The Stack

I’m still playing Gemcraft: Chasing Shadows. Because the game isn’t strictly ordered, there are still multiple levels I haven’t beat — in particular, the “Vision” levels, optional strategy-puzzle challenges where you don’t have access to the skills and XP you’ve accumulated. Without the option of just bludgeoning a level to submission with superior force, the game can actually be pretty challenging.

But also, if I finish all those, there are still the Achievements. There’s a lot of them. Do I want to achieve them all? I don’t know. Maybe. It depends on how many are left after I’ve completed all the levels. But there’s one particular achievement that I definitely want to try for, and that’s because it’s a riddle. Its name is “Grey Trees” and its description, where most of the game’s Achievements give you explicit instructions on how to get it, is simply “11331791”.

Some possible leads: I’ve seen some grey trees in a level or two; there was one level in particular where all the trees were grey. The in-game Achievements page can be filtered by various keywords, such as “Gem” or “Enhancement spell” or “Destroy”, and the only keyword for Grey Trees is “Click”. Most levels display gameplay tips while they’re loading; a few instead show a row of gem shapes. Since the shape of a gem indicates its “grade”, this is a way of representing a sequence of numbers. And every one of the levels with the gem shapes also contains a mysterious compass embedded in the ground, which rotates to point in a new direction every time you click on it. The direction of the compass has no obvious effect, but the game considers them important enough that compass levels are marked with a special icon on the map screen.

I’m assuming that I’ll have to turn the compasses to some particular direction, but what? I’ll have to do some experiments, find out if the gem shapes vary from level to level and if changing the compass direction changes them. And once I get everything into the right orientation, what then? Is that the only step needed? Will it open up some extra-secret level? I don’t know.

It all reminds me of the special post-Mastery levels in the later DROD games. But DROD was already a puzzle game; adding in additional secret puzzles was far from unexpected. But then, neither is it incongruous here. It may not match the gameplay, but it fits right in with the fiction, a story of wizards facing uncertainty, fighting a shadowy foe who outsmarts them at every turn.

The Gemcraft series is pretty minimalist about its creeps. It uses just three archetypes: normal enemies called “reavers”, weak but fast and numerous “swarmlings”, and tough but slow “giants”. Any wave will consist of just one enemy type, with their appearance randomized from wave to wave, probably to help justify their increasing stats. Sometimes a wave will have randomly-assigned special powers. And that’s it, for regular enemies.

But there are also special monsters. Gemcraft: Labyrinth, the third game, had various special bosses like Arcane Guardians on key levels: you’d think you’re done because the last numbered wave is over, but then something large and glowing and very hard to kill would start making its way very slowly down the path. The final level was protected by a Shadow, a creature made of particle effects that moves outside the paths and has a fairly complicated repertoire of behavior. Also floating free from the paths were the ghostly Apparitions, which are kind of like the saucers in Space Invaders: they don’t attack at all, but you can shoot them down for a bonus. Apparitions aren’t bosses. They just appear at random from time to time.

And it’s these random appearances that Chasing Shadows adopted as the basis for all of its special monsters! There are no bosses per se here: special challenge levels are instead done by giving you special tasks, like destroying locks or activating ancient devices. But we get random appearances from boss-like creatures. In particular, the Shadow from Labyrinth, toned down a bit, becomes just another thing that happens once in a while.

The most interesting randomly-appearing boss-like enemy is the Forgotten. A demon that manifests sometimes as a tentacle monster and sometimes as a woman with skeletal arms, the Forgotten is the main antagonist of the series, but doesn’t appear in the game levels until about halfway through Chasing Shadows. In fact, she arguably doesn’t appear in the game levels even then. She appears to be in some way outside of the game, like the player. When she appears, you just see her silhouette on the screen, as if she’s passing in front of a movie projector, which would place her in the player’s physical space.

Because the Forgotten isn’t inside the scene where your gems and spells have their effects, she cannot be fought. When she shows up, she just takes a semi-fourth-wall-breaking action and leaves. Sometimes she enrages some of the upcoming waves, making them tougher to beat, which frankly never seemed all that bad to me — as I noted previously, I was enraging most waves myself by the end, so when the Forgotten does it for me, all she’s really doing is sparing me a little effort and expense. Ah, but the other thing she can do is fearsome: sometimes she takes control away from the player. For about the duration of a single wave, all the controls are simply removed from the screen and all you can do is sit and watch events unfold. Which you normally spend a lot of time doing anyway, but you usually at least have the ability to spring into action if there’s a sudden need, and she temporarily takes that away.

