The Stack

I’ve talked about this before, but it’s impressing me anew: in comparison to other DRODs, it’s striking how much of The City Beneath is spent without using your sword. It’s not just the time you take exploring the city proper, where you’re forced to keep your sword in its sheath lest it be devoured by oremites. It’s also the way that the new game elements provide other things for the puzzles to be about. Take the rafts and floating platforms, which are really the same thing in two different contexts. A raft or floating platform is a contiguous group of floor tiles that move under your control when you’re on them, provided that there’s nothing in the way. ¹That’s a bit of a simplification, but I’ll spare you the details. Once that’s in the level designer’s toolkit, they can use it to make polyomino assembly puzzles that you solve from the inside. You don’t need a sword for that.

In a weird way, it reminds me of Portal 2. There was reportedly a point in the development of Portal 2 when the developers were seriously contemplating leaving the portal gun out, because they had a number of other novel mechanisms to replace it with, such as magic paint. But they wound up including the portal gun as the means by which you interact with the other novel mechanisms. TCB sort of does it the other way around: using your sword is the goal, the thing that the other puzzle-solving is done in service of. But then, the game also introduces new ways to kill things without using your sword, such as leading them onto hot tiles, or knocking out the supports of a bridge, or letting them step on your raft and then pulling it out from under them.

I mentioned oremites. When you first learn about the city’s oremite infestation, it seems like they’re nothing more than a tongue-in-cheek in-game excuse for the narrative decision to keep Beethro from killing important NPCs. But fairly late in the game, there’s a level called “Oremite Breeding Grounds”. Here, most of the floor is covered with oremite nests, except for a few tiles — and consequently, those are the only tiles where you can draw your sword. The puzzles are thus all about getting monsters over to the places where you can kill them, with enough time to turn to face the right way once you get there. I just love how it turns what seemed like a throwaway joke into a mechanic that produces puzzles, and I especially love how long it waits between introducing the concept of oremites and paying it off in this way.

I mentioned the sequence where you play as the Negotiator, who doesn’t even have a weapon, and has to kill monsters by controlling Fegundos. There’s a second such interlude where you control a goblin, who doesn’t have a sword, but who’s capable of killing things anyway (as we know from the puzzles where Beethro fights goblins). This is the sequence that introduces Aumtliches, the Empire’s latest vat-grown weapon of war. An Aumtlich’s eyes emit glowing, sparking beams that paralyze you if they touch you, leaving you able to turn in place but not able to flee. There’s a sense of powerlessness there that’s worse than simply being killed outright, and the player learns to fear the sound effect that accompanies getting stuck in the beams. And it all leads up to the moment when you get Beethro back, and have a sword again. To emphasize the importance of this, Beethro has to surrender his sword to be allowed into a military encampment just before the goblin sequence, and it’s only returned after that’s over. It’s important because, as we soon discover, your sword can block Aumtlich beams, or even reflect them back. You can still get stuck without recourse, but only if you stand in a place where two beams cross, so that you can’t block them both at once. Even so, being able to block them at all feels like a superpower after experiencing helplessness as a goblin. At first, I took care to orient my sword correctly before stepping into a beam, just to avoid the scary noise it would trigger otherwise. But the game gradually makes this impossible, and after a while, you learn to not let it bother you.

I’m managing to replay The City Beneath at a pretty good clip — I haven’t gotten to the harder levels yet, but I’m pretty sure that the game is overall much easier than Journey to Rooted Hold, even with the Challenges. The central organizing structure of the game is that your progress through the city is blocked by sundry obstacles requiring subquests. You explore more or less freely for a while, triggering plot events as you go, and then, when you can’t do that any more, you go off and solve a series of puzzle-rooms, then come back and resume exploring. The most significant of these obstacles is a series of three great gates. I’ve just passed the second one. This doesn’t mean I’m two-thirds of the way through the game; the first gate comes at the very beginning, and only requires that you sheathe your sword to pass. Also, I recall there’s a significant number of levels after the third gate. Still, it’s good progress.

