Hey everyone! Today I have a good-old-fashioned formalist-ish game design article. It’s been a little while since I’ve really done one of those, unless it was attached to Push the Lane.
This article is also a little bit different than a lot of my other work because I usually talk about rulesets: what the actual rules are. I tend to talk less about, within a set of rules, what players can do. Today, I’m talking about designing strategy space, and a specific way to think about the strategies that players can pursue in your game.
If you’re into strategy games, you probably at least loosely know the basic idea behind “rushdown” (or “rush”), “economy” (or “econ”), and “defense“. A lot of us first heard these terms in RTS games like StarCraft, wherein the “zergling rush” was a very common and easy-to-understand manifestation of a “rush strategy”. Terrans building a ton of bunkers and missile turrets and siege tanks was a pretty clear example of “defense”, and expanding (getting another base with another source of minerals) was an “economy” play. In some games, it can be seen as a triangle, or rock-paper-scissors relationship, with rush beating econ, econ beating defense, and defense beating rush. It’s worth noting that “rushdown” is not, itself, a strategy, but rather a family or style of strategies in a given game. There may be many different rushdown strategies. Also, it’s spectral. You may pursue a strategy that’s like 60% rush-y, or 80% rush-y, etc.
Those are useful enough terms in StarCraft, at least with regards to playing and perhaps balancing StarCraft. But in my experience, it’s difficult to apply that way of thinking about strategy to new games we design. And it’s much weirder to apply it to single-player strategy games (which, as you may already know, I’m most interested in).
When I’m introducing these terms to someone, I find that the best one to start with is rushdown. It tends to be the easiest to explain. Here’s Giant Bomb’s description of “rushdown”, which I think accurately describes the way most people think about it.
Rushdown refers to a both a general philosophy across all fighting games as well as a number of specific, aggressive strategies and play styles in particular games. It can be considered the opposite of turtling. The general goal of a rushdown player is to overwhelm the opponent and force costly mistakes either by using fast, confusing setups or by taking advantage of an impatient opponent as they are forced to play defense for prolonged periods of time. Rushdown players often favor attacking opponents in the corner or as they get up from a knockdown; both situations severely limit the options of the opponent and often allow the attacking player to force high-risk guessing scenarios.
Rushdown characters are often fast and require a certain degree of physical dexterity and educated guesswork to use effectively.
While I agree that this is how people see it—and I agree that it’s how it manifests in most games—there’s a lot of stuff in there that is extremely problematic from a Clockwork Game Design point of view. More broadly, I’d even just say that for “new strategy games that aren’t copies of existing designs”, this description will cause problems.
I just went through and bolded all of the problem areas in that quote, and all of that sounds really terrible. And the other two strategic prongs aren’t much better; after all, they’re just the other sides of this crappy triangle.
Also: Giant Bomb does not have an entry for “econ” or “defense” in the way that it has one about rushdown. This may support what I’ve long suspected: that in most gamers’ (and hence game designers’) minds, there isn’t a very formalized or well-understood strategic “triangle”.
A better triangle
I completely wrote off this whole way of thinking for many years, because I only saw it appearing in games that I was generally skeptical of. It seemed to me that it ended up being a dumb rock-paper-scissors thing a lot of the time, particularly in a game like StarCraft. “Oh, they were going for a rush, and I was going econ, so now I lose.” “Oh, they were going for a rush, and I was going defense, so now I win.” (Specific to StarCraft: granted, there’s scouting, wherein players try to extract information about whether the player is going to play rock or paper or scissors, but it can be somewhat random sometimes whether your scouting happens to find that key hidden barracks in the middle of nowhere or not.)
For a long time, it wasn’t a part of how I thought about strategy game design. Around 2010 or so, David Sirlin and his community were formalizing this idea into something they called “the triangle”, which manifested really clearly for me in Puzzle Strike.
I’m not at all sure that Sirlin or his community were necessarily the first ones to formalize the language in this way; in fact, I’m pretty sure they weren’t. But what I do think was unique about Puzzle Strike was that it took this Eurogame type of strategy game—this engine-building machine thing—and applied the triangle to that.
In Puzzle Strike, someone “going rushdown” or “going defense” is very clear, and you can always respond almost immediately. What this means is that instead of this big “guessing game”, you have a careful dance of both players rotating slowly around the triangle.
You went a little more rushdown, so I’m going a little more defense.
You went a little more econ, so I’m going a little more rushdown.
Or, even more interestingly, bending the limits of the dance, like:
You went a little more rushdown, but I’m confident that I can survive it, so I’m gonna go a little harder econ.
It’s true that this kind of thing does exist in something like a StarCraft somewhat, but so many other things (execution, guessing, and all kinds of other random stuff) get in the way. By the way: I recommend reading this thread on the Puzzle Strike triangle.)
Early in Auro‘s development, I started trying to implement an early version of my own take on this strategy design model. It didn’t work. At the time, I wasn’t sure why, but now I know it’s because Auro was never a strategy game to begin with. Auro is a pure tactics game, all about making short-term decisions repeatedly, with very little in the way of medium or longer length strategic arcs. The triangle, on the other hand, is a description of strategy, not tactics.
I had a few prototypes ongoing which played with the idea, but it really wasn’t until about 2014 or 2015 that I really formalized how my version of the triangle would work. The Puzzle Strike interpretation, while better than versions which came before, still didn’t really make tons of sense when applied to a single player game. Or, at least, it’s not at all obvious how someone could go about applying it to one.
Dinofarm’s currently-shelved Alakaram, which is the strategy-game version of Auro, as well as my own Push the Lane, are the two latest and best examples that demonstrate how I intend to express this stuff in a single-player game. But much of this theory developed during the 5+ year long design phase of Auro.
