placed undue demands on the platformer, a genre that historically relied on a stable, all-seeing perspective and a precision of movement that two-dimensional models made easy. “As soon as you move to 3-D, trying to re-create that accuracy is almost impossible because you can’t see your character as accurately and you can’t judge distances as well,” Mayles says. Data from the MobyGames database shows that platformer frequency as a percentage of all titles dropped off drastically at the end of the 16-bit era, just as Mayles was casting about for a new direction after Donkey Kong Country 2.

As much as he admired Mario 64, Mayles thought that the game had half a foot in each era and suffered slightly from being beholden to Mario veterans who expected it to provide an addictive test of quick-twitch timing that could rival the reflex-straining stages of, say, Super Mario Bros. 3. Rather than try to replicate a style of gameplay that 2-D did best, Rare embraced 3-D’s exploratory possibilities and deemphasized the demanding parkour of previous platformers. “It was a very conscious decision not to have pixel-perfect jumps,” Mayles says, adding, “[Banjo] wasn’t supposed to test your ability to land on really tiny places. [It was] more a case of rewarding exploration and discovery.”

That shift in focus led to one major innovation that Nintendo would one day adopt. In Mario 64, the player collects power stars that are scattered across 15 worlds, each of which holds seven stars. But those stars aren’t distributed in a way that allows the player to collect them all at once: Instead, the player tackles one distinct challenge after another and, after earning each task’s associated star, is booted out of that environment and returned to the Peach’s Castle hub that links the levels. Upon reentering the same stage, a new challenge — and a new star — is available. Although Mayles says he understands why that structure was simpler from a design perspective, it was one of his pet peeves about an otherwise-great game. “It was like a stage, where they’d rearrange all the props and they’d let you back on the stage again,” he says. “Personally, I really didn’t like that. I think it ruined the sense of immersion in the world, so that was one of the very first things, was, ‘Let’s keep players immersed in the world. Don’t throw them out. Make sure players can complete everything while they’re in there.’”

In Banjo-Kazooie, when a player collects one of the 10 puzzle pieces hidden in each environment, the game goes on without interruption. The player can clean out each world without being bounced back to Gruntilda’s Lair until he or she is ready to leave. “It’s very different, because Mario games are usually about, ‘Oh, here is the start, here is the end. Now move from the start to the end,’ while Banjo-Kazooie doesn’t have that at all,” Jonas Kaerlev, the founder of indie developer Gears for Breakfast and the designer of 2017 platformer A Hat in Time, says via Skype.

Gregg Mayles

Nintendo stuck to its flagship franchise’s N64 formula for many more Mario releases. But last year, for the first time in a 3-D Mario title, the company abandoned its traditional layout and followed Banjo’s lead. In Super Mario Odyssey, the critically acclaimed and massively successful Switch installment of the series, the player no longer leaves the world after nabbing each power moon. Odyssey allows players to explore each stage at their leisure, with no forced exits or countdown clocks to destroy the sense of continuous discovery. Like Banjo before it, Odyssey is about exploration more than coordination.

That’s not the only aspect of Odyssey that seems to owe a debt to Banjo: Odyssey features hidden portraits that reveal new elements of (or connections to) other worlds, a staple of both Banjo-Kazooie and its 2000 sequel, Banjo-Tooie. Even more notably, Odyssey includes a capture mechanic that allows the player to possess and control other objects and creatures, much as shaman Mumbo Jumbo transforms Banjo throughout the Banjo-Kazooie series. In Odyssey, Mario morphs into some of the same forms that Banjo had assumed years earlier, including a tank and a Tyrannosaurus rex. “As soon as I saw the first one of those in Mario, I thought, ‘Oh, this sounds a bit familiar,’” Mayles says.

Mario’s cap companion, Cappy — which hitches a ride on his head, similar to how Kazooie rides on Banjo’s back — even has googly eyes, a staple of many an inanimate object in Banjo’s world. “[Super Mario Odyssey] is more Banjo-Kazooie than it is Mario,” Kaerlev says. Decades after Nintendo Power’s preview, the franchise comparison tables have turned.

