The Stack

Sometimes, the environment in Gish looks downright photorealistic. Which is strange, because it’s not. At all. It’s a tile-based world with lots of regularly-repeated textures, and while it contains moving physics objects (such as blocks suspended from flexible ropes), those objects are largely made from the same size of square tile as everything else in the environment. But sometimes, just sometimes, it gives an impression of looking close to real. I think I can identify a few factors contributing to this.

First, there’s the darkness. This is a dimly-lit game, and this helps to hide the imperfections. Second, there’s the light. Gish features dynamic lighting. Now, “dynamic lighting” usually means “moving shadows”, which may or may not make things look realer. (In Diablo II, for example, I felt the shadows just made things look odd.) But in a dimly-lit game like this, it more often means moving brightly-lit areas, sometimes shadow-striped. Add a moving shadow-casting physics item in the middle of this and you get a convergence of visually convincing stuff, forming an island of concentrated realism.

And that “island” effect is a big contributing factor. When things suddenly look real, it’s partly due to the contrast with the way they normally look. The monsters are grotesque cartoons, visibly hand-drawn. Gish himself looks like an animation cel, with big yellow eyes and a toothy mouth that’s only rendered when it’s open. But sometimes, when the light hits him just right, he gets spots of dynamic reflection that suggest a curved and shiny surface. It’s particularly striking when the light is colored. This is one of the best visual effects in the game.

At a casual glance, Audiosurf looks a lot like an early Harmonix game, like Frequency and Amplitude: the player drives a little spaceship down a twisting multi-lane highway in an abstract environment, trying to hit colored spots in time with the music. But the similarities end there. Harmonix, even in the days before Guitar Hero, has always been about capturing the feeling of performing music. The player’s goal in their games is to add something to the soundtrack, to build up a piece of music note by note, by hitting the right buttons at the right moments. In Audiosurf, the music is there regardless of what you do, and buttons are strictly for voluntary use of special powers not directly related to the music. You don’t even necessarily want to activate all of the colored spots, like you would for a perfect performance. That’s because you’re not in any sense performing the music. You’re reacting to it.

Or, to be precise, you’re reacting to the level design, which is generated automatically from the music. Procedural generation of game content from music has been done before — Vib Ribbon, released in 1999, may be the earliest released example, but it came too soon to take advantage of ubiquitous networked digital distribution of music the way that Audiosurf does, providing built-in iTunes and Last.fm integration, as well as a small weekly roster of songs to download from the Audiosurf servers. (As I write this, it’s Jonathan Coulton week.) Not that you’re limited to this content; any song you have in a DRM-free format is useable, provided it meets certain requirements such as a minimum length.

And that’s a snag for such as me. This game is really meant for playing with your own music collection, and I don’t own a lot of music. I never went through a music-collector phase like most people; my collector instincts attached themselves to games instead. I have a handful of CDs left over from my college days, back when people still bought CDs: several They Might Be Giants albums, some Satie and Prokofiev and Philip Glass. I have a few recordings of bands that friends of mine were in. And I have the DROD soundtrack CD. This is little enough that I don’t even really consider it representative of my own musical tastes. Still, there are a bunch of songs there that I haven’t listened to in a long time, and this is as good an excuse as any to drag them out.

So, how well did it handle the music I had available? It varies. It’s probably at its best with dance music, or things resembling dance music. Playing Satie’s piano works, mainly quiet and slow things, the burbling electronic sound effects of the game itself felt very weird; I suppose I could turn them off, but then I’d lose a valuable channel of feedback. Moreover, the whole way it detects the tempo and intensity of the piece seems tuned more for the way pop music works than for classical-ish stuff. For example, in one of the Glass pieces, a very steadily-paced work throughout, the path tilted straight downward simply because a bassoon joined in. This is supposed to be what happens in more intense sections; sedate stuff has the path tilting upward. (Why not upward for rising tension? Because the slope determines how well you can see what’s ahead of you. Downward slope means limited visibility.)

Still, some of the less modern stuff works well. The “Montagues and Capulets” theme from Prokofiev’s Romeo and Juliet — you’d probably recognize it if you heard it — made the road satisfyingly bumpty in just the right places. And sometimes even TMBG was awkward, as in the opening of Ana Ng: the speed at which you go down the road varies with the music volume, and the unnaturally sharp and echoless cut-offs here made the vehicle jerk and judder like it was having engine trouble.