All special monsters, including the Forgotten, are heralded at least a wave in advance by glitches and flickers, as if their eldritch presence is interfering with the magic you’re using to view the scene. So you at least get some warning when the Forgotten is about to show up and mess with your plans, a trick that the game uses to make you blame yourself for the outcome.

Earlier today, Paolo Pedercini hosted an “interactive movie night” on the Molleindustria Twitch channel, using Twitch’s features to poll the people watching about choices. I of course had to watch. Four pieces were screened: Kinoautomat, I’m Your Man, an adaptation of the shoot from Steven Spielberg’s Director’s Chair, and The Immoral Ms. Conduct. As of this writing, the recorded stream is still available, although obviously not interactive.

I mainly want to talk specifically about Kinoautomat, a black-and-white Czech comedy that first screened at the 1967 World’s Fair in Montreal. This is a piece I had been curious about for some time, but had never had a chance to see. As far as anyone knows, it’s the first work of interactive cinema ever attempted — although later claimants to that distinction can be forgiven for not knowing about it, because it was simply unavailable for decades, having been banned by the Czech Communist Party. In its original form, it would have had a live presenter narrating the choices, telling the audience which buttons to press for each choice. Interactivity was thus seen as something operating on the film from outside, rather than a part of the content. It’s a bit ironic, too, to put “automat” in the title and then make it depend on direct human interaction. In the form presented to us today, however, the presenter is part of the video, clearly a later addition. The original filmmakers could have done it this way — the film was still running while the presenter talked — but chose not to. Call it a UI decision.

In form, it’s very close to linear: despite some pretentions of being controlled by a complex computer (which speaks directly to the audience at the end), the interactivity was originally created by running two film projectors simultaneously and blocking one of them at a time. (Apparently it was also once televised using two channels.) So, your choices can’t affect the sequence of events for very long. This isn’t much of a surprise, though, because the bulk of the story is told in flashback: the whole thing starts with an apartment block on fire, followed by recounting the events that led up to it. Apparently it’s been read as a satire of democracy. No matter what you vote for, it all ends in flames.

The style is goofy and extremely 1960s in its sensibilities — most of the story concerns a respectable middle-aged man dealing with being caught in the company of a naked woman for completely innocent reasons and Everyone Getting The Wrong Impression. That man, Mr. Novák, is the viewpoint character, but interestingly, not all of the choices concern his actions: in the end, the audience is asked to simply pass judgment on him, decide whether he’s culpable for the fire or not, a choice that basically asks you to end any identification you had with him. (Personally, I was strangely disappointed when it turned out that he didn’t start the fire deliberately.) Also, at one point the doorbell rings and the audience is polled for their guesses about who it is. Neither of the options offered is right, and the only effect of the choice is the display of how many people chose what. So even at this early date, designers were using completely fake choices to split up lengthy noninteractive sections.

I feel like the filmmakers showed a lack of confidence in the format when they decided to make Kinoautomat a goofball comedy, unsure about whether audiences would be willing to take interactivity seriously. That’s pretty definitely the case for I’m Your Man, the second “world’s first interactive movie”, which was specifically intended as a test of audiences, and which turns up the goof factor to the point where it’s using cartoon sound effects. I’d also like to note that I’m Your Man doesn’t bind its audience to making choices for a single player character, switching characters freely — indeed, its first several choices are choices of whose story to follow. Film, after all, has no technical constraints limiting who can act on your decisions. And yet Kinoautomat mostly sticks to the CYOA model, despite predating the Choose Your Own Adventure series by a decade.

As much as the details change over the course of the Gemcraft series, there are some things that are curiously constant. In a typical tower defense game, different types of weapon fire differently: you’ll have some equivalent of machine guns and sniper rifles and laser beams, differing in how frequently they fire, how far, whether they do instant damage or fire a slow projectile, whether they hit a single target or everything within a certain range. The magic gems that are your weapons in Gemcraft basically do all of that the same. No matter what the gem, they fire the same sorts of projectiles in the same way. Oh, there’s a little variation: chain hit gems have a longer reach, poison gems have a higher base damage, things like that. But the difference between different types (or colors) of gem is never that great, and it’s completely overshadowed by the difference between grades of gem. Upgrading a gem improves it in every regard: its range, its fire rate, the speed of its projectiles and how much damage they do — and the power of its special effects.

