Probably Dance

I’ve had an obsession recently with a mod for Warcraft 3. It’s called Survival Chaos. I want to talk about it because it’s part of a genre of strategy games that hasn’t had a big success yet, and this feels like a big evolution, maybe even a breakthrough. It’s rare to see a new video game genre emerge like this, and nobody ever writes about this while it’s happening. The history of Auto Chess, the other recent genre to come out of Warcraft 3, is almost completely lost. (I was able to find a very similar map called “Pokemon Defense” from 2010, but that’s about it…)

I am not sure if the genre has a good name yet since it’s never been big. In StarCraft 2 it’s called “Tug Of War” so I’ll go with that. The basic idea is to make a RTS where you don’t control your units. You just build buildings, the buildings automatically make units, and you watch the units fight automatically. You mostly make decisions about the macro: When to invest in your economy, what units you should invest in, what upgrades you should get. Before we go any further though, let’s just watch a video of someone playing the game:

Skip to 2:40 in the video. This is from a live stream. At first not much is happening, other than the choice of which “bonus” to go for. (Phoenixes in this case) Otherwise you just have a couple buildings and you can decide which buildings to upgrade or which upgrades to research. We also see early on that you can do a lot of manual defending: You can control which enemies your towers shoot at, your central building has a defensive spell to damage incoming enemies, and you can buy extra units to help defend your base.

But mostly you just spend a lot of time observing how your various fights are going. Are you winning in the middle? How fast are the side-lanes progressing or retreating? What are the enemies upgrading? This is actually the part I like most about Survival Chaos: You have time to think about what to do. It’s got the slow pace of a turn based strategy game but it’s real-time nature does keep the game going. And when you’re defending your base, it has the same fast gameplay of other real time strategy games, but only for brief moments.

The other big important parts are the heroes (which you also don’t control) and you hear the streamer in the video talking a lot about “ultimate weapons.” The ultimate weapon is supposed to be an ability that you can research at the end of the game to end it more quickly. Unfortunately one of the many things that Blizzard broke in the Reforged upgrade is research requirements, so you can now get ultimate weapons much earlier, that’s why he is talking about them so much. I don’t think anyone uses them in this video, at least not before you’re supposed to, so I’m just explaining it in case you’re wondering if you’re missing something important. (but I could write a whole separate blog post about how Blizzard needs to treat the Warcraft 3 ecosystem with more respect and not just introduce thousands of bugs into mods. So many new video game genres came out of Warcraft 3 that I would consider it important for video game history. The long stability of Warcraft 3 was important for that. It’s not OK to damage so much video game history. It’s unknown at this point how many old mods got major bugs like this from the Reforged upgrade, but I know of two major bugs in just Survival Chaos (early ultimate weapons, invisible neutral buildings), several medium bugs (missing race-specific terrain colors, broken custom models, broken effects, broken shadows) and many smaller ones (wrong unit scaling, wrong model for Furion, stretched loading screen…), but I digress too much…)

Why it’s Fun

At this point some of you may already get the appeal without me having to explain it, (fans of Dungeon Keeper or Majesty probably get it immediately) but for those who don’t, let me try: A lot of people don’t like the micro of strategy games. “Micro” is the part where you control units during fights. The part of RTS controls that gets messy and stressful once you have more than five units. All of the previous big genres that came out of Warcraft 3 came out of the motivation to reduce the amount of required micro: Tower Defense and Auto Chess require no micro at all and in DotA you only control one hero, allowing you to actually get good at microing. Previous games in the Tug of War genre also tried to eliminate micro entirely, but Survival Chaos keeps a little bit: When defending your base you get to control your buildings. But for most defenses you only need to control four buildings: Your main building, a barracks and the two towers facing the attack. And those buildings can’t do much. Mostly you just have to get good at “redirecting” heroes because heroes are more likely attack the building that’s shooting at them.

And then since you don’t have to micro much, you can entirely focus on the macro, just like in many turn based strategy games. And the macro in Survival Chaos is surprisingly layered:

The Macro of Survival Chaos

The first interesting trade-off you encounter comes from you having three lanes. In the side-lanes you are only fighting one opponent, in the middle lane you are fighting all three opponents. You get gold for killing enemy units. The side-lanes are zero-sum games in that regard: Any units you send will give the opponent on that lane more gold, any units they send will give you more gold. But the middle lane is a non-zero-sum game. If you can dominate the middle lane you can get gold from all three opponents at the same time, and if you do badly there you end up just feeding the other players, getting no gold yourself. But if everyone focuses on the middle lane, and everyone does roughly equally well, it turns into a zero-sum game and nobody gains any benefit over anyone else. So to get the most income you have to balance which lanes to invest in.

