I’ve done my best to ensure this article contains no spoilers for The Last Guardian, be they story spoilers or specific puzzle spoilers. I do discuss some of the central game mechanics, and very broadly allude to the content of some of the puzzles, but I don’t believe there’s anything here that impede your ability to go into the game completely fresh if you haven’t played it already.

I wonder if Fumito Ueda plays videogames. One would assume that as a games designer he must, but if so he’s very good at hiding his influences; no one taking inspiration from anything released in the last 10 years would go on to develop something like The Last Guardian. Some critics have accused the game of being old-fashioned, but this isn’t really true; it’s not that Ueda’s third game doesn’t feel modern. It doesn’t feel like any other game at all; excepting, of course, Ueda’s other projects, 2002’s Ico and 2005’s Shadow of the Colossus, from which Guardian draws many mechanical, thematic and aesthetic cues.

Modern game design is full of often unspoken rules, usually developed by consensus between developers over many different projects. This has become increasingly true in the last ten or so years, as the rising cost of game development has seen studios more inclined to stick to tried-and-true design tools. There is now, for example, something resembling a standardised model for designing an open world and populating it with content. There’s a similar standard for multiplayer shooter design; an established set of progression systems that players have come to expect from these sorts of games. While an over-reliance on established models can lead to fatigue and frustration, these design trends have generally developed because they work, and if a designer is going to deviate from an established format, there needs to be a compelling reason for them to do this.

The Last Guardian disregards these hazily established design rules in numerous ways. It uses triangle as its jump button, for starters, when everyone knows by now that X is the correct jump button for a PlayStation controller. More damningly, it eschews a standardised climbing mechanic honed over numerous Assassin’s Creeds, Uncharteds and Tomb Raiders for a finicky physics-based solution that’s liable to send players plunging towards an unearned death at least once. There are numerous modern player experience standards the game simply ignores, such as the idea that a player should be able to achieve basic actions with a minimal number of button presses; it sometimes feels like it takes about 15 inputs to get The Last Guardian’s child protagonist to go down a ladder (this an exaggeration, sure, but the kid does not like descending anything in this game).

Trico is super nice though. Here he is having a sleep.

Perhaps this is what people mean when they call Guardian old fashioned; not that it feels like an old game as such, but rather that its design philosophy seems to harken back to a time when designers built mechanics that supported the type of game they wanted to make, rather than assembling a game from a predefined box of established mechanics. This is a grossly oversimplified view of modern AAA game design, sure, but it certainly feels like a lot of modern games are built by mashing together bits of other games. We’ve got working models for ‘third person shooting’ and ‘open-world city design’ that come up again and again in different games. That’s not to say that these games lack vision, but it’s difficult to argue that a standard ‘over-the-shoulder’ third person cover system was built from the ground up to specifically suit whichever of the twenty-or-so games that will have used it in the last few years.

Ueda’s games, in contrast, feel considered, as if every mechanic has been designed to specifically reinforce a theme. Mark Brown of Game Designer’s Toolkit goes into some detail on how The Last Guardian‘s carefully considered mechanics cement the relationship between boy and beast that is at the heart of the game. It’s well worth checking out his video; there aren’t any story spoilers, but there is discussion of core puzzles and mechanics as well as reveals of certain key moments that you might want to avoid if you want to go into the game completely unspoiled, which I would recommend doing. If you’ve already played the game or just aren’t bothered, there’s some great insight into how considered design can reinforce a game’s core theme.

Following on from this, it was surprising to learn that Ueda doesn’t actually consider theme that important when developing games. For someone who has developed three games now that many people hold up as the gold standards for portraying relationships through gameplay, he’s remarkably unattached to what people consider to be the core themes of this games.  From an interview with Keith Stewart for (appropriately) the Guardian:

“From our side, we didn’t have a strong intent to portray relationships as a theme of our games. We actually tried not to do that. But as a lot of people played through Ico and Shadow of Colosuss, they said, ‘Oh, these games are about building a bond, they’re about trust.’ From my perspective, I just say, here, go ahead, play it – then at the end, the player can come to their own conclusion about what the game is trying to tell them. I don’t like to force feed themes.”