I think the most satisfying ride I’ve had yet was in Floe, by Philip Glass. It’s almost like he wrote it with Audiosurf in mind. I mean, just look at that intensity graph in the upper left of the screenshot. (You’ll have to click on it to see it; it’s invisible in the thumbnail.) Most songs have spiky and irregular graphs, but here, it looks like it’s made of circular arcs. The effect on gameplay is a smooth progression from easy to difficult.

According to the leaderboards, only four other people have tried that song. Most of my songs, no one else has ever tried. But that might be misleading: there may be other people playing Prokofiev, but they wouldn’t show up as competing with me if they’re playing a different recording. (And quite right, too: different performances could vary the level design in nontrivial ways.) This is another way in which my music collection isn’t ideal for the game. But I suppose it could be worse. With the company I kept in college, I could have easily wound up a fan of Karlheinz Stockhausen. In fact, I’m kind of tempted to download some of his stuff to see how well the game copes with it.

Gish is a goth/cartoony 2D platformer starring a sentient ball of tar, a hero that stretches and splats and sticks to things. In some respects it’s very traditional, based around Mario-old conventions like the kidnapped-girlfriend plot ¹No, Gish’s girlfriend isn’t a ball of tar. She’s a cartoon goth girl. Don’t ask me for details. , the linear sequence of levels grouped into “worlds”, enemies you kill by jumping on their heads, etc. I think the designers chose to adhere to old and even outmoded conventions as much as they did, not out of lack of imagination, but to give the player something familiar to cling to as they pulled the rug from under you on the mechanics and controls. This is a platformer where jumping is difficult to execute.

Oh, sure, there’s a “jump” button, but it doesn’t do much of a jump by itself. To make Gish do a jump worth jumping, you have to hit the button while he’s compressed — the more compressed, the better. And he’s at his most compressed when he’s in the middle of colliding with a hard surface, such as a floor. So standing jumps are a no-go, but sequences of long bounds are doable, once you’ve attained sufficient facility with the controls to carom about with confidence. Momentum is your friend, and hesitation is your biggest enemy.

It takes a while to become that confident, because, aside from the jump, Gish’s core abilities are not standard platformer fare. The core controls allow you to make Gish stickier than normal (useful for climbing walls), slicker and less viscous (useful for sliding down narrow chutes), or heavier and more resistant to deformation (useful for breaking things or sinking in water). These can be combined arbitrarily: sometimes, for example, the easiest way to shift a pile of loose blocks is to trickle into their midst through slickness and then tense up to resume ball-shape and force the blocks apart. It’s all very physics.

I may not be playing it much, though, because it’s still crashing for me. Not as badly as it was before, but I’m typically getting in about 10-15 minutes of gameplay before it exits to the desktop without so much as an error message. This is, however, long enough to make permanent progress, so I really could just keep playing, as long as I exit the game every once in a while to force it to save. (There is no manual save option.)

I think my mistake in previous attempts at the final level was underestimating the efficacy of grouping your heroes into parties. Oh, I had tinkered with parties before, but the rules don’t let you do so until you’ve upgraded your palace to level 2. I suppose this should have been a signal to me that the designers considered it to be too powerful a technique for the early stages of a scenario, but in practice, it just meant that I seldom tried it until my heroes were pretty well advanced individually. At any rate, hiring a couple of cheap elite Lords and tethering them to healers seems to be a winning strategy. As in previous levels, there came that turning point when I realized that I had managed to clear most of the map of monster lairs, and that my base was therefore no longer under serious attack. Even then, my trepidation about actually sending my heroes into the final assault against the final enemy caused me to delay more than was really necessary, building up cash for on-the-fly resurrections and spellcasting, creating more parties, etc. But the deed is done, and the Barlog is dead.

That’s not a typo. As in Ultima (Balrons) and D&D (Balors), what we have here is a game that isn’t under license to the Tolkien estate and therefore has to make do with a Brand-X Balrog knock-off. Unlike the others, though, they make a joke of it: “Barlog”, we’re told, is short for “Baron of Logic”, Hell’s embodiment of merciless rationality, who taunts you with the logical inevitability of his ultimate victory. I suppose this means that the means of defeating him — building Temples to enlist the aid of the Gods — is a matter of superstition triumphing over reason. It doesn’t really feel that way, though, because the Barlog’s real weakness is that he’s easily distracted: on this level only, you can periodically summon a colossal “Spirit of Kings”, causing the Barlog to drop whatever he’s doing and rush off to fight it, even if it’s in the far corner of the map. It’s hardly rational behavior, so in addition to being a demon from the pit, he’s also a hypocrite.