The second gate is called the Gate of Namedness. The only people allowed through are citizens with names, which is to say, jobs, those two concepts being synonymous in the Empire. As an above-grounder, Beethro is automatically ineligible for this status, until the Negotiator moves heaven and earth to get the law changed. Really, Beethro severely underappreciates that woman. He remains as hostile towards her as he was when they first met, just as he remains belligerent towards everything in the Empire, even the people who are trying to help him. Nonetheless, getting deeper into the city requires him to integrate himself into the Empire’s systems, at least nominally.

I described the process of obtaining a name during my first pass through the game. Beethro applies for the post of Slayer. There are five other applicants for the position, and they all compete for the job with a fight to the death, in which everyone else naturally agrees to target the above-grounder first. Now, the room contains an obvious mechanism for beating them, involving a bomb and a decoy potion. But this time around, there’s a Challenge scroll, which predictably asks you to do without these aids.

I am of course an expert Slayer-killer by now, thanks to the Challenges in JtRH. What once seemed impossible is now child’s play. But there’s a particular difficulty with taking on five at once, and oddly enough, it’s exactly what made the same fight so surprisingly easy back in 2007: the tendency of the Slayers to bunch up. When you’re trying to get them to cluster together around a bomb, that’s helpful. When you’re trying to manually pick them off one by one, not so much. You want them isolated. The act of getting past a Slayer’s guard with nothing but a sword and a wall is a delicate dance, with precise timing, and easily ruined by another Slayer barging in from the wings. The first kill I managed was almost entirely accidental. Somehow, an interloper wound up in a position I could take advantage of. I have no idea how that happened, but it was a big help. Each enemy you take down makes the rest that much easier.

So! Replaying DROD: The City Beneath. This puts me in the new position of blogging a game that I’ve already blogged. Well, it’s been, good grief, ten years since those posts. My perspective will be different.

Coming on the heels of Journey to Rooted Hold, the most noticeable difference is that it isn’t simply a series of levels any more. JtRH played with nonlinearity a little, giving us up staircases and quests that looped back to previous levels, but the notion of “previous level” doesn’t even apply very well to TCB. Grouping rooms into levels is still important for the game’s organization, if only because of the “level clear” gates, but levels here are named, not numbered, and only partially ordered. I seem to recall that the later parts of the game turn into a linear descent like the previous titles, but for now, it’s hub-and-wheel, with some forking. And, while I recall once pooh-poohing the significance of the hub to the whole, it seems impressively large to me right now.

I had forgotten that the Negotiator from the beginning of JtRH shows up again here, in an expanded role. She turns out to be a bit of a rebel, in her own bureaucratically-bound way, and actually does a lot to help Beethro along, even though he doesn’t appreciate it. More importantly, you get to play as her for a little while, in an interlude between Beethro’s scenes, tutorializing the use of the Fegundo. Switching characters mid-game like this is a new feature for DROD.

And of course there are challenge scrolls now! As before, I’m trying to tackle all challenges as I find them. So far, I’ve found one worth describing briefly here. Recall that TCB contains occasional appearances by Slayer trainees, eager for promotion. In behavior, they’re just like the Slayer from JtRH, but they’re stupid enough to approach Beethro in rooms where they can be killed easily. The first such encounter is in a room full of “hot tiles”, which kill anything that steps on them and doesn’t step off on its next move. This makes killing the Slayer there not just easy, but nearly automatic. Once he gets close enough to be a threat, he’ll pause to reorient his weapon, and consequently die. So the Challenge there is: Keep the Slayer alive. So neat, so elegant.

Those hot tiles are just one of a bevy of new environmental features. I feel like the introduction of new stuff has accelerated. Yes, JtRH had new elements, but not this many, not this fast. And a fair number of them were new monsters, some of which were variations on stuff from King Dugan’s Dungeon, like Rattlesnakes and Awakened Mud. Whereas TCB is focusing less on new monsters and more on new room features that let you interact in new ways. Consider mirrors. These are things capable of reflecting the gaze of an Evil Eye (or, later, an Aumtlich), and which Beethro can push around the floor to block passageways or weigh down pressure plates (themselves a new element). They’re a bit Sokoban-like, except that Beethro can use his sword to push them around corners. The point is, the previous two games didn’t have anything remotely like them. I feel like JtRH was basically about fulfilling the potential that the familiar DROD ruleset obviously had but didn’t quite live up to in KDD, while TCB is more about striking off into fresh territory, and discovering new potential in the unfamiliar.

