Half-Life 2 gave me one mesmerizing playthrough, but every time I return to it I feel like I could be doing something else. There are impressive setpieces and environments, but the world overall feels much less coherent than that of the first Half-Life. In that game, the setting was as much of an antagonist as the game's enemies, throwing all kinds of environmental hazards at the player, forcing them to contort their bodies to fit into tricky places, etc. This kind of interesting movement has largely been replaced by puzzles which show off Source engine tech, but are uninteresting in their own right. AI and enemy variety are clearly worse here as well—not that they were necessarily exceptional in the first game, but it made more effective use of what it had. But the worst change Half-Life 2 brings is its particular emphasis on exposition and dialogue. Half-Life 1's technique of placing talkative NPCs alongside a silent protagonist worked well because Gordon Freeman's characterization is very abstract: he is defined by his immediate situation and little else. Half-Life 2 gives him a backstory, relationships, attempts to shoehorn him into a stupid and uninteresting mythology: in short, it puts us the awkward position of being party to social interactions in which our silence is extraordinarily unnatural and off-putting. I'm far from the biggest fan of the videogame-spliced-with-prestige-TV-serial trend that Naughty Dog helped popularize, but there's a reason no one does it like Half-Life 2 anymore and that's because it sucks.

"The greatest indie metroidvania"—so here's the 8/10 it deserves.

The main thing Hollow Knight has going for it is of course its massive size and true nonlinearity. When I was deep in the process of exploring, finding secrets, getting lost, finding my way again, seeing how different regions link up—these were the times when I thought I was playing one of my favorite games. The charm system is pretty cool as it is, but it along with all the other goodies in the game gives plenty enough reason to explore every nook and cranny (and overlook some pretty bad Super Meat Boy-esque spike maze level design along the way). There's so much to admire in how distinct each region feels and how elegantly they all fit together, and the amount of freedom you're given after a certain point is something few games arrive at even when they strive for it.

But.

The issue is that Hollow Knight's very scale tends to emphasize one of its genre's inherent limitations. Backtracking is built into the genre's DNA—and this is no bad thing, in fact it's kind of the whole point: revisiting areas with new tools and with a new purpose develops a much more organic relation to space than is usually the case in more linear games. The tradeoff is repetition: not only is the space itself usually unchanged, but obstacles are usually static as well. In a game this big and this long, in which backtracking is therefore not only mandatory but extremely time-consuming, seeing the exact same enemies respawn in the exact same places over and over again starts to feel like a drag—much moreso in a 40-ish hour game than in a 10-ish one like Super Metroid. Moreover, Hollow Knight, like Super Metroid, wants you to be immersed in a massive and seemingly "alive" world, but the world and its inhabitants are essentially static, existing only as a backdrop to the player's "hero's journey." Having no internal dynamics or logic of their own, they change, if at all, only in response to the player's actions, and then only rarely. (I was almost ecstatic when I first got to the Infected Crossroads, and then massively disappointed when I realized that other areas of the game would not get the same treatment.) Size aside, the difference between the two games is that only one of them came out in the same year as Rain World: a game that shows us how the player-game relationship can be reimagined in the context of a 2D platformer with a massive persistent world, and how decentering the player character can result in a more rewarding and immersive experience.

There are other aspects of Hollow Knight I'm ambivalent about. The "soulslike" elements are a decidedly mixed bag, and despite the variety within the player's toolkit, the core moveset feels kind of generic. Samus may have fewer abilities at her disposal, but the ones she does have define her as a character to a far greater extent than... (wait, what's the little guy's name again?). I also have to admit that I balked at the endgame boss rush, but that's more because I took a long break between getting the three thingamugugs and facing the Titledrop Monster, and completely lost any momentum in between. I may return to Hollow Knight—some years down the line—but if I do it'll be from the very beginning.

As good as I was expecting given the seven-year development time and the pedigree of the development team (a virtual who's who of the current crop of Thief fan mission authors), which is just insanely good.

The campaign continues my favorite trends in Thief custom level design (high levels of verticality; rehabilitation of the less... popular themes that characterize Thief: The Dark Project levels, namely supernatural horror) while bucking the older habits endemic to fan missions that target the "hardcore" set (obtuse riddles and key item hunts). Thankfully.

Before this I'd have pointed to the 20th Anniversary contests as the ideal entry point to what is, in my opinion, one of the single greatest PC game modding scenes of all time. Now I'd point to The Black Parade for the second and fourth missions alone (respectively the best Constantine's mansion and Bonehoard-style missions out there, and I've played all the good ones).