The effects, now. That is what distinguishes the type of gem. Exactly what types are avalable varies a little from game to game in the series, as do the details of what they do. For example, when bloodbound gems were introduced in the third game, they became more powerful the more kills they get. In Chasing Shadows, they become more powerful the more hits they get, which makes it a lot easier to bring a new bloodbound gem into play late in the level. (Bloodbound gems also changed color between games, from red to black. This is the sort of thing you only notice when you play the entire series in a row.)

For most of Chasing Shadows, I found it prudent to have multiple types of gem in play whenever I could. For swarmers, you want red chain hit gems. For heavily-armored giants, you want either purple armor-tearing gems, or green poison gems (because poison damage bypasses armor), or both. I found it effective to have two towers with blue slowing gems: one targeting the enemy closest to the base, as is the default, to make sure the one most urgently in need of slowing gets slowed, and the other set to target at random, to spread the slowness around. There’s cyan gems whose special power is to suppress healing, which isn’t actually all that useful, because things tend to die before they can heal, but I’d gladly throw one of those in once the more essential gems were in place for the few cases where it was useful. And everything could benefit from being combined with a white poolbound gem, which enhances the other special attributes.

By the endgame, though, things were a lot simpler. With sufficient power, I was relying on just two gem effects: orange mana leeching gems and yellow critical hit gems. (Both enhanced with poolbound, of course.) Mana leeching gems are weak in the beginning, doing slightly less damage than other gems for a marginal gain, but if you keep upgrading them, they become your main source of mana, taking in tens of thousands with every hit. Place it in a trap instead of a tower to maximize its yield, and spend most of its mana output on upgrading it. A sufficiently powerful critical hit gem kills everything that doesn’t perish on the mana leeching trap. Critical hit multipliers just keep on increasing as you upgrade the gem, so that eventually it’s got a multiplier in the millions or billions, and fires fast enough to guarantee that everything it hits gets hit by a crit, multiple times. With damage like that, who cares about armor or healing?

You’ll notice that getting these things up to superpower levels requires upgrading them a lot, which costs a lot of mana. So, yes, most of my endgame involves not just two types of gem, but two gems. But I liked to make one exception to this: a second yellow gem dedicated set to target structures, specifically to destroy beacons. Beacons are enemy buildings that aid the monsters in various ways: there are beacons that heal monsters within range, ones that grant them shields, ones that prevent you from building in a certain area, etc. They’re rarities when you first encounter them, appearing only in certain levels. But one of the special powers sometimes found in enemy waves is “spawns a beacon on death”. And beacons give tons of XP when destroyed. So once you’re strong enough, it makes sense to make sure you “enrage” those waves (sacrificing gems to increase the number of monsters) to get lots and lots of beacon-spawning enemies. It gets so that the level is saturated with beacons, popping up as fast as you can destroy them.

In fact, enraging waves in general is one of those things that I didn’t see the point of at first, but which became a key part of my tactics by the end. I mean, spending precious mana on a gem only to give it up to make things more difficult? But more enemies means more enemies getting killed, and also more enemies walking over my mana-leeching traps, both of which mean more mana to spend on getting stronger for later waves.

So, that was where I stand as of the final level. But I do still have some suspicion that it’s not the final best strategy. I mean, look at traps. I used traps a lot in the early part of the game, the better to deliver unblockable poison damage to lots of foes at once. Then I abandoned them for a while as increases in poison damage didn’t keep up with my needs. But by the end, I was making heavy use of traps again for the mana-leeching gems. I can imagine that eventually, as the stats reach even farther into the ridiculously astronomical, I might start seeing armor that blunts even my strongest critical hit, prompting me to bring out the armor-tearing gems again. Things may well be cyclical.

So, I’ve played a bunch of hidden object games this season. And I’ve played a whole lot of Train Valley 2. But the main thing I’ve played, the biggest constant throughout the pandemic, has been Gemcraft ¹Officially, the title is capitalized as “GemCraft”, but I find that less pleasing, ambiguous in how to pronounce it properly. Besides, I called it “Gemcraft” in all my previous posts, so why stop now?. And when I say Gemcraft, I mean all of it. The near-simultaneous releases of a new Gemcraft sequel and a couple of standalone Flash players with bundled games in response to the long-awaited Death of Flash on the web spurred me to try to actually play every game in the series to the end for the first time. This experience has played the same role in my life this year that Creeper World 3 did a couple years back.

