Being the follow-up to a masterpiece is no small order. Being the follow-up to two masterpieces is a Sisyphean task. Ico is a masterpiece of its time. Fumito Ueda and SIE Japan managed to follow that game with Shadow of the Colossus - a masterpiece of even higher order. The bar was set tremendously high for the team's third project: The Last Guardian. Multiple delays, a change in platform from PS3 to PS4, and Fumito Ueda's departure from Sony squashed a lot of the hype for the game. Might the game turn into vaporware? Or might it release in a condition analogous to Metal Gear Solid V?
On the surface, The Last Guardian comes off as being a mash-up of Ico and Shadow of the Colossus. Superficially, it's much more in-line with Ico: you play as a small boy who must guide a companion through a maze of environmental platformer obstacles and adventure puzzles. The catch this time around is that the companion happens to be a giant animal that you can climb and ride on.
The Last Guardian share more with Ico, but your companion is a giant creature that you climb and ride on.
The big difference though, is that The Last Guardian is sort of an inversion of the gameplay of Ico. In Ico, the player character had to guide a helpless (some even speculated she is blind) princess through a castle and defend her from shadow monsters that try to drag her away. In The Last Guardian, however, it is the player character - the boy - who is mostly helpless. True, you have most of the agency and are guiding Trico through the maze. But Trico is the one with all the power, and your progress is often dependent on Trico getting you past obstacles.
This point is most hammered home by the game's combat mechanics - or rather, its almost complete lack thereof. The boy can't fight off the stone knights that hunt him down. You can only run away, or let Trico smash them into dust for you. If they catch you, they drag you off to a nearby mysterious blue doorway (a parallel to the smokey portals that the smoke monsters dragged Yorda through in Ico), and all you can do is mash buttons to kick and squirm. Ico and Shadow of the Colossus experimented with player agency by making the player question the motivations of the character and wonder if maybe you're doing more harm than good. The Last Guardian toys with agency in other ways. In this game, you, the player, are the helpless tag-along character in an escort quest. You get a glimpse through the eyes of Yorda from Ico or Ashley from Resident Evil 4.
The boy can't fight back, he can only kick and squirm - much like Yorda from Ico.
Not entirely though. The player and Trico make mutual contributions to progress, and their contributions are shared much more than Ico and Yorda. Much like how Yorda could occasionally open the magically-locked doors, the boy in Guardian also has to pull levers and open doors for Trico to pass from room to room. The boy also has to destroy glass eye murals that mesmerize and terrify Trico to the point of paralysis. The boy also hunts down barrels of [supposedly] food for Trico to eat whenever Trico is tired or wounded. But then there's also parts of the game in which the boy simply hops on Trico's back, and Trico leaps away to the next puzzle area without the player having to do anything.
Going nowhere fast
Probably the first problem that I started to notice with The Last Guardian is its length and pacing. The game just doesn't really seem to go anywhere, and it takes its own sweet time getting there. Pacing and length are a major issue in puzzle games, and the best puzzle games tend to find a perfect balance of length, variety of content, and progress. Ico and Portal, for example, work so well - in part - because they are short. They are very tight, concise games. My first playthrough of Ico and Portal was roughly 5 hours each, but The Last Guardian dragged on for roughly 12. Most of the dead weight seems to come in the first half of the game, as the second half finally started to provide a drip feed of intrigue to keep me interested.
Ico had a clear objective: get out of the castle and take Yorda with you. The opening cutscene establishes that you were brought to the castle against your will (for some kind of ritual or sacrifice) and gives you an idea of the space you'll have to traverse and the challenges you'll have to overcome. As you progress, the large open areas will give you a view of areas to come, and as you reach those foreshadowed areas there's a definite sense of progress; you're getting somewhere! There's also early cutscenes that keep the narrative going at a steady pace. Of course, the game throws a wrench into everything when you get to the front gates, but that's fine.
Ico and Portal provide clear delineations of progress and accomplishments.
Portal on the other hand, provides the implicit objective of completing all the puzzle challanges for GladOS. The game even posts signs telling you exactly how many puzzle rooms there are, but it becomes obvious pretty quickly that GladOS isn't going to let you just walk out when you're done. So you're not quite sure where you're going, but the game makes up for it by being divided up into clearly delineated puzzle rooms. Each puzzle room either adds a new mechanic or requires the player to combine previously-learned mechanics in new and novel ways, providing a constant sense of progress and learning. And when the room-16 twist happens, and the game goes "off the rails", it's actually a pleasant surprise! Portal 2 felt weaker than the first game [to me] because it seemed to go on for too long and had some gimmicky, repetitive puzzle designs.
Behold: the first new mechanic in hours!
The Last Guardian doesn't have any of the strengths of Ico or Portal. It's implied that the boy is trying to escape the castle, but there's no indication as to where the exit is or how escape would be accomplished. He has no idea where he is, how he got here, or why he's here, nor does he know where the Trico came from or why it seems to like him so much. You progress through the castle, therefore, feels aimless and meandering.
The game also introduced practically all of its mechanics in the first half an hour or so, and then even takes a major mechanic away for no real reason and with little justification (only to give it back to you nine hours later with equally little justification). You jump and climb over obstacles, pull levers to open doors, push blocks, and climb onto Trico to let him take you to the next challenge - all while avoiding capture by the stone knights. I was playing for hours doing those same repetitive tasks before the game introduced anything even remotely new: knights with glass eye shields that the boy has to push or shove in order to disarm them or knock them off of ledges.