Bust out your black hoodies and diegetic compass/map combos, folks. This is IT.

This is not a review so much as a brief commentary or analysis of a certain aspect of Yume Nikki I find interesting. I have no intention of providing the (n+1)th narrative interpretation of the game's imagery etc.

--

A large subset of "art games" (a term I dislike, but I won't go into that here) turn on the central act of moving through space. While derisive commentary on such games generally draws attention to the lack of interactivity between the player and environment, I find it more interesting and helpful to think about the ways in which these games allow movement to be expressive, or how they fail to do so.

In a certain sense many, if not most, games already have movement at their core, and arguably handle it more profoundly than many art games by complicating and problematizing the nature of movement itself. One trick is to split movement into distinct actions (jumping, walking, dashing, etc.) which are required in different combinations at different times; another is to confront the player with AI agents ("enemies") that impede the player's movements unless they are avoided or overcome. If this is obvious, the point is that these complexities make the act of movement itself expressive in ways that are easy to forget when everything is viewed abstractly in terms of genre conventions.

The question is, when such things (whether we insist on calling them "gameplay" or not) are subtracted, what is left to be expressive? Many authors will rely on the strength of their writing or art, and this approach can be fruitful, but it sidesteps the question I'm interested in. Yume Nikki answers this challenge with the very structure of space itself.

Partly it does so by drawing on the conventions of a kind of game that predates digital games, i.e. the maze. Even putting aside its scale and interconnectedness (and the occasional gate puzzle), Yume Nikki is a particularly forbidding maze in that it has no end, at least in the usual sense. The "effects" scattered throughout its interior (some of which aid traversal while others are purely cosmetic) unlock an ending, true, but many of Yume Nikki's depths contain no effects at all. The game's world is many times larger than it needs to be to accommodate the effect-finding game. Certainly, each "unnecessary" area serves its own narrative or affective purpose, but in terms of my argument they all serve just one: they are places to be lost in.

The other, complimentary aspect of Yume Nikki's approach is its austerity. Yume Nikki has no concern for making a first impression of variety or density; looking at the areas immediately accessible from the Nexus, the impression is rather one of emptiness. Partly it's the repeating tiles, a technical convention of RPGMaker and the games it draws from, but also the scarcity of features, interactive or otherwise, that would distract the eyes or hands in the way we normally expect.

The upshot is that Yume Nikki requires two things of its players: firstly, not to become bored (something of a dare in all honesty), and secondly to be quite deliberate and systematic in their approach to exploration. All this results in a level of serious engagement the typical "walking sim" does not require, and in the long run that other aspect—Yume Nikki's art and aesthetics—only gains from this.

Notes:

1. One could reasonably argue that much of what I have said about Yume Nikki's structure derives from its unfinished state. While I doubt an updated version would drastically alter this structure by filling in currently "empty" areas, I'll concede that I have nothing to base this on besides intuition, and that other additions or alternations could change the core in unforeseen ways. (That is, of course, assuming the project has not been abandoned.)

2. I will also concede that certain secrets are definitely /too/ well-hidden.

In designing System Shock and Thief: The Dark Project, Looking Glass Studios aimed to achieve a "role-playing" experience that was quite different from videogames that drew their mechanics directly from tabletop RPGs. A key insight was that much of the arithmetic that in a tabletop environment might burden players and DMs alike could, in a digital environment, be handled more or less exclusively by the computer. A certain minimum of numerical awareness on the player's part is still necessary (health, ammo, etc), but there is simply no need for the usual RPG rube goldberg machine of having to do things to cause some numbers to go up to get some other numbers to go up in order to get the numbers you actually want to go up to go up.... Far more elegant approaches to player progression are now possible, and what Looking Glass achieved with System Shock—progression based on collecting equipment and suit upgrades—converges onto territory analogous to what Jeremy Parish charts in the transition from the console action-RPG to the metroidvania.

True, SS2's leveling system (by another name) is not exactly the sadly familiar contraption described above, but there is still a diegetically nonsensical experience currency that muddies the intuition driving the equipment-based elegance of SS1. True, grinding is not possible, but like all RPGs even with grinding, the whole system progressively locks the player out of options as the game advances, which by a sleight of hand is disguised as "choice." It is expressive enough that I may choose which weapons to fill my limited inventory space with, attending to the limitations imposed by the available resources (ammo, energy, etc), without forcing me to put points into a skill tree to use them effectively, or at all.