It’s been nearly a month since the Steam sale, so I think it’s about time to wrap up. But first, I want to revisit a couple of games I mentioned previously.

I had some harsh words for Loot Hunter. Well, it turns out that a recent Humble Bundle contained a strikingly similar game called Windward. That is, Windward doesn’t abstract ship-to-ship combat into a match-3 the way Loot Hunter does, but the rest of the game is fundamentally similar. In both games, you explore an age-of-sail world viewed at a large enough scale that your avatar is a ship rather than a person, and you attempt to make enough profit to upgrade your ship by trading goods between ports, doing quests, and fighting off pirates. But Windward does it all so much better. It’s designed more like a MMORPG. The quests are often elaborate multi-stage affairs, even if they are obviously procedurally-generated. The world is divided into zones geared towards different levels. If you’re too low-level for a zone, you’re allowed to go there, but none of the towns will trade with you, and that gives you a strong motivation to go back where you belong. This division lets the designers better control the pacing and keep multiple upgrade tracks running in parallel.

Now, I haven’t given all my sale purchases an honest try yet, but of those that I have, the one I’m most satisfied with is Creeper World. I played through the whole thing pretty quickly, including the bonus levels (which seem to be the part that differentiates the “Anniversary Edition” from the free online Flash version), and then was pleased to discover that it has a couple of sequels. I purchased the first sequel while the sale was still going, and now I’ve played through that as well.

Creeper World 2 pulls the same strange trick as the second Zelda game: it takes a game in a top-down view and turns it on its side. The battlefields this time are systems of caves and underground tunnels, exactly as wide as the screen but many times taller. It’s still basically a game about fighting a fluid called “creeper”, but instead of just spreading generally outward, it flows downward if it can, filling cavities. When completely enclosed, it can become pressurized, spurting out and expanding at an alarming rate when it finally dissolves a hole in the wall. The most effective way to combat this is to build up some pressure of your own to resist it. In other words, you can generate your own creeper this time around. Once you have this ability, it almost seems like an oversight that it wasn’t in the original game. Here it is, the game’s most distinctive feature, but wasn’t something that the player could do. Being able to create your own creeper gives the game a greater sense of completeness, in the mathematical sense.

For all these big changes, the tactics are fundamentally similar to the first game. It’s still primarily about finding the opportune places to cut off the enemy’s creeper flow so you can move in on the things generating it. One change that’s kind of mechanically trivial but makes a big difference to the feel of the game: In the first game, you couldn’t actually destroy the creeper generators. You could effectively disable them by clearing the vicinity of creeper and then parking a blaster near the generator, so that it would destroy any creeper the generator emitted the instant it appeared, but that’s as far as you could take it. (And even then, there was the risk of forgetting that you needed that blaster there, and moving it with disastrous results, as happened to me more times than I’d like to say.) In Creeper World 2, you have a new device that actually destroys the generators.

And it is this ability that made me think: This is a lot like fighting a Tar Mother in DROD. I mean, it’s not identical, sure — Tar Mothers make all the tar in the room expand, while these creeper generators are just point-sources of the stuff. But in both cases, I’m cutting my way through a mass of viscous blue stuff to kill a thing in the middle that makes it keep expanding. And with that said, I think it’s time to get back to DROD.

Another threshold breached: For the first time, I have requested and received a refund on Steam, something that has become a lot easier lately.

It wasn’t a big refund. A mere 49 cents. This sale price is of course a big part of why I bought the game in question; that’s well within the “I don’t know much about this game, but I’m not risking much by buying it” range. I don’t want to identify it here, but it was sort of arcadey, and 49 cents is slightly less than the cost of two plays on a classic coin-op machine, so that was informing my sense of value here. And in fact I did get two substantial plays out of it, so under most circumstances, I’d call it even and leave it at that.

But I started having misgivings about the game when I first started it, and saw the logo was flanked by two images of Pedobear, like heraldic supporters. This is basically the creators proclaiming “Our tastes and sensibilities were formed on 4chan”, which is to say, identifying themselves as jerks. That’s a minor matter though; I’ve played plenty of other games by jerks. But apparently the flanking images are randomized, because the next time I went back to the main menu, I got Pepe the Frog. This is more or less equivalent to putting swastikas in your logo. This is the creators all but outright telling me they don’t want my money.