What Ueda’s designs convey, then, if its not a specific theme or a specific relationship, is physicality. No games communicate the physicality of one conscious body acting upon another the way Ueda’s games do. In Ico it’s the way Yorda stands passively as Ico tugs her hand, only to be dragged forward into a run when her arm reaches it’s full extension. In Shadow of the Colossus it’s the way Wander can be battered like a rag doll against the side of a creature the size of a building, only barely able to maintain his grip until the beast stops shaking. In The Last Guardian it’s the sheer difference in size and power between Trico and the boy, how Trico can leap effortlessly up a wall it might take the boy a full minute to climb, how the force of Trico alighting on a platform next to him can bowl the boy off his feet.

It’s the believability of these physical interactions that make the relationships between these characters feel so meaningful, and the fact that they are written into the mechanics of the game make this something that couldn’t be delivered in a different medium. Detractors of The Last Guardian lament the limb-wheeling uncertainty with which the boy moves, or the way Trico will ignore an instruction to sniff quizzically at the air, but it’s these consistencies of embodiment that make these characters so believable, and make the unique chemistry of the game work.

Much of this physicality is delivered through animation, which makes sense when you consider another quote from Ueda’s interview in the Guardian:

“In general, the games that are appealing to me are the ones where the character animations are very intricate and beautiful. It’s not just about the asset itself, it’s how I perceive the character…”

This indicates that Ueda does play games, at least, even if what he takes away from them isn’t quite the same as the majority of designers. There’s a term used in studies of the social psychology of games called ‘relatedness’, which describes how a player will treat an on-screen avatar much as they would a real living creature providing there’s a suitable level of believability. Research, including my own, has looked at how this believability can be achieved through effective use of animation, and it’s this that makes Ueda’s characters feel so real.

The Last Guardian features some of the best creature animation I’ve ever seen, in games or otherwise. Anyone familiar with animal behaviour will look at Trico and instinctively understand it as an animal. This is the real achievement of The Last Guardian; anyone familiar with animating for games will have an idea of how difficult achieving this must have been. Seeing how flawlessly Trico interacts with the game world, Guardian’s notorious 7-year development hell starts to make more sense. This isn’t something a sensible developer would attempt; centring an entire game’s design on the believability of a single virtual agent is a ridiculous design goal.

Come down the stairs Trico, I’ll find a nice barrel for you. Trico loves barrels.

Ueda acknowledges the challenge in another interview for Eurogamer:

“AI, working with AI, it’s not easy. But it’s more doable than most people might think. It’s the process that happens after, when the AI decides to do something. That’s incredibly difficult. Trying to get that motion right after the AI’s decided to do this or that, trying to get it look as natural as possible, that was something that was painstakingly hard.”

The animation approach used in Guardian is one that Ueda’s been using since Ico, and one that’s now popular in the AAA game space. Animation layering and blending is a solution that allows pre-built or pre-captured animations to be effected by the physics of the game world. It’s what allows Trico to adjust his ‘eating a barrel’ animation depending on the location of the barrel in front of him, or have his turning animation change depending on the size of the ledge he’s standing on so none of his limbs cross into empty air.

This breakdown from Naughty Dog shows their approach to developing animation for the first Uncharted game; the final 10 or so slides address animation layering and how that allows Drake to move realistically and consistently through a complex game world.

Ueda’s games have always used similar systems. In Ico, the speed and weighting of Yorda’s run is governed by the speed at which the player, as Ico, pulls her. Wanda’s climbing animations in Shadow of the Colossus adjust dynamically to the shifting terrain of the colossus beneath him, seeking out natural handholds and shifting his weight appropriately. The same is true of the boy in The Last Guardian, whose weight is constantly shifted by the much greater bulk of Trico.