This entire business is jokey in a way that, to me, doesn’t fit entirely comfortably with the game. It’s strange that this is so, because there are bits of humor throughout the game — the royal advisor’s introduction to each map typically includes comical digressions, and a lot of the things the heroes say during gameplay are hammed up enormously. (I particularly like the elves, who look post-Tolkien but talk like excessively enthusiastic children.) But the advisor typically only says anything at the beginning and end of the scenario (and, due to a bug, sometimes the end speech doesn’t play), and the hero quips, which fundamentally serve to signal status changes like “I just gained a level” or “I’ve decided to flee this encounter”, are repeated often enough that after a while you basically stop noticing the words and just process their relevant content. Fundamentally, the player’s attention is going to be on the gameplay most of the time, and the gameplay itself isn’t funny. So when the Barlog talks, interrupting me in the middle of gameplay mode, my reaction is “Huh? What? Oh, right. Comedy.”

Still, for all my complaints, I think overall I’m glad I got this game, especially given the pittance I paid for it. In addition to the campaign mode, there are several standalone missions. I’ve already dipped a toe into them, and will probably wind up playing through them all eventually.

And another night passes without defeating the demon and reclaiming the high throne. I’ve made a little progress, though. The level contains sub-quests to build one of each of the six types of temple that I mentioned earlier, or rather, to build two groups of three, as the gods of Ardania split naturally into two trinities. I finally managed to complete one triad, and was told that the demon was greatly weakened, and would remain so as long as the temples stayed up. Then, of course, the elementals managed to knock two of them down. So I’m starting to formulate anti-elemental strategies. It seems like what I really need is to pull a high-level Ranger and a high-level Blademaster out of my Lord pool. Rangers respond eagerly to Explore bounties, which should enable me to find the elemental lairs, and Blademasters respond eagerly to Attack bounties, which should enable me to destroy them once found. (The high levels on both are simply to let them survive their missions.) The one big problem with this plan is that it leaves me with scant funds to build the temples, but we’ll see how it works.

Still on the last map. I’ve definitely spent by far the majority of my time on this game toward the end. Fortunately, that isn’t just because there’s an impossible end boss. It’s because the end is where the correct strategy becomes non-obvious, and therefore the game becomes both more difficult and more interesting than in the previous parts. It would probably have been better for the game’s reputation if the developers had front-loaded this more — as it is, I doubt most reviewers played it this far before reviewing it. Which is fair, because most players won’t either. But it does mean that the game as a whole is probably underestimated.

Even my own judgment so far, that the original Majesty was better, is probably at least partly based on overly-sunny memories. The one thing that I can point to as definitely changed is the lack of randomized layouts in Majesty 2, but in a way, the final level brings that back. The whole level is based around a demon who alternately smashes your city and disappears to the far unexplored reaches of the map. It takes a few play-throughs to figure what he’s doing out there, but he’s creating new monster lairs, ones that spawn elementals. If you let him just keep on creating them indefinitely, the elementals become too numerous to fight. So you have to find them by exploring, just like with every single lair in the original Majesty.

The strategy I was starting to formulate at the end of my last post worked like a charm. And the reason it works all comes down to money. There are six specific hero types (Paladin, Blademaster, Beastmaster, some special kind of archer, and two flavors of specialized priestess) that we can think of as “elite” (although the game doesn’t use that term) — they’re more powerful than normal heroes, especially at higher levels, and they cost more to hire. Every elite type costs 1000 moneys at level 1, ten times as much as the cheapest unit (rogues), and a high-level elite-type Lord from a different map costs positively unwieldy amounts to hire: if you can afford them, you’re probably in good enough shape that you don’t need them. But a low-level elite is good enough, because it doesn’t take them long to become high-level elites. Now, you start most maps with 2-4000 in cash, which is enough to hire a fresh elite and still have some left over for buildings, but before you can do that, you need to build the temple that produces that type (cost: 3000), and the prerequisite for that is a level-3 palace (2000 for the first upgrade, 5000 for the second), so that’s 10k before you can get started, and in the meantime you have to build other units just to defend yourself. But with a level-1 elite Lord in the wings, you only have to build a 1k Lord Tower, well within the starting budget, and well worth it.