I mean, yes, Pepe has a history as both a character and a meme that predates its appropriation by white supremacists. I can easily imagine that including it here was meant as just another 4chan shout-out, and that the developers simply don’t mind being mistaken for neo-nazis. This would make it an act of stupidity rather than a declaration of fascism, but you know something? Even in that scenario, I’m comfortable with asking for my 49 cents back. If I’m wondering whether it’s a sincere expression of solidarity with the alt-right, there are definitely going to be alt-rightists who read it as definitely sincere, and take it as yet another sign that their ideology is acceptable now. I don’t want to support even that, not even with such a pittance.

At any rate, the refund request form was very straightforward, except for the bit where they ask the reason. There’s a list of options, and none of them are “it turns out to expresses support for political loathsomeness”, useful though that would be for a number of games. I can report that I received one Steam trading card drop before the refund period expired, something that I didn’t think was supposed to be possible. Also, I’ve experimented a bit with the consequences for the Steam UI in general, and can report that it doesn’t quite know how to deal with Achievements for games that you no longer have.

Having been through the process once, will I be more willing to exercise it in the future? I don’t know. It bucks decades of habit. Even when it turns out that a game won’t run on my system, my reaction is typically “I’ll probably get it running at some point, maybe after an upgrade”. We’ll find out.

Given how much effort Open Sorcery takes to bridge the gap between the human player and its nonhuman protagonist, I suppose it’s appropriate that it’s basically an AI awakening story. Indeed, NPC dialogue in the intro makes the player aware of the possibility of BEL/S attaining human-like self-awareness before it’s even made clear exactly what BEL/S is. Decker, one of the two magician/programmers responsible for BEL/S’s existence, considers this possibility to be a threat. You can convince him otherwise over the course of the story, but there are branches that justify his concern.

BEL/S learns and grows with every encounter, whether by simply leveling up her firepower from combat experience or by learning new attributes, new elements that can be applied to later encounters. Engage a chaos spirit in conversation, and you might gain access to the Chaos element. It strikes me that the mechanics work towards channeling the player’s story along specific paths here: choosing clever diplomatic solutions gives you the tools you’ll need for later clever diplomatic solutions, choosing violence just enables greater violence.

The spirits you encounter are mostly already self-aware in a human-like way, which raises an interesting point: Without the computer code structuring her awareness, what would BEL/S be? I may be reading too much into things, but I think there’s some implication that before the start of the story, she was just a raw piece of spiritual force made out of fire. Some of the spirits in the game recognize her as something novel, a spirit elevated into consciousness by technology. And, having been so elevated, she has the possibility of leaving that technology behind, attaining freedom but abandoning any connection to her creators.

But — and this is a crucial point — doing so ends the game prematurely. For all its branching, this is a story with a definite ending, a confrontation with a final boss and an epilogue that follows it. This is clearly the true, correct ending, and any branch that misses it is the narrative equivalent of death. Now, I’ve replayed the game enough times to get several alternate endings, and I have yet to see one where BEL/S actually dies. All the “bad” endings I’ve seen are ones where she transcends her limitations and abandons her tasks.

That final boss, by the way, a powerful death spirit, has been called “a personification of depression” by the author, and, as always, the game is extremely clear about the author’s intention. Not all of the spirits are allegorical in this way, but there’s another — a ghost of Decker’s dead lover — that’s clearly an avatar of his grief. The purpose of a magical firewall isn’t just to protect people from external threats (including, at one point, aliens), but to help them with the problems they bring with them. Even if that wasn’t the original intention of her creators, it’s her purpose, the purpose she discovers for herself. I suppose this is why those alternate endings are failures. You can gain immense freedom and power, but once you’ve cut yourself off from humanity, you can’t self-actualize.

I’ve been meaning to check out Open Sorcery since meeting its author at Indiecade last year, and I finally got around to purchasing it in the recent Steam sale. Since then, it’s become a finalist for Best Game in the Xyzzy Awards. And having played through it now, I see why. It’s a very engaging work, and one that goes to great lengths to meet the player halfway.