The full strategy of which lane to invest in at what time is complex enough for me to talk about a lot longer, but notice how even within the game of “balancing three different lanes” there is a sub-game of “trying to dominate the middle” which in itself is big enough that I could write whole paragraphs about how to play it. You have to invest resources and know when to send a hero. If you just throw money at the problem a more skilled player might be able to outmaneuver you with better timing. Or you might “push through” too quickly and see your investment being fed to the player on the other side. And if you see another player investing lots of money and you’re able to thwart their investment repeatedly, that can be really rewarding as well.

And of course there is the higher level game of you having to play against three other players and having to decide who to focus on. If a player gets too strong they can snowball the entire game, so if you’re currently in second place, maybe you’ll go easy on the players in third and fourth place so that they can help you overwhelm the player in first place. But the moment you sense weakness in either of the other players you have to strike to get the gold from killing another player, because that might allow you to catch up to the first place. (also you have to prevent the first place player from getting that gold) The players further behind have to play different strategies to not get wiped out, and the player in first place can have the most stressful game of all if all others attack them, so they have to find ways to de-escalate against at least one player. It’s a complex set of interactions that’s different in each game, depending on the personalities of your opponents.

And then there is the game of which heroes to play in which situations. Some heroes are great for sending in the middle, some are great for killing other heroes, some are great for defending against giant waves of enemies, some are great for dealing lots of damage to an enemy base. And all of this has a lot of uncertainty since the heroes have a bit of a mind on their own. If you don’t take the uncertainty into account you might easily lose the investment in a hero by sending one out only to die quickly. But in a game that goes on long enough, being good at using your heroes will win you the game.

The better you get at Survival Chaos, the more of these games-within-a-game you have to be playing at the same time. You start doing trade-offs that you wouldn’t even have considered as a beginner (in the game of “defending your base” it’s sometimes OK to trade off building health to save gold while defending, where a beginner player would never consider building health to be something you can trade off against anything) and you end up discovering more and more layers to all the other games you’re constantly playing.

Most of the time the actual “gameplay” is just watching, planning, predicting. You see how your battles are going, you’re trying to figure out what your opponent is doing, making plans of what upgrades to get yourself, oh and you better get ready to defend because there is a huge wave of enemies getting close on one of the side-lanes. You usually have multiple options to defend an attack and you also need to think about which of those options to choose. Then when it gets to the actual defense you’re busy for a minute, then it’s back to watching, planning, predicting to figure out how things are going and which investments you need to make.

Estimating “Depth” of the Game

The main benefit of Survival Chaos over other games in the genre is its depth. It’s hard to measure the “depth” of a game but one thing you can do is measure how high the Elo ladder is for a given game. The name comes from the Elo rating which allows you to compare how good two players are in a game. Let’s say your average player wins against a bad player 90% of the time. A good player wins against the average player 90% of the time, an expert wins against that player 90% of the time. How many times can you do that? Can there be a “master” level after that that wins against the “expert” player 90% of the time? Can there be a “grandmaster” level after that that wins against “master” 90% of the time? In very simple games like Tic Tac Toe there are only two rungs to this ladder. (a good player wins against a bad player 90% of the time, but then there can be no “master” level player who wins against the “good” player 90% of the time. Instead you get 100% ties) “Connect Four” might have a few more, checkers has more still and chess has a lot of rungs to that ladder.

Survival Chaos has a lot of randomness and unpredictability so it won’t have as many as chess. The community is small enough that I would have no idea how to measure this, but based on my personal experience I estimate there to be four 90% steps. It took me more than 50 hours before I won my first game. But since you’re always playing against three other players you can still measure your progress during those 50 hours: At first you finish in last place or third place, then you end up making second place reliably, and then it still takes a big step up to actually win. Once I had won my first game it took a while before I could win 25% of my games, which is what you would expect for an average player in a game with four people. I didn’t know if I could get much better. There just seemed to be too much randomness and uncertainty. Now I’m over that and if I play against random people on Battle.net, I win most of my games. I still get screwed over by randomness all the time but now I know how to react and I make sure that I can’t get screwed over too much. But if I play against people who play in tournaments, I still lose a lot.