Some paths are literally marked out for you.
Reading the signs
The other major contributor to the game's pacing feeling wrong is that the puzzles alternate between stupidly-obvious sign-posting that takes all the challenge away, to obtuse or esoteric sign-posting, and sometimes even to outright red herrings. The extent of solving a puzzle is usually just figuring out where the game wants you to go. So if you find some differently-textured line of bricks sticking out of a wall, then it means "climb here". But if the path ahead isn't as obvious, you can spend large chunks of time just wandering around in circles.
And some of the areas are quite large, and wandering in circles looking for a clue you missed can be quite tedious. The size and openness of the areas isn't really used for creative problem-solving or to allow player-driven exploration of them. They're all very simple, linear designs requiring little thought, but a lot of patience. There's one bit near a minecart with a chain that goes up to a ledge that is a dead end. I spent about half an hour climbing and re-climbing that chain trying to find something that I could climb onto because I thought that was the path out. In another area, I spent a good 10 or 20 minutes trying to figure out how to get Trico to dive under water while I'm grabbing its back.
The character can tip toe or run. No in between. He can't simply walk...
There's no real mechanical challenge to solving any of the puzzles. There's no grip gauge to limit how long you can climb or hold onto Trico, and combat is literally just mashing buttons. So the only real ludic challenge comes from having to fight with the game's camera and controls. Movement is fidgety and imprecise. If you nudge the stick slightly, there's an absurdly slow tip-toe movement, despite there also being a crouching slow-walk. This implies [to me] that two different developers on the team thought that this was maybe supposed to be a stealth game, and they both implemented different movement methods that both somehow ended up in the game, despite the game requiring neither. If you push the stick all the way, then the boy does a full run. Oh, there's also a silly-looking run-in-place animation whenever you're giving movement commands to Trico or when there's enemies around. There's no casual walk speed.
Cat owners know the feeling...
Controling Trico isn't much better. The animal sometimes just doesn't want to cooperate. I can't count the number of times that I gave a command to Trico, only to have the beast quizzically stare at me. This often leads me to wonder if maybe Trico isn't supposed to go that way, and I'm supposed to find another route or solution. But no; it's usually the right way, the animal is just stubborn. If the delays were to improve the beast's A.I., then this game could have used a few more months of polish. On the other hand, if the intent was create an A.I. for a split-brain cat that occasionally thinks it's a dog, then the dev team deserves a bloody award.
It is telling that Trico doesn't become more cooperative or smarter as the game progresses. In fact, I had even more trouble getting the animal to do what I wanted it to do in the later stages. So I doubt that its failures to follow orders are a deliberate element of the A.I. design. If the A.I. were deliberately finicky, then I would expect control over Trico to become more reliable as the game progresses, and their bond and rapport develops. But that isn't the case.
Trico routinely fails to follow instructions, leaving me to wonder if I'm trying to do the right thing.
It's also frustrating that there is no easy way to dismount Trico. Hypothetically, the X button is supposed to let go of stuff and drop down. But since there the boy automatically grabs onto things, he ends up just letting go and then grabbing onto the next clump of feathers or fur. So you can mash the X button until the boy eventually just faceplants into the ground, or you can leap off the beast and hope you don't break your leg or fall into a bottomless pit.
The controls get even worse as the game nears the end. The Circle button is already overloaded to serve as the generic "grab" and "interact" button for when you want to pull a lever, push/pull a box, or pick up an object in the game world. But it's also overloaded to pet Trico, and to remove spears from him when he's attacked. Then the geniuses [sarcasm] on the development team thought it a brilliant idea to also map the magic shield to the Circle button, and then force you to use it while standing on a bunch of narrow walkways with Trico standing right next to you. So it's a crap shoot whether the boy aims the shield or grabs onto Trico's fur, or pets Trico, or picks up a dead enemy's head off the ground.
Shooting lightning from Trico's tail is a silly mechanic to begin with, but it also controls very poorly.
The whole "shooting lasers from Trico's tail" is a silly mechanic to begin with (which is why they have to take it away from you for 90% of the game), but it's made absolutely intolerable by the developers' abysmal control decisions. It's practically impossible to use when you're mounted on Trico, as he'll move and squirm and usually cause you to either lose the thing you're targeting, or the boy will stop targeting altogether and grab onto Trico. Also, does the game ever give any reason as to why the tail shoots lightning at the beginning, but shoots laser pulses at the end?
Why did you need to overload so many functions onto one button?! There are twelve buttons on this controller, plus clicking the analog sticks, plus the trackpad! The triggers and touchpad aren't even used for Pete's sake, and you thought it necessary to pile so many conflicting functions onto the circle button...?!
Doesn't hold a candle to its pedigree
I've also heard reports that the game has numerous stability and performance issues. I only suffered one crash and a few noticeable instances of framerate loss. But then again, I didn't start playing the game until a month after its release, so I may have benefited from the game being improved by stability patches.
Regardless, perfect performance and stability would not save The Last Guardian. With just a few seemingly common-sense changes to the controls (like using a trigger button to aim the shield and having a dedicated "climb" button), and with a better sense of direction and purpose at the beginning, The Last Guardian could have been a very good game, and Trico would be a memorable and beloved video game mascot. Instead, the game is mediocre, and whether you even like Trico by the emotionally-traumatic conclusion will depend entirely on how frustrated you are when the game finally ends.