The basics: This is a story of AI awakening in a world of modern magitech. You play the role of BEL/S, a firewall that’s also a fire elemental, conjured and programmed by two friends and charged with keeping them and their kith and kin safe from evil spirits. Unusually for a commercial release, it’s written in Twine ¹I know of only two other Twine games on Steam: Depression Quest and This Book is a Dungeon. And only one of those costs money. Firewatch gets an honorable mention here for having an intro that was prototyped in Twine and then put into the game more or less as-is. — there are a few pictures, but for the most part, output takes the form of text, and input takes the form of clicking on text. (There’s at least one moment of of freeform text entry as you guess at a chaos spirit’s riddles, but even this is done by clicking letters on the screen.)

Now, the point of view is decidedly nonhuman, unfamiliar with human things, and prone to describing stuff in a mixture of technical jargon and poetry. That has the potential to create confusion and communication problems — particularly in a choice-based game. I’ve certainly seen Twine pieces involving AIs or aliens or whatever where I had no idea what was going on, even after playing all the way through. Parser-based IF usually manages to avoid this by making progress contingent on understanding. As such, its failure mode is the player not making it all the way through, but even this is often avoided, because most information takes the form of room and object descriptions, which the player can view again when confused.

All of which is to say that Open Sorcery does an incredible job of keeping the player oriented. To start with, it makes use of routine. The game plays out over a series of days, and your primary task on each day is the same: visit the same four locations, look for signs of evil spirits, scan for any you find, and deal with them. On most days, there’s exactly one location with a spirit to confront, but you have to visit the other locations too, the better to experience plot elements that haven’t reached a crisis yet. There’s freedom within this structure, of course — once you locate and identify an evil spirit, you get to decide what to do about it.

Now, every spirit has two attributes, Matter and Motive, that you have to choose from a list of possibilities to identify it. (For example, BEL/S’s Matter is Fire, and her Motive is Order.) This is a necessary part of initiating an encounter; if you don’t know what you’re looking for, it remains invisible to you, and your progress in the story is blocked until you get it right. How do you know which Matter and Motive to choose? By paying attention to its effects on the environment. Notice that this is more or less the condition that I just described as typical of parser-based IF, that progress is contingent on understanding. In form and in purpose, the identification step is a puzzle — just not a puzzle you’re likely to get stuck on.

There’s a Twine technique I’ve described before wherein clicking on a word or phrase replaces it with a different word or phrase, going through several possibilities before settling on a final one. As I described it before, this is used to represent the narrative voice trying out different descriptions, different ways of thinking about things. Here, it’s used largely for clarifications to bridge the gap between player and protagonist, human and nonhuman. In fact, it bridges that gap in both directions. There are links that clarify BELS’s native jargon, as when the words “physical password” are replaced with “key”, and there are also links that give BEL/S a chance to google things, learning that the “edible organic matter” is a peanut butter and honey sandwich.

One way that the game really goes out of its way to be clear about what’s going on: Whenever there’s a conditional based on your character stats, it lets you know. If there’s an option that’s unavailable because you don’t meet the requirements, it’s displayed but grayed out. If there’s a response that’s conditional, it gets an entire page to itself, with the stat identified. This game wants you to know exactly how it works. I find it intriguing how different my response to this was depending on the stat used. There’s basically two sorts of stat: there’s your firepower, used to determine success and failure in blasting spirits with fire, and there’s your relationship level with several characters, used to determine their reactions. I’m accustomed to games where the mechanical details of characters’ minds are hidden from the player, the better to maintain the illusion that they’re not machines. So seeing the game call attention to the relationship stats so vehemently felt a bit strange. Stats in combat situations, on the other hand, I take in stride — so much so that I didn’t even notice that it was being treated the same way until I started writing about it here.

I haven’t really gotten into the story yet, but this is already a longish post about a shortish game, so I’ll leave off for now. I’ll probably have more to say about it tomorrow.