Nintendo declined to comment on the similarities between Banjo and Odyssey, but if the resemblance isn’t a complete coincidence, Mayles wouldn’t mind. “I don’t think good ideas are the property of anybody,” Mayles says. “Any designer that says they’re not influenced by other games or other designers is either an outright genius that never plays any other games or is being economical with the truth.” In this case, Mayles subscribes to the saying that turnabout is fair play. “We took enough influences from Nintendo’s games, so if they were influenced by Banjo, then that’s kind of a nice thing,” he says.

Rare’s pursuit of seamless immersion extended to the music that company composer Grant Kirkhope cooked up for the game, as sound is central to Banjo’s appeal from the instrumental medley in its intro video on. Mayles asked for a soundtrack that would mimic the innovative early-’90s iMUSE system developed by LucasArts, which cleverly synchronized sound with on-screen events by transitioning smoothly between themes as players moved between areas. “We set Grant an even more difficult challenge where we wanted different worlds but having different layers of music,” Mayles says. “So if you go really high, we want some of the channels to fade out. If you go underwater, we want to make it sound like it’s underwater. If you go down to the beach, then you can hear coconuts clapping.” Mayles expected that all games would adopt such a system eventually and expresses some surprise that most modern games are “going for these almost grand pieces of music, but it’s very clear where the lines are between the music and when it changes from one [track] to another.”

Rare even hoped to blur the lines between games. One of Banjo’s other lasting legacies is a planned, fabled feature that failed to come to fruition: The Stop ’n’ Swop. Rare, which had a history of experimental engineering, discovered that when an N64 cartridge was powered down, its flash memory would retain data for seven to eight seconds before it was wiped. When the Banjo team learned that, Mayles says, “We just started thinking about what would happen if we could use the ability for the machine to remember what game was in there first, that when you put a second game in … there could be a link between the two cartridges.” It was an unheard-of idea in an era before game consoles came with internet connectivity.

Mayles made a plan for Banjo to link up with Donkey Kong 64 and, someday, Banjo’s sequel, inserting several mysterious items — six secret eggs and an ice key — into the code for Banjo-Kazooie that could be unlocked in Tooie. But after Banjo came out, an alteration to new N64 systems’ hardware made it almost impossible to pull off the cartridge swap before the flash memory faded. Rare’s plan wouldn’t work, but because the Banjo-Kazooie cartridges couldn’t be patched (again, no internet), the signs of the Stop ’n’ Swop scheme were still there, tantalizing players like the distant Dam island in GoldenEye. Crucially, a cutscene at the end of the game teased the existence of the eggs, fueling a Ready Player One–style search for their locations and purpose.

In 1999, a then-24-year-old U.K. player named Alan “Ice Mario” Pierce used an Action Replay device (and a lot of trial and error and reverse engineering) to access the hidden items. When he found the first egg, he says via email, “the whole Stop ’n’ Swop thing went mental. … People were frantically trying to find out what these items did, or whether there were any more secrets to be found.” To further their efforts, Pierce cofounded a game-hacking community with the extremely 1999 name The Rare Witch Project. Egged on (so to speak) by occasional cryptic comments from Rare, the collective went on to make more discoveries and still exists today. Although many such groups have mined subsequent non-Rare releases for R-rated mini-games, jetpacks, aliens, and other Easter eggs, this one was ahead of its time. Via direct message, current Rare Witch Project site editor Andre Brown calls the Stop ’n’ Swop saga a “precursor to a lot of the [alternate reality game] stuff that’s so common in game marketing nowadays — except this was an accident.”

Not every of-its-era aspect of Banjo-Kazooie seems quite so quaint today. For one thing, the game’s writing is off-putting at times to 2018 eyes: Not only does the story hinge on rescuing a helpless damsel in distress, in traditionally regressive Mario fashion, but it often fixates on its female characters’ outward appearance. Gruntilda’s greatest sin, it seems, is being smelly, ugly, and overweight, and her sole motivation for kidnapping Tooty is to steal her looks. Given the frequency with which Banjo body-shames her, it’s hard to blame her for wanting to be better-looking. In the game-over video that appears when the player runs out of lives, Tooty packs on pounds while Gruntilda emerges from her transmogrification machine in a miniskirt, stick-thin but as buxom as Lara Croft. Upon seeing the newly slender and conventionally attractive witch, a smitten Mumbo Jumbo (whom the insensitive Kazooie calls “skirt boy” and “mask midget”) instantly hands her a flower. It’s not the sort of message that one would want to imprint on impressionable gamers today, but two decades ago, games were less often designed with women in mind.