Where Guardian differs from something like Uncharted is in the priority it gives animation over player utility. In Uncharted, player intention gets priority over animation and believability; if the player attempts to make an achievable jump, then Drake will make the jump, even if it involves breaking the animation and in-game physics slightly. Animations will stretch to allow Drake to cover a slightly greater distance than he might realistically be expected to, or quickly crunched into another animation to ensure a smooth landing on a ledge. The result feels floaty, faintly unrealistic, but it ensures the player is always able to get where they need to go with minimal frustration. This approach has been widely adopted by many modern platform games, from Assassin’s Creed to the modern Tomb Raiders.

In The Last Guardian, however, the game’s physics are sacrosanct. If the boy attempts a jump he’ll reach towards an oncoming ledge just as Drake would, but if his hand doesn’t connect with the edge there’s no auto-correction; he misses the jump and plummets downward. It means that if he falls off Trico, he’ll land in a rag doll sprawl of limbs rather than ‘correcting’ out of his fall into a smooth landing animation.

This approach gives a much greater weight and believability to the game’s characters. As someone who’s maybe a bit too into fancy animation, it was a joy to steer the boy into stumbling down stairs and ledges, or having him tumble off Trico to appreciate the organic way he was jounced across the floor.

The effect on the game’s design, however, is a double edged sword. Where the game leans into the physical relationship between boy, creature and world, it works gloriously. Puzzles that revolve around making Trico engage physically with the game world, which make up a majority of the game, are a joy. Several brought a rare sort of smile to my face; without wishing to spoil any of it, one section involving the effect of a large creature on a body of water and another based on the physics of a see-saw were particular delights. This kind of direct physicality is something rarely see outside of Ueda’s games, because it requires such dedicated attention to animation and physicality.

Here Trico is being nice and helpful and saving me from death.

Where the game falls back on more traditional game-like mechanics, however, this approach feels cumbersome. Object physics puzzles, whether you’re rolling a large cage or attempting to transport a barrel across a series of ledges, feel finicky as overly-detailed physics mar player intention, objects tottering over edges or getting diverted by tiny blemishes in the environment. Straight-up platforming sections, which lamentably increase towards the end of the game, are worse; the physical animation system that makes climbing Trico such a joy is an absolute liability when it comes to making a jump between one flat surface and another, with the boy regularly missing a hand-hold for no visible reason and plummeting into space.

It’s perhaps ironic that The Last Guardian is a less enjoyable videogame when it’s trying its hardest to be a videogame. These platform sections and barrel puzzles add little to the overall experience. They seem to be there as a concession to videogame design; this is a videogame, and to prove it here are some videogame things to do.

Given this the game’s mixed response is understandable. People who appreciate the affordances of modern game design to seamless player experience are justified in their frustration at Guardian’s deference to its own physics and animation. A few sections of the game trend towards teeth-pulling frustration. However, without such a robust simulation the game simply wouldn’t work. The entire experience hinges on Trico’s believability, and Trico’s believability hinges on his convincing physicality in relation to the boy and the game world.

Fumito Ueda doesn’t design games like anyone else. Most designers would have compromised somewhere along the development cycle, and we would have ended up with a game that would be perhaps more playable but less engaging for it. I for one consider The Last Guardian a triumph. As a player, it offers an experience with an emotional weight that more conventionally designed games can’t match. As a designer, it offers an insight into a design process that shows little regard for defined forms and theories, and in an industry of increasingly homogenous designs, that refreshing to see.

Nice Trico.

Further reading:

The Last Guardian Fumito Ueda Interview – The Guardian.

“Never in my wildest dreams did I think The Last Guardian would take this long” – Eurogamer.

Fumito Ueda Interviews – Shmupulations.

Uncharted Animation Workflow – Naughty Dog.