The nondescriptively-titled Loot Hunter is a pirate-themed game in the genre that I’d call “puzzle-quest-like” if that weren’t so ambiguous. It’s ambiguous because Puzzle Quest was two-layered, an RPG wrapped around a match-3, and you can generalize that into either “RPG wrapped around an arbitrary minigame”, as in Runespell: Overture, or “arbitrary overgame wrapped around a match-3”, as in Hunie Pop. Loot Hunter is basically the latter, although the overgame is pretty RPG-ish in every way but presentation.

In presentation, it’s Grand Theft Galleon. Your ship freely roams a map of an archipelago, trading goods from port to port, doing optional quests, hunting for buried treasure, and attacking other ships for their booty. The game’s intro text makes it sound like you’ll have to choose between the path of the honest trader or that of the freebooter, but in fact there’s very little motivation to refrain from pursuing every path at once. Even if you’re just doing quests, the British can give you quests to attack French or Spanish vessels and vice versa. The three-sided faction system reminds me a little of GTA2, but only a little.

The main reason I’m singling this game out for a blog post is that it’s a very clear demonstration of something I’ve observed before: if you stuff a match-3 into an overgame, I will be unable to stop playing it, even after it stops being fun.

To be clear, it does start off fun. It starts off by giving you a bunch of things to explore! There’s a terrain exploration element, with most of the map shrouded by clouds until you sail them off. There are special ship-to-ship combat skills, essentially spells, to acquire and upgrade. There’s a series of increasingly large and expensive ships to buy, and items you can install in them for bonuses. And of course there’s the matter of figuring out the best tactics in match-3 ship-to-ship combat. The thing is, none of this lasts. It doesn’t take long to explore the full extents of the map, and once you’ve done that, there’s no more exploration. It doesn’t take long after that to max out your spells. Once that’s done, there’s nothing stopping you from discovering the optimal tactics, which you don’t have to vary ever.

Well, that’s not quite true. Sometimes you’ll be attacked by a ship that’s simply too powerful for you — say, if they have enough cannons to take you down in three volleys or fewer — in which case your tactics should shift to rapidly acquiring enough mana to cast the “escape from combat” spell. But in all other cases, including cases where the enemy outguns you only a little, you can treat all encounters the same, because they basically are. There’s none of the variability in rules or abilities or obstacles or even just board shape that I expect of a good match-3; the only differences between ships are differences in power. Even when the game describes an encounter as being with a group of ships instead of only one, it’s treated like just one larger ship in combat.

Once you’re that far, the game is really significantly reduced, and you’ve still got hours before you can afford the biggest of ships. And yet I wound up playing long enough to do that, and a little while longer besides. At this point, I’m not really getting anything out of the experience, apart perhaps from the opportunity to listen to it — the music is really lovely, varying from Jolly Little Nautical Tune to Stirring Adventure Music to Dark Orchestral Menace, and the sound of splintering wood from the cannons is really well-done. The game doesn’t seem to strictly have an ending, but there’s one more goal — maxing out your social status — that’s so close that I may well wind up playing to that. After that, there are some Achievements that are basically only achievable by playing for a very long time after the game has been rendered completely trivial and devoid of further goals.

The lack of a definite ending may be the game’s biggest design problem. An ending provides a basis for balance: ideally, in a game based around leveling and upgrades, everything the player is doing should reach its peak at around the game’s end, which is to say, at around the same time. Lacking anything like this to tether the game’s systems to, different parts of Loot Hunter have their endings at wildly different times.

OK, obviously I was going to take a little break from DROD when the Steam Summer Sale came up. Like last time, I’d like to take a brief look at a few of the things I bought for cheap.

As usual, I started by just getting the cheapest things on my wishlist — and from the results, I think I must have really been nostalgic for xbattle recently. This is a game I remember playing back in college, at a time when online multiplayer games were only easily played at an academic computing lab. Apparently there are multiple games called “xbattle”, so I should specify that the one I mean was a primitive RTS, named for its use of the X11 windowing system, and notable mainly for its high degree of abstraction. The battlefield was a grid of squares (or hexes, but we mainly played it in square mode), and any troops were represented by a colored square within that, the number of troops indicated by the size. Certain points on the map produced more troops at a steady rate for whoever captured them. You moved troops around by pointing squares at adjacent squares, causing the color to slowly drain from one to the other over time at a rate governed by their size difference. Consequently, the game was mainly about supply lines, each player trying to keep troops flowing efficiently to the front and interfere with the opponent doing the same, to keep them from pumping in reinforcements as fast as you.