I'm admittedly being perverse in framing the benefit of such systems as a disadvantage when it's really a tradeoff: these systems allow for distinct paths of player character growth which, to be meaningful, must be mutually exclusive. That's nice. I prefer it the other way. ¯\(ツ)/¯

While I wouldn't call Yoshi's Island one of my favorite platformers, I certainly wouldn't bat an eye at the many who do. Yoshi's moveset is fairly complex, and the level design built around it is consistently inventive throughout the game's six worlds. I'm less convinced of the need for the need for, say, Yoshi's helicopter powerup or super baby Mario than Mario's various suits and hats in the older Mario games, but it's not a huge issue. The much-derided crying Mario "health" mechanic is one of the game's bigger strokes of core design success, providing leeway for mistakes in a way that's a bit more interesting—and less punishing—than in earlier Mario games, while still enforcing a degree of consistency and quick, smart responses from the player. If the level themes get a little repetitive at times, the art and especially the backgrounds are consistently beautiful. Whether it's the best-looking Mario game probably comes down to personal aesthetic preference (there's something to be said for the constrained strangeness of the NES visuals, especially in SMB3), it's certainly the most impressively detailed.

My biggest gripe with the game come down to the persistence of arcade-style design conventions (coin collecting, limited lives, etc). While I don't have any issue with these conventions per se, provided they are smartly integrated into the game design as a whole, here they feel almost vestigial. It smarts especially when the reward for unlocking a semi-hidden door turns out to be a slot machine minigame. This stuff feels strange in an otherwise (relatively) slow-paced, exploratory game, and gets at why my favorite SNES games tend to be the ones that abandon these conventions in favor of something else. Then again, the games I'm thinking of have direct antecedents back on the NES which I've never played, so back to gamer college for me I guess?

Doom should be understood not as a "first-person shooter" but as a highly asymmetrical real-time chess variant. While the first Doom is arguably the better experience right out of the box, mainly due to the sequel's less consistent level design, the addition of the super shotgun and a richer bestiary add a great deal of depth to the game. Of course, Doom II is not just the game out of the proverbial box: it's also the entry ticket to perhaps the single richest archive of fan-made levels for any game, and it's this fact that makes it an essential game. Many people tend to regard fan-made content as necessarily inferior to "professional" design work, especially when it dares to push the difficulty ceiling beyond their comfort zone, but this is ignorance and laziness. The truth is quite the opposite: the collective effort of iteration over years and decades inevitably pushes a game's level design in directions that no single team could accomplish or even imagine in the span of a release cycle or two. Of course that isn't possible if the base game isn't good at its core. Doom II just happens to be perfect.

Myth is a unique series in the annals of the strategy genre. The factors that make it unique limit its appeal to what I imagine to be the typical RTS fan, but they have also granted it an enduring cult status in the eyes of the few (enough, at least, to secure updates to modern systems in the form of Project Magma, though not enough to save it from the abandonware limbo). They superficially resemble RTS games that were popular at the time, but any base building or unit construction has been stripped away. There are no resources to manage other than your units, which number in the low tens at most, and there is no opportunity to replenish them during a mission (outside of scripted reinforcements, which are rare). It goes without saying that each unit is therefore very, very precious. Myth also boasts of an innovative engine that treats each projectile as a separate physics object, subject to gravity and weather effects. All of these factors together place overall gameplay emphasis squarely on battlefield tactics: formations, positioning (to take advantage of terrain as well as avoid friendly fire), and various kinds of tactical maneuvers. The number of unit types is not large, but there is no redundancy, and nothing really missing aside from a scout-type unit which could add a thin layer of strategic depth.

In typical Bungie fashion, there is a lot of effort put into story presentation in the campaign. Granted, it is typical dark fantasy shlock. For, uh, mythical reasons, an army of undead is on the march, set to annihilate the human race. Hope is, naturally, all but lost. Thankfully, the actual missions sell this scenario as much as the impressive narration and drawings that introduce them. The campaign is notoriously hard, confronting the player with overwhelming amounts of undead resistance from an early stage. Early missions have a good variety of objectives, and the player is well-advised to choose their battles wisely in completing them. As the campaign drags on, however, more and more missions force the player to hold a position while fighting off wave after wave of opposition, or march from point A to point B while killing everything in between. These become extremely punishing, with the last mission in particular being a brutal slog where all but the most minor mistakes force a reload. Mission briefings are more about storytelling than anything else, so success becomes a matter of memorizing opposition forces rather than any kind of strategic planning. The game's tutorial basically tells you to savescum, so there's that.

Soulblighter is a strict upgrade over The Fallen Lords, but the games share technical flaws like shoddy pathfinding (Soulblighter's is way better but still) and a camera that is too close to the ground (which also tends to limit the terrain depth promised by the game's physics). Soulblighter's map design is much better as a whole, but should you for some reason want to play The Fallen Lords' campaign it's been ported to Soulblighter as a mod.