There is also enough depth to the strategy that I would still see new strategies even after more than a hundred hours of play. Often highly creative ones. There are 14 factions in the game, each of which has 12 different bonuses. Some of the bonuses just give you a stronger unit or hero, some are just for fun and some are for trolling. But there are also some bonuses that can completely change the game when used creatively. (like early suicide barracks on the side lanes) Even for the normal bonuses, the sheer amount means that here is lots of replay value in just trying all the different ways the game changes. (14 factions multiplied by 12 bonuses multiplied by 1 hour for your average game means it’ll take you 168 hours just to play each bonus once with each faction)

How to Play It

If you want to play the game, you need to have Warcraft 3. The author of the map recommends going with the classic version, not Reforged. The custom models in the game were made to fit well with the normal Warcraft 3 models, so they look out of place in Reforged. The gameplay is the same and you can play with people who have the other version.

Then download the map from the website or find a game on Battle.net to get the game downloaded automatically. You can usually find a game on there, or if you host one yourself it will fill up within a minute or two. Before you do that I might recommend playing a game against computers a few times just to get the hang of it. If you can beat the “insane” difficulty computer, (which cheats to have more gold than you do) you’re more than ready to play online. Just know that you might learn some bad habits while practicing against computers since they have different weaknesses than human players, so maybe start playing online once you can beat a “normal” difficulty computer.

There’s also a mod for DotA 2 called “Survival Chaos: Devotion“. It’s a clone made by different people and far less finished. I haven’t played it, but I hear it’s not as good yet. But it’s free to play since DotA 2 is free. So if you just want a feel for the genre, you can try that one.

Evolution of the Genre

I also want to talk about the “Tug of War” genre and how it evolved. You might look at Survival Chaos and say “it’s like DotA, but you control the base instead of controlling the heroes” but that was actually tried fairly early on and it doesn’t usually work too well. Well lots of things have been tried. I’m going to skip over all games that are like “DotA, but you can also upgrade the units.” Those all don’t work very well because if you can control the base and the heroes, the heroes always dominate. The gameplay ends up revolving 95% around the heroes and the upgrades are a total afterthought. Sometimes games try to force you to care more about the upgrades, like “War of the Races” but if you can only control that one unit, your focus will be on that one unit. It’s an unsolvable clash.

So instead for most of it’s evolution the genre has looked a lot like Castle Fight. Which simply has no heroes and instead it expands on you building your own buildings and deciding which units you want. I don’t know if games like it already existed in StarCraft,  but in WarCraft 3 they’re old, at least fifteen years old. Here is someone playing it:

This map is always popular and is fun to play, but it has one major problem: once one side has a slight advantage, it often snowballs and the other side can’t really do much to defend itself. To solve this each player has one “rescue strike” that kills a lot of enemies at once. This resets the progress and allows your side to snowball the other side. Unfortunately this often means that the game just turns into “left side snowballs right side, then right side uses rescue strike, then right side snowballs left side, then left side uses rescue strike, then left side snowballs right side…” There is more strategy to it, but among two evenly matched teams it often degrades like that. Once each player has used their rescue strike once, whichever team snowballs next wins.

How do you add depth to something like that? Castle fight has some depth with items you can buy and some units countering other units, but it all doesn’t really help. The snowball dynamics are too strong. Another map, Preschool Tech Wars, tried to add depth by having a very complex upgrade tree. The first time you’re playing, it is interesting to try to figure out which building requires which other buildings to upgrade. It also makes towers much stronger and that cuts the snowball short. (there are towers in Castle Fight but they’re expensive and don’t help against the really big snowball waves) It also adds heroes that you can control late in the game. Here is somebody playing it:

The problem this map has is that it always turns into a mess. Since you’re able to stop the snowball by building towers, the game keeps on escalating until it turns into an unstoppable mess. This map was always less popular than Castle Fight because it adds complexity but runs into similar problems in the end. In terms of the Elo ladder that I talked above, Preschool Tech Wars might just have two rungs, like Tic Tac Toe. Because the outcome is such a mess that once you learn the basics, nobody is able to win against anyone else 90% of the time. Castle Fight is better in that regard.

So how else could you evolve this genre? If you know anything about it you’re probably screaming at me because I haven’t mentioned the biggest game: Direct Strike in StarCraft 2. It was one of the only two paid mods for StarCraft 2. (it was a thing Blizzard tried) Here is someone playing it:

To briefly explain how it works: Instead of building buildings, you have a board on which you directly place the units. Then the units get spawned each time that a wave starts. Also there are three players on each team, and the waves spawn one at a time from each player. So if you are player 2 on your team, your units will always spawn against the units of player 2 on the opposing team.

How does it solve the problems that Castle Fight has? Mainly by tightening everything up.