The tricky part of this approach is getting a level 1 elite Lord in the first place. You pretty much have to hire a new elite type just before winning, not giving the new hire enough time to either level up or engage the boss and get killed. Also, of course, you can only use each such Lord this way once, because once they’ve done a mapworth of foe-slaughtering, they’re not level 1 any more. Still, it worked so well on the level where I was stuck before that I took special care to do another last-minute hire there to use on the next level, where it also worked beautifully. There’s a tremendous moment when you suddenly realize that your peasants’ houses are actually staying up for significant lengths of time — yes, the poor are the primary casualties and all that, and a good buffer zone of peasant housing around the buildings you actually care about is often the simplest way to preserve them — and that you can start thinking about going on the offensive.

That leaves the final map, which I haven’t yet won. It shakes things up by giving you a ton of cash at the very outset — not an endless supply, but enough to open up options that I hadn’t had to consider before. Hiring a high-level elite lord (or an even higher-level non-elite lord) at the outset is suddenly possible, but so is upgrading the castle and making an elite factory or two. What you can’t do is pursue both of these routes at once. I’ve got some experimenting to do.

Three levels from the end, the difficulty spikes. I have now spent long hours on a single map. Forget what I said last time about making steady progress once you’ve survived the initial onslaught. At this point, it seems like the initial onslaught never ends.

The premise of the level is that you’re caught between two rival clans, one of elves and rogues, one of priests and paladins — which is to say, a subset of the hero types available to the player. They’re quite willing to fight each other if they meet, but since your palace is in the way, they both gang up on you instead. Even if you could manage to get them fighting each other, it’s not clear that it would be a good idea: hero types tend to flee battles before they die, and that which does not kill them gives them experience points. Yes, even for non-lethal encounters. Maybe they’re not so much enemies as sparring partners. Regardless, the last thing you want is for them to level up faster than you.

Paladins in particular are very hard to kill, and overpowered for their nominal level. Creating some paladins of my own would be a fine thing, but just getting to the point where you’re allowed to create the building that produces them takes more time and money than I seem to ever have. The only way I’ve managed to last any length of time is by bringing in Lords. Lords are a device peculiar to this game: basically, when you finish a map, you get to choose one of your heroes to become a Lord that you can call up in other maps, for a fee that fee depends on the Lord’s class and level. (Any experience gained after becoming a Lord persists, too, so if you use the same Lord a lot, it becomes more and more expensive.) It’s an interesting touch that gives a little bit of overarching strategy in a game where levels are otherwise self-contained.

Generally speaking, I’ve been promoting my most powerful warriors to lordhood. On the level where I’m stuck, I can basically afford one of my weakest Lords at the very beginning, and perhaps one more later on if I’m willing to ruin my kingdom saving up for it. I’m starting to think that I’ve taken the wrong approach here — that what I really should do is go back and replay a couple of earlier maps and lordify some level 1 paladins, ones that will be cheap to re-hire and then level up in the map where I need them. At any rate, it’s an experiment that wouldn’t take nearly as long as I’ve already spent replaying the level I’m on.

Majesty 2‘s campaign mode has 16 maps, each of which is a satisfying length for a single play session, provided you’re not determined to play the whole thing through in a burst like me. And on most maps, one session is all you need: if you do it wrong, if your heroes die and your city crumbles, that’s just a setback. You haven’t lost, you just need to regroup and try again. Time makes you stronger — partly because your heroes are constantly fighting monsters and leveling up (even if you’ve destroyed all the monster lairs, your own city’s sewers and graveyard keep emitting low-level crawlies), partly because more time means more income, which means more upgrades for your surviving heroes. In theory, you can lose if your palace is destroyed, but generally speaking, if you can survive the very beginning, when you don’t have guard towers or powerful heroes yet, you can survive until the designer decides you’ve had enough time and ends it.

Most maps don’t seem to have time limits. In fact, I’ve seen only one that has an explicit time limit, a novelty level where your goal is to accumulate a certain amount of cash on a deadline (while spending enough to stay alive, of course). This, for me, has been the hardest map in the game so far, because of the self-restraint it requires. If I have a failing in strategy games, its my urge to upgrade everything to max, rather than take a considered look at costs and benefits. Really, though, any time-limited level makes you prioritize what you want to do with your money. It’s just that, on most, the worst thing you can do with your money is hoard it.