As a single-player experience Myth has a lot going for it: a strong theme and unique playstyle that is accessible to non-strategy fans, but challenging enough to take some dedication to get through. It's certainly not deep, but there's a reason I've kept knocking my head against it over the years. Multiplayer Myth looks like bloody chaos in the best way possible, but unfortunately I have been unable to get any of my friends to be interested in the game. Miraculously there are still servers running for online play, so maybe I'll hop onto one of those at some point.

On paper, this is absolutely what I wanted from a sequel to the original Knytt—the same format, tapping into the intrinsic pleasure of wandering with no other purpose than to see what's around the next corner, over the next hill, on the next screen—but in an even bigger world this time. And the new, more sombre but still very colorful graphical style is an interesting step from Nifflas's earlier games.

But I have to admit I really dislike the platforming here, which is a shame because the game places a much greater emphasis on platfoming challenges compared to the earlier Knytt titles. There are hints of this in Knytt Stories (even putting aside Kaizo Knytt stuff), not to mention Within a Deep Forest (which has its advocates, though I never got into it myself). Bouncing ball mode is the main culprit here: I just find it impossible to control it with any degree of precision. Who knows, maybe proficiency and even mastery is totally possible and I just suck. If that's the case I'm willing to live with that.

But I do enjoy putting this on and wandering around for a half hour every now and again, and that's worth something.

This is the first Zelda game I've ever beaten (on an emulator of course). I didn't grow up with a Nintendo in the house, and the mentality required to get through a lot of those really old console games is kind of alien to me. So you can rest assured that exactly zero nostalgia is informing my statement that A Link to the Past is to this day the gold standard in the action-adventure genre. It's frankly baffling to me that it could be be considered dated when it has at least as much going for it as many of the 2D indie darlings of recent years, or overly cryptic or punishing considering the massive popularity of a certain developer of action-RPG titles.

The overworld is large and open enough to feel expansive, and small enough to backtrack through without feeling like a slog. There are enough secrets and side areas available from the beginning to encourage poking around right away, while tantalazingly inaccessible ones hint at future abilities.

Dungeons are on the whole immaculately designed. Each one has a distinctive identity, iterating on concepts and challenges from previous dungeons while introducing new ones. The bosses are just as consistent.

Link's inventory strikes a rare balance between offering a lot of options without being too much. Dungeon items remain useful long after they've solved their initial puzzle, with most of them having multiple uses inside and outside of combat.

Combat itself is as good as it gets. Your basic attack remains useful throughout the game, and mastering it requires careful management of both your position and orientation. Your expanding inventory unlocks a variety of options without risking redundancy or trivializing the challenges you face. Controls are crisp and responsive, and enemy types are both numerous and differentiated.

A Link to the Past is imperfect, as games of its kind inevitably are. All things considered, there's remarkably little about it to actually criticize without going into detail about specific puzzle solutions or boss strategies, all of which recede in the light of what the game gets right—which is damn near everything.

Windjammers is pretty much Pong spliced with the DNA of a fighting game. Anything about different movesets, combos, specials, etc you could say about beating up dudes and dudettes in a fighting game you could say about tossing a frisbee in Windjammers. It's certainly more on the accessible side, but considering how, uh, loud it can get as a couch multiplayer game I'm sure it could hold up pretty well in a more competitive setting.

Apparently the game was big in France, and it was a French company that handled the Switch port and the subsequent sequel. The sequel is exactly what you want from any such thing. The basic moveset is unspoiled, and the new moves are integrated in a way that enhances the core. Ditto for the new court gimmicks. And since the company is French they made sure to make the new French character the hottest.

It's Windjammers 2 and it fucking rules.

The most exposed single player Quake mod out there by far, and also thankfully the best and coolest. Between the levels that come out of the box to the still-growing number of third- (fourth?) party maps that have been made for it (including my own: google Threat Assessment on Quaddicted dot com, and look out for Azure Anesthetics "when it's done") you have every type of Quake experience imaginable, from tight gameplay-focused maps to massive adventures where you can spend hours combing for secrets. While I'm not in love with all of the additions, the amount of added content from monsters, items, and traps to under-the-hood mapper goodies is as impressive as it gets. But to me, the single most important change is converting the shotguns to projectile weapons, which raises the skill ceiling while also subtly upping their raw power (or at least making gibs come more often?). And when I thought it couldn't get any better, the Jump Mod's jump boots were added.