Banjo’s gameplay also seems dated in certain respects. In 2002, Microsoft purchased Rare, making the company’s games exclusive to Xbox consoles. Six years later, piggybacking on the release of Banjo-Tooie’s divisive, vehicle-based sequel, Banjo-Kazooie: Nuts & Bolts, a remastered Banjo-Kazooie joined the Xbox 360’s downloadable lineup (later to appear in the 2015 Rare Replay compilation on Xbox One). Even in 2008, when Banjo-Kazooie was only half as old as it is today, writers regarded it as a relic. IGN’s largely positive review of the remaster warned that “many people will find this style of gameplay outdated by modern standards” and concluded, “There’s a reason developers stopped making platformers like this one: Most gamers have had their fill of simple fetch and hop games and like a bit more action and drama on their consoles.”

The graph above showed the extent to which platformers fell off the table in the mid-’90s, but the graph below, based on data from the Internet Game Database and limited to releases on Nintendo, Sega, Sony, or Microsoft systems with at least five user ratings — a proxy for prominent games on non-handheld hardware — reveals that not long after Banjo, the console platformer sank into a second, deeper trough that it didn’t emerge from until 2012.

That additional decline may have stemmed in part from practices that Banjo’s success helped cement. Banjo could be frustrating: It forced players to collect the 100 hidden notes in each level in one life, wiping away their progress if they ran out of health. No one knows this pain more intimately than the dedicated players who, in a testament to Banjo’s replayability, difficulty, and depth, still speedrun the game, using optimized routes and game-changing glitches to shave hours off the time it takes to complete it. Although the average Banjo playthrough takes roughly 14 hours, speedrunners can complete the game in a small fraction of that time.

Last month, Stephen “Stivitybobo” Domke, a New Jersey speedrunner, made a record-setting run that lowered the bar for a “100 percent” run — one in which the player collects every available item — to 2:01:03. A South Australian speedrunner named Jake, who goes by “The8bitbeast,” holds the record for an “any percent” run at 1:11:39, although he says via direct message that he’s holding out hope of finding “the holy grail of BK glitch-hunting,” which would allow him to bypass the game’s final note door and take that time down to roughly 40 minutes. “I think the game challenges patience the most out of anything,” DMs Domke, who estimates that he finished the game more than 50 times even before he started speedrunning it several years ago. “This is because it is one of the most punishing speedruns out there.”

Banjo is pretty punishing for regular players, too, but Rare could get away with making players retrace their steps because late-’90s gamers had fewer quality titles to choose from. “Back then, if people bought a game, they were really, really committed to it,” Mayles says. Now, he notes, players are “just absolutely spoiled for choice,” which, Kaerlev says, means that modern gamers have “much less tolerance for time-wasting.” Although Banjo-Tooie relented and let players preserve their progress upon leaving a level, Mayles says he still believes in Banjo-Kazooie’s less forgiving approach. Even so, he acknowledges, “I certainly wouldn’t want to try those levels these days. … I think I’d end up smashing my controller very quickly.”

Subsequent platformers walked back Banjo’s difficulty, but they doubled down on its “collect-a-thon” tendencies. Rare’s Donkey Kong 64, which arrived the year after Banjo, made Guinness World Records for offering the most collectible items in a video game (3,821) — at least one example of which was well-hidden enough to elude detection for 17 years — and Banjo-Tooie also subscribed to a self-defeating “more is better” philosophy that satisfied compulsive completists but turned off everyone else. “I did not enjoy that game at all,” Kaerlev says. “It’s very different from Banjo-Kazooie in that it tries to be so much bigger, and as a result everything gets kind of diluted.”

Kaerlev continues that at Gears for Breakfast, “Our rule is that if something has genuine purpose, players will not feel like it’s a waste of time.” A worthwhile collectible, he says, “needs to provide the player with something that advances them beyond just being a gate, because gates are arbitrary, and players can see through the fact that a gate is arbitrary.” Banjo sometimes adheres to that rule: A few of its items — including eggs, feathers, honeycomb pieces, and Mumbo tokens — grant power-ups and abilities, although the notes and puzzle pieces simply provide players with quotas to fill.