On most maps, the mission objective is to destroy something, either a building or a boss monster of some sort. And if I read things right, these boss monsters are how the game imposes time limits without making them explicit. For example, I just finished (on my second try) a map involving an undead king, who either wanders the map or periodically appears and vanishes, I’m not sure which. (There were still substantial sections of the map unexplored when I won.) What’s certain is that he eventually shows up at your base and starts demolishing it. There’s no point in trying to rebuild when this happens. All you can do is set a hefty bounty on his head and hope that your forces are strong enough to whittle him down before he runs out of buildings to smash. The resulting donnybrook feels a lot like the “Armageddon” spell in Populous, the finishing move that makes everyone in the world rush to the center and fight until only one side is left, except that you can still meaningfully participate by, for example, resurrecting heroes that die in the dogpile. Now, both of the times that I played this map, the skeleton king found me on day 86 or so. ¹There’s a little inconsistency about how time works. The UI reports a number of days elapsed, but the graphics display some maps as taking place in daylight and some at night. Perpetual daylight is something we accept as an artifact of the way time is represented in RTS games, but throw in nighttime and it starts to seem a little weird. Also, there’s one level where you’re told in the beginning that it’s an opportune time to attack because it’s raining, and the rain then continues for however many months you need to finish. This could be coincidence, but it seems more like a time limit. A soft time limit, sure, because it doesn’t end things immediately, but if you haven’t built up your forces enough to win the battle by that moment, you never will.

The problem is that it doesn’t read like a time limit. The first time around, it seemed like I simply had a stroke of bad luck and the boss just wandered into my camp before I was ready. If it’s a time limit, it’s a surprise time limit. And while I can see the need for time limits to create challenge in a game like this, I don’t see any need to be coy about them.

I first became aware of Majesty 2 through a series of posts on rockpapershotgun.com. Looking back at them again, one particular passage drew my attention:

The great irony/joy of [setting bounties] is that the only things heroes have to spend their money on is stuff you’ve got up for sale. Build a marketplace to flog potions, a blacksmith to sell armour and weapon upgrades and an inn to booze the night away in. Dimwits – all that gold they’ve earned, given straight back to its source. So, you eventually get a decent portion of your spending back…

…To then in turn spend on new heroes, buildings and upgrades. Nobody gets rich here – the money just cycles around and around. I’ll guess this tongue-in-cheek futility is deliberate…

Fools? Futility? Hardly! Money is not the same as wealth, as anyone in Zimbabwe can tell you. A man trapped on a desert island with a million dollars in a suitcase isn’t “rich” in any meaningful sense, because the lack of things to buy makes it worthless. ¹The presence of a half-dozen additional castaways who still believe that the money has value may change this, of course. Likewise, a Hero who’s decked out in the best weapons and armor, with a backpack full of healing potions and charms, is better off than he was before he spent the time accumulating the cash that he spent on it — and the kingdom is better off for having him.

And ultimately, you’re controlling a kingdom here, not an individual. Individuals may value money as an end in itself, but from a national perspective, the whole purpose of money is to keep it running through the engines of commerce like water through a mill-race, generating useful work through its movement. If you can do this without increasing the money supply, that isn’t futility. That’s more like a miracle.

I don’t want to overstate the case here. It’s not like the game has inflation or a business cycle or anything. In sophistication of economic modeling, we’re somewhere above most RTS games, but below the likes of Sim City. The strangest part, when you stop to think about it, is that the peasants in your domain seem to just generate tax revenue out of thin air. These are people who spend most of their time toiling away at new buildings for you, but you never actually pay them. They pay you. It works as gameplay, though, because it gives you an incentive to protect them.

Come to think of it, there’s an interesting contrast there between Majesty and its evil twin, Dungeon Keeper. In DK, your minions are gold-sinks. You have to keep paying them periodically to keep them from leaving, and gold paid in wages just disappears, as if eaten. If I were a more suspicious man, I’d think that they were somehow sneaking it to the peasants in Majesty 2, explaining two mysteries at once. But no, I think it’s just illustrating the nature of good and evil. Good produces; evil consumes. This means that evil’s preferred strategy is the quick strike, achieving victory before the money runs out and then, presumably, leaving things behind in an unsustainable state. Good, on the other hand, wins by enduring. You’ll be unstoppable if you can last long enough. It’s almost as if they really are two different sides of the same game.