As my choice of words here indicates, it feels a lot more like a game about fluids than a game about armies. Supply lines visibly have a pulse — we sometimes called it “the pulsing squares game” back in the day. Directing troop movement is a matter of opening and closing valves.

Of the games I’m recognizing as xbattle-like, Five Elements, a casual game with slick production values, is the closest and clearest. It uses a semi-randomized freeform graph of nodes instead of a grid, but it the essential mechanics are the same: opening spigots from node to node, pouring substances to fight. If anything, it’s a purer expression of the idea than xbattle is, because it makes no pretence that the fluids are armies. The battles are called “meditations”, and are supposed to be taking place inside the player character’s head. The enemies are aspects of his mind that he’s trying to subdue, such as Laziness and Fear and Anger, with appropriate special abilities. Instead of troops, you’re channeling abstract essences drawn from traditional Chinese alchemy, the five elements of the title. These are bound up in an extended rock-paper-scissors cycle of vulnerabilities and resistances: water beats fire, obviously, but also water “devours” metal, meaning that attacking an enemy’s water node with your metal node will strengthen it. Every pair of elements has one of these relationships — either one destroys the other or one devours the other. I find this difficult to keep track of, in part because the visuals don’t suggest the elements very clearly. The UI provides some help here, displaying an icon indicating what you’re in for before you open a spigot, but it’s still hard to make long-range plans under time pressure when you’re querying every possibility this way. Perhaps for this reason, the levels tend to be small. There’s a lot of them, though. This is clearly meant to be played in little bits over a long period of time, and I don’t know if I’ll want to keep up with it to the end.

Energia, a shortish lo-fi indie piece, also uses a freeform graph, because the graph is one that you build yourself, by placing nodes and blasters and linking them up to pre-existing power sources. It’s basically a tower defense, and it doesn’t much feel like you’re piping fluids around — it goes for more of an electricity feel, with energy leaping through nodes of the graph instantly. But I count it as xbattle-like for two reasons. First, it’s largely about maintaining those supply lines. Attackers can come in from any direction and sever your graph at any point that isn’t adequately defended. Secondly, it’s extremely abstract. This is a not really a representation of a battle, and the game takes advantage of the fact when serving up variations. For example, one level is based on the idea that the battlefield is constantly scrolling to the right. Your base is fixed in place on the right side, the enemy’s base is on the left, but everything else, including anything you build, moves past you and then is lost. Playing this level feels like trying to walk up a descending escalator. Another level puts all the power sources on a pair of large rotating circles. Basically, the game’s greatest strength is the variety it can get away with because it’s nonrepresentational.

The highly satisfying Creeper World is mechanically similar: it’s basically a tower defense in which you build a network of nodes that rely on their connection to your main base to receive power. “Energy packets”, for construction and ammo, visibly flow along the lines of the graph, giving it a little bit of that fluid-and-pipes feel. The part that really feels fluid, though, is the enemy, called “creeper”. Creeper is quite visibly a liquid. It’s like you’re fighting the ocean. It’s viscous enough that it takes a while to level out and can even pile up around its outflow pipe, but it quite visibly flows like a liquid, ripples when you drop a mortar shell into it, fills up holes in the ground in preference to advancing toward your base — elevation is key in so many ways. (A note in the game says that it’s actually modeled on heat flow, but physics has a way of repeating itself in different contexts.) As a result, you get to play offense in the supply-line game. Cut off all the inlets into a lakebed, and it’ll dry up, at which point you can claim it for yourself. That’s one of the more satisfying things in the game: painstakingly cutting through a crucial stretch of creeper and seeing sudden and dramatic effects.

There’s a room pattern introduced back in level 16 of Journey to Rooted Hold, of rooms entirely filled with bombs except for a narrow and squiggly path. Accidentally nudging a bomb with your sword makes the entire roomful blow, emptying it in an instant. It’s the game’s strongest expression of overwhelming power, so of course the designers bring the pattern back when it’s time to kill the Slayer. The climactic one in level 25 is easier to navigate than the ones on level 16, though, which strikes me as a smart move. The main purpose of this room is story, not puzzle, and it wouldn’t serve the story to frustrate the player right on the verge of victory over the main antagonist.