I’ve written a few posts before about the fourth Gemcraft game, Gemcraft Chapter 2: Chasing Shadows. The fact that the fourth game is labeled “2” is a little peculiar, but not unprecedented. To recap, it’s a series of wizard-themed tower defense games based around two innovations: the ability to move your weapons around from tower to tower, and press-your-luck gameplay where you can make levels more difficult for greater reward, both before starting the level and while playing it. Once you’ve leveled up a little from the beginning, the only reason you ever fail is overconfidence. Right now, literally between starting this post and finishing it, I’ve played to the ending of Chasing Shadows. This was quite unexpected. A conspicuous gap in the overworld map made me think that I had some way to go yet, but that gap only fills in on victory. It looks like it may be the setting for Chapter 3.

The other chief thing of note about these games is that they’re long. Far longer than is comfortable for my normal binge play-style. They’re really meant to be played a bit at a time over a long period, but even then, you’re going to level up to the point where the challenge is gone long before you reach the end. This is part of how the game tempts you to turn up the difficulty. But it’s also part of the appeal when you’re in a certain state of mind. If I’ve found myself playing Train Valley 2 a lot lately, it’s because it offers a fantasy of control, of making plans and executing them. Gemcraft offers a fantasy of mastery, of not having to put in the effort you once did. Of waving away even the most absurdly overpowered attackers.

It also offers to contradict that. After you win a level, you can keep on going in “Endurance Mode”, where it just keeps on sending enemies in increasing numbers until they finally overwhelm you. Endurance Mode is one of the keys to gaining XP fast, and gaining XP fast is also one of the game’s great joys. When you’re powerful enough, you don’t just work towards gaining levels one at a time, you get dozens at once, the XP bar at the main screen swiftly filling repeatedly, the ding turning into a jingle.

Chara would love this game.

And it seems to know that. The story underlying the series is one of repeatedly being morally compromised, of being tricked into doing the work of demons. The second game, Chapter 0: Gem of Eternity, has you playing the character who will become the antagonist of the first.

I’ll have more to say about Chasing Shadows tomorrow. I may have won, but I’m not done with it yet.

Train Valley 2: Seldom have I seen a sequel so thoroughly change the fundamental character of a game without altering its basic gameplay.

That gameplay consists mainly of laying tracks to join stations. Games based on that idea run a spectrum from abstract puzzle games like Trainyard, to realistic simulators like Railworks, and Train Valley is toward the abstract end of that, but not quite as far out as Mini Metro. You have fanciful toy-looking locomotives on a grid of big chunky squares, dotted with obstacles and color-coded stations. Trains randomly materialize at the stations, and if you get them to their destinations in good time, you earn money that you use to build more tracks to cope with increasing demands on your network and the gradual appearance of additional stations. Your chief enemy is the constraints of the grid: tracks can only turn 45 degrees per tile, and each tile can contain only one switch or crossing. And on top of that, it’s prudent to have redundancy, to keep any trains going from point A to point B from blocking trains departing point B. Sometimes you wind up making a complicated web of junctions to cope with the constraints, and once you have that, it’s very easy to leave something switched the wrong way and send a train to the wrong place. The scale goes together with the art style to make it all look and feel like playing with a toy train set: sometimes the distance between stations is barely longer than the trains running on them.

Train Valley 2 shifts towards realism. Mainly it does this by adding more ways to affect and be affected by terrain. You can build bridges and tunnels, at great expense. There are slopes, which you can build tracks to ascend or descend but not running along laterally. There are steeper slopes that are just plain impassible. This variability makes it easy for the designers to make the kind of congestion that dominated the first game local to a part of the playfield. Distance is now a big problem. The tiles are smaller, or, equivalently, the levels are larger, and trains can take a significant amount of time to get where they’re going. To intensify this, you have a limited number of trains that can be running at one time. You can purchase more, or upgrade them to run faster, but this comes at a considerable cost, which can delay the purchase of essential bridges and tunnels.

Most of all, though, the trains in Train Valley 2 are purposeful. It’s not just a matter of “Train at blue station arbitrarily wants to go to orange station”. Rather, stations are associated with commodities. Each level has one or more towns, which produce workers at a steady rate and which have icons indicating certain needs, like “this town requires 12 copper ingots, 16 books, and 6 airplanes”. Fulfilling those needs is the goal of the level. Other stations will take specific resources to produce others: workers + grain = cows, for example. (Like all resources, workers are absorbed in these recipes.) It’s a bit like Hero of the Kingdom but with trains. And it has a profound effect on how the game feels. The first Train Valley was all about reacting to events. Train Valley 2 is all about planning. The first thing you do on loading a level is scan the map to see where the towns are, and where the resources that require nothing but workers are. The hierarchy of dependencies is like a story, a sequence of events with an optimal ordering that gets you everything you need within a par time.