Gregg Mayles

Mayles blames the platformer’s long malaise on the genre’s lack of evolution. “You’re still doing pretty much the same thing you were doing 20, 25 years ago,” he says. “I think there is still room to do something different, but I’m just not seeing it at the moment.” However, Kaerlev pins the platformer recession on cyclical market trends more than tired design decisions. “I don’t really think anything in particular happened,” Kaerlev says. “It’s just, people felt that everything had been explored in that genre and wanted to move on. But now it’s made a return, it’s become a trend again. It’s a chance to go and revisit that genre and look at the old games and see if there’s anything that can be improved for a modern audience.”

Kaerlev, 27, has made that his mission: He was weaned on games like Banjo-Kazooie, and he’s trying to bring them back. The data shows that developers are making more platformers as a percentage of total games than they have since the early ’90s, in part, perhaps, because they work well on mobile platforms, hit the nostalgia sweet spot of today’s typical gamer (average age: 34), and are simple enough to program that one person or a small company can create them on an indie developer’s budget. The audience demand is out there: A Hat in Time, which billed itself as a spiritual successor to late-’90s Nintendo and Rare (“Think Banjo-Kazooie, The Legend of Zelda, and Super Mario 64 all in one!” the game’s Kickstarter description said), exceeded its crowdfunding goal by nearly 1,000 percent.

A Hat in Time features two early sequences explicitly based on Banjo: one in which the game’s young female protagonist, Hat Kid, collects scattered tickets, and one in which she races another character. Neither one was a hit. “Overall, the feedback that we got was anything that remotely resembled Banjo-Kazooie was not very much liked,” Kaerlev says. “So I thought, ‘Huh. Well, OK, I guess that kind of game does not work today.’” Kaerlev made the tickets less painful to find and concentrated on keeping the positives he remembered about Banjo — charm, puzzles, freedom, clever levels, a hub world that deepened the connection to the characters — while adding variety and stripping out some of the unmotivated collecting and the often-clunky camera (a casualty of an overstretched team) that Mayles calls “OK for a first attempt, but probably no better.” A Hat in Time harnesses modern processing power to employ a CPU-pushing “algorithmic camera” that never would have worked on the N64. “For every frame in the game, we calculate the camera like 100 times and find the best position,” Kaerlev says. It’s a 2018 solution to a 1998 problem that never went away.

A Hat in Time generated strong reviews and, despite being made and marketed by a small studio and coming out the same month (and featuring the same fashion accessory) as Mario, has sold a third of a million copies across all platforms, validating the idea that there’s an appetite for games inspired by the best parts of Banjo — or even, perhaps, starring Banjo himself. With other late-’90s platforming figures such as Spyro and Crash back in action and a new Battletoads slated for 2019, rumors and requests still swirl around the idea of a Banjo revival, and Mayles hears regularly from “aggrieved people on Twitter” who want to know when Rare will bring back the bear and the bird. (Don’t @ him.)

Twitter mentions aside, Banjo is never far from Mayles’s mind, because it taught him “how to imagine game worlds in 3-D inside my head.” Although his latest game, the shared-world pirate adventure Sea of Thieves, evinces a recognizably Rare sense of humor, it also, he says, reflects a shift in philosophy toward “giv[ing] the player tools in order to create their own fun, rather than Banjo’s approach of us curating the fun.” To justify a sequel, Banjo would have to prove flexible too.

“If we were to look at Banjo again, I think I’d take what players love about the traditional Banjo … but I’d still be keen to do something different with it,” Mayles says, adding, “another Banjo game in the exact vein of what we did 20 years ago … wouldn’t challenge me as a designer.” Judging by Banjo-Kazooie, it probably would challenge players. But even if Banjo himself is history, the legacy of his early adventures remains an indispensable part of platforming’s present and future — and that future appears promising for the first time in years.

Thanks to Arthur Rudolph, Alex Dancho, Randy Gleason, Simon Carless, and Jorgo Stavroudis for research assistance.