In fact, that could apply to most of the level. Apart from the ending, level 25 is mostly about a new enemy, the red-uniformed “Guards of the Poppy Brigade”, brought in by the Slayer for a last-ditch holding effort. These guys are kind of like Slayers, but not as smart. Like the Slayer, they wield weapons, and they know how to navigate around walls (which makes them smarter than most monsters), but they’re vulnerable to some really basic swordfighting tricks. They like to keep their swords pointed toward Beethro even when that’s not the right thing to do, letting you kill them with maneuvers that they could have blocked if they knew the right dances. So they’re only dangerous when they mob you from multiple directions. The rooms do build up to that, but only after a whole bunch of easy tutorializing, and even when the puzzles get hard, the guards still kind of feel easy just because I’m mentally comparing them to the Slayer. Being mobbed just enhances that, because it’s an opportunity to kill them in droves.

Now, in that final Slayer room with all the bombs, there’s a moment of hypocrisy so pointed that it has to be deliberate. When Beethro lights the fuse, the exasperated Slayer complains that Beethro doesn’t know what he’s doing and that he’s putting “many thousands of innocent lives” in danger. And he has a point – Beethro really doesn’t know what the consequences of his delving will be, and we know from subsequent games that a whole lot of people, both surface-dwellers and citizens of the Empire, wind up getting killed. But anyway, Beethro retorts that the Slayer isn’t in a position to lecture Beethro about killing people. The Slayer starts to reply that Abovegrounders don’t count, but catches himself, apparently aware for once of what he’s saying. The thing is, at this point they both act as if Beethro has won the moral high ground, even though he’s just cut his way through a battalion of Poppy guards, treating them like they don’t count because they’re Belowgrounders. Really, it’s all just furthering the same who’s-really-the-monster stuff that got started when Halph befriended the roaches, or even earlier, with the Neather.

After that, it’s on to the sequel. Beethro swears to get to the bottom of this whole Empire business, which he eventually does, literally. But the player isn’t done with the game yet. There’s still Mastery. This means solving every single room, including the hidden ones — and some of them are hidden much more cleverly than I remembered. (Completing all the Challenges is not necessary for Mastery, in part because Mastery existed before Challenges.) As I’ve said, the game gives you a lot of help here. You can reload any room you’ve visited, and once you’ve completed the main game, the “Restore” menu tells you the number of secrets on each level, making it a lot easier to find them. Hunting down your missing secrets on every floor is a nice way to look back at the whole game. I like to do it from the bottom up, as if returning from the journey Beethro hasn’t really finished yet.

Mastery gives you access to the “Dreamplane”, a sort of a museum of concept art and rejected room designs, which you explore in the same engine as the rest of the game. It’s impressively large, easily the largest level in the game, and I find it oddly engaging. Usually I’m not much interested in concept art galleries in games (unless the concepts are very different from the finished work), but piloting Beethro around the floors, looking for interactive bits but not really worrying about puzzles, is a relaxing end to a sometimes frustrating game. Moreover, there’s an element of the uncanny to it. See, the Dreamplane is physically accessible from the dungeon. There’s a stairway to it on level 13. You can’t explore beyond the entrance room until you achieve Mastery, but you can at least enter the level, which gives it something of an aura of in-fiction reality, despite being completely fourth-wall-breaking. Furthermore, getting past the Master Gate and into the rest of the Dreamplane effectively requires time travel. Now, when you jump around from level to level looking for secrets, you’re effectively rewinding time back to earlier points in your explorations, but this is non-diegetic time travel, more like turning back the pages of a book. The resulting Mastery doesn’t violate the narrative because even if you didn’t get all the secrets on your first pass, you theoretically could have done so. Getting past the entrance to the Dreamplane, on the other hand, isn’t theoretically possible within the narrative, because there are multiple points of no return between level 13 and level 25. Time travel isn’t necessary for Mastery, but it is necessary for getting back to that Gate once you have it. And once you have a physically real place that’s only accessible through metanarrative trickery, you have something very strange indeed.