I haven’t been posting much lately, and what I’ve posted has mainly been looks back at games I played a long time ago. You’d think that being stuck at home due to a pandemic would be a perfect opportunity to play lots of games and blog about them, but somehow I’ve only done the first part of that. As usual, I’m hoping I can get back into the habit of short daily posts.

So, what have I been playing? Several things. For one, I got onto a sizeable Hidden Object Game kick. These are my comfort food, the ludic equivalent of a popcorn movie. They come in a wide variety of shapes and colors, but there’s a sameness to them all. I recall that a bigwig at Marvel Comics once attracted some flak for saying that he wanted his company’s output to be the equivalent of vanilla ice cream. Hidden object games are like that.

And yet, one of the things I really appreciate about them is the variability! Every game does its own little tweaks on the formula, and over time, those tweaks accumulate into an evolution. I remember playing time-limited free trials of the first generation of the genre, downloaded from the Pop Cap website. Those were pure hidden object games, just a sequence of cluttered scenes with lists of things to find. There was a tendency to hide objects by altering their scale, making a pencil into part of a city skyline or whatever. That’s fallen out of fashion. The trend nowadays is more to exploit ambiguity in the object names: a “bow” could be a hair bow, a violin bow, or an archery bow; a “pipe” could be lead or meerschaum. For that matter, it might be a picture of a pipe, which is very easy to just look past of you’re not careful.

(Mind you, not all such word problems are deliberate. These games tend to be produced in foreign countries — one of the best studios making them is Artifex Mundi, which is based in Poland. And this means that every once in a while the object names don’t match English usage. The game might ask for a shovel, but the item you’re expected to click on is clearly a trowel. I assume that this means the designer’s native language uses a single word for both.)

Eventually, the genre started taking on features of adventure games, placing the hidden-object scenes into an explorable space. Big Fish’s Return to Ravenhearst (2008) was a landmark in this regard, pioneering the conceit that fuses the hidden object and adventure genres: that hidden-object sections represent searching for a specific object that you need to solve a puzzle elsewhere. Not that you can just pick up the one object you need if you happen to find it first, of course. And at some point, I don’t know when, some games started putting special-cased interactions inside the hidden object sections: you might be told to find a “lit candle”, and the scene contains an unlit candle and a book of matches, which you can pick up and click on the candle. There’s something oddly recursive about this.

Once hidden object games and adventures had hybridized, the resulting genre started de-emphasizing the hidden object sequences, using more diverse minigames. Some games even make the hidden object sequences optional, by, say, letting the player substitute a game of Mah Jongg Solitaire, which resembles a hidden object game in that it involves searching for things. There are even some adventure games, such as the Drawn series, that I’d put in the same genre because their UI, interaction, and art style all fit in perfectly, but which don’t have any hidden object scenes at all.

So the genre isn’t really defined by hidden objects, although those are a very common feature. Rather, it’s a family of cheap casual-positioned first-person adventure games marked by easy puzzles, frequent mini-games, full voice acting but limited animation, and a very distinctive art style: it’s a painted look, but with sharp focus and lots of rim lighting, and it’s distinctive enough to come as a surprise on the few occasions when I’ve encountered the style elsewhere. Often the UI uses lots of particle effects. Usually there’s an inventory bar at the bottom of the screen with a map and/or journal on the left side and something you can click on to get hints on the right. It’s all quite formulaic.

The stories, too, are mostly pretty formulaic, in a pulpish way. Usually there’s someone or something you’re pursuing, usually because they’ve kidnapped someone close to you, either in the opening cutscene or about a third of the way through the story. That’s your call to adventure. It’s almost the standard videogame Princess Plot, except that the player character is very often a woman. I’m tempted to say usually a woman, but this may be perceptual bias. At any rate, it’s a template that can be fit over basically any genre, and the designers exploit this by fitting it over as many genres as they can. There was a tendency for the early pure hidden object games to be about detectives, presumably because that was considered to be a reasonably good fit to the gameplay. There’s still a fair number of detective stories in today’s hidden object adventures, but you can just as easily be given the role of an elven hero, or a steampunk adventuress, or an ordinary woman who was minding her own business when her daughter was abducted by ghost pirates.

At any rate, there’s a zillion of the things, and they frequently get swept up into huge bundles. Usually I can finish one in about a day, or a single long evening — maybe two, if I get stuck. But I’m in no real danger of ever running out of them. They’re just always there, for when I need the comfort of the familiar.