1. Counters are much stronger, so if you see an opponent do something, you can build the counter and it will actually make a difference. (in Castle Fight you also kinda can, but since the counters aren’t very strong, they get lost once the snowballing starts)
2. The lines are ridiculously short. Your units get to the other side very quickly. This prevents the snowball from building up. (also there are towers on the other side that end waves quickly)
3. Waves come all at once. Combined with the short lane, this means that the fight for the previous wave is usually over by the time that the next wave comes through, once again reducing the snowball size.

The fact that there are strong counters also adds lots of depth. The one downside is that it’s a kind of depth that requires memorization. You have to memorize which unit beats which other unit. Survival Chaos also has some of that, but most of the time when you get better, it’s because you get better at the systems, not because you memorized what to do in what situation.

Survival Chaos’ Contribution to the Genre

So how did Survival Chaos solve the problems that the genre has always had? The biggest fix is actually that you don’t get to decide which units spawn. A tier 1 barracks always spawns two melee units, one spellcaster and one ranged unit. You can only change that by upgrading to a tier 2 barracks. But you can’t build a second barracks or five more or decide that you want three more spellcasters, like in other maps. This immediately solves the problems of the waves turning into huge messes. I think other maps never had the courage to limit you like this because deciding what units to build was literally all you did in those other maps. In Survival Chaos the upgrades are more important instead.

The second thing that really helped was giving you more tools in your defense. You can cast spells, summon additional units, control your towers and you can summon heroes. It’s not a huge variety of options, but it’s enough to solve the problem of ever-escalating waves that you can’t do anything against. It also helps that your barracks is the most forward-facing building. If you lose it you’ll be permanently crippled, but you’re not out of the game. (and you can definitely still win with just two barracks) This forces you to actually invest in your defense. In other maps you had no failure spectrum like that. Instead you wouldn’t lose anything until you lost the whole game. This meant that defending wasn’t worth it, because it would only make sense if you were able to defend against a game-ending wave, which you’d never be able to do. In Survival Chaos a fairly small wave can destroy your barracks early in the game. It’s cheap to defend against and the downside of not defending is huge. So you’ll invest money in your defense. This balance is only possible because while the downside is huge, it’s not game-ending. (some players don’t know that though and leave out of frustration when they lose one barracks. So the game could potentially be improved further by adding more stages to this failure spectrum)

Another really big thing that helped is that there are four players. Previous games in the genre had tried to add multiple lanes but it never quite made sense when there was only two teams fighting against each others. Having three lanes with different characteristics adds whole new layers of strategy. The fact that there is a non-zero-sum lane in the middle is what made heroes really viable. If you added the heroes of Survival Chaos to Castle Fight, they would only be useful as a defensive tool. Otherwise they’re just a huge investment that is guaranteed to lose you money and to feed the opposing team. But in Survival Chaos they have a lot of depth to them. Not only can you use them to try to win the non-zero-sum game in the middle, but even if you only use them defensively on a side-lane (like you would in Castle Fight) that may be worth it because you’re not just in a conflict with one other player but with three other players. Pressuring one player a lot may just knock them out, or it can buy you breathing room which you can use to focus on other battles. All these considerations don’t come up in Castle Fight or other previous games where there is only one fight.

Will this Genre Make it Big?

Let’s say I have convinced you that this is an interesting genre and that Survival Chaos is a big new evolution. I said at the beginning that Survival Chaos might be a breakthrough for the genre, so do I think it’ll suddenly become very popular, maybe the next DotA or Auto Chess? I’m not sure. This genre has been around for many years and has always had a steady popularity, but it was never as big as DotA or Tower Defense. But then again neither was Auto Chess, and that also took some evolution before it got really popular.

There have been previous attempts at turning this genre into standalone games. Here and here are two examples. But those all followed the Castle Fight/Direct Strike model. So they all died off, probably because they couldn’t have enough depth.

How big is the potential audience? What comes to mind is people who like playing Dungeon Keeper or Majesty or Kingdom. Those games were popular, but never huge. But they also never had a strong multiplayer, and Survival Chaos shows a model for how to make strong multiplayer for games like that.

So overall I’m cautiously optimistic. I have always enjoyed the genre, but Survival Chaos is the first time that I found a game that I can really invest time into. And that is a major change. If you’re into strategy games, I definitely think you should check this out, simply because new strategy genres don’t come along all too often. I do think something bigger will come out of this genre eventually. It’s good to have a type of RTS game that only focuses on the macro. For me, and a lot of other players, that’s what they really want to do in an RTS.