It’s cool to look back at a lot of my strategy-design charts for Auro. That process is what led me to a lot of what I’ll be prescribing in this article. Take a look at these charts!
These above two show an early idea that I had, which was that there was a “countering” system, wherein the game would give you more flying monsters (“Next Monster” shows a new monster type that’s added to the map generation), and you’d “respond” by choosing more anti-air spells, or the game would give you more heavy monsters, and you’d counter that with anti-heavy spells, and so on. Looking back at it now, it’s obvious that this flat “pick the counter” thing is uninteresting and obvious.
(Of course, Auro is a tactics game, but bear with me, because there are lessons here that would absolutely apply to a single player tactics game.)
This experience mirrored my StarCraft experience, wherein enabling “direct, hard counters” is bad. But it also is bad because the player is simply reacting, and reacting isn’t making decisions. The player can’t “go econ” or “go rushdown” or “go” anything. They just have to keep picking the best available stuff based on what’s happening with the monsters.
Another one of the questions it brought up was: is this just an inherent problem with single player strategy games? Is the answer doing some weird thing where the computer also responds to your picks? Like if you pick an Air spell, next time it will pick something that counters Air? That sounds functional, but it still sounds kind of flat. None of it sounds like strategy to me; none of it sounds particularly rich soil for strategy, either.
Here’s another chart that implies the same thing, but I tried to add in some secondary soft counters in the middle. (Man, I did so much game design for Auro—I have dozens of these things.) These ultimately didn’t change anything about the situation, and eventually I dropped the entire concept and just made Auro a tactics game: gave Auro tactical spells, and the monsters tactical abilities, and everything kinda worked out.
But all the prototypes I’ve worked on since then have been attempts at strategy games. So what’d I do next?
A much better (non-)triangle
The new “triangle” probably shouldn’t be thought of as a triangle at all. The problem with the triangle, as a concept, is that it’s inherently based around two or more players (usually two) and some kind of zero sum interaction with them. It’s basically all based around Rock, Paper, Scissors, which is not a strategy game. You did this, but I did this, so I win and you lose. This is not fertile soil for a strategy game.
And it is also problematic for single player strategy games. One of the ways that my general game design philosophy is different from other people’s is that I am very much centered on “the player, their experience and their decisions”. That is to say, when I design a game, I am thinking about ONE INDIVIDUAL’S EXPERIENCE, whether it’s a multiplayer game or a single player one.
One manifestation of this is that I see “other player’s input” as a source of randomness, whereas other people, it seems to me, think of two-player games as like this together-soup of interaction where both players are together engaged in something. I do think this is a valid way of thinking, and a great way of thinking when it comes to social games and party games, and I want to make some of those myself. But I think for strategy games, we should think entirely about “the one singular player and their experience”. The information that hits them and their ability to make creative decisions is all that matters – what the other player is doing, thinking, or whatever doesn’t matter for this player.
Rushdown is going for an early-game win.
Econ is going for a mid-game win.
Defense is going for a late-game win.
The pink line represents the computer’s “power” – their ability to influence the game state, over the course of the game. For now let’s just think of that as ATTACK DAMAGE, for simplicity’s sake. The pink line could also represent other players’ power, if it’s a multiplayer game. For simplicity’s sake, the pink line is moving up linearly, which is how I tend to at least start off my “computer power escalation”. Classic example is something like, “every 10 turns, all enemies gain 10 attack damage” or something like that. The gold line is the player’s power.
(This article also assumes that all games are moving towards a conclusion at all times. Design your games in such a way so that there is some relatively hard limit on the overall length of a match.)
The basic idea is that winning the match happens when you spend a significant enough amount of time with your power above that of the opposition. Here is “rushdown”. In this one you can see that the player gets a huge spike in power early. This starts falling off pretty quickly, and the rush strategy works if they can win before the power spike falls off. You can see in the dotted line how it would go for the rush-er if they fail to win in time. The dotted line shows a failed rush strategy.
This shows a DEFENSIVE game, one in which the player is going for a late-game win. Their objective here is to stay near the line for as long as they can until the opponent runs out of resources. In this illustration, it shows the player’s power just underneath the pink line, but it could go a little further down sometimes, as long as it doesn’t go too low, for too long – that would be a failed defensive strategy. Note that there would probably have to be some kind of “finite” resources in a game for this to work, so I advocate that all strategy games have that. There has to be a point where the resources run dry. Put in the most extreme way, if both players go defensive, there need to be forces which bring the game to an end anyway; one that both players can kind of see coming and play around.
Here is econ, which looks a little bit like rush, except there is a notable dip in the early game, with a mid-game payoff that results in a win. Notably, if it doesn’t win with its power spike, it should lose soon after.
In practice, things are almost never this cut-and-dry—or at least, we should try to design things in such a way so that it’s not so cut-and-dry. Also, the power-level of a given player at any point should itself be somewhat ambiguous, since even stuff like “where the player is at which specific moment” does dictate, to some extent, the player’s power. As the system itself is unsolved, the known power level should too be unsolved.
However, if we had a Laplace’s demon or something that could really know the actual power level of a character, it would look more like this:
It’s a careful dance, and the individual arcs shouldn’t be very long. Think of it more like, there are loose trends in a noisy graph. Also remember that the pink line could also be (and probably should) be more complicated. In Push the Lane, different difficulty spikes happen at different (slightly random) times, resulting in a slightly different line shape each game, but a same, or similar amount of difficulty overall.
A temptation is to say something like “econ is vulnerable early game to rushdown”, and connecting them in that sort of a rock-paper-scissors way. That may be in-effect-true, but I would caution against thinking of that as the defining or important factor. Try to think of that as a secondary quality. The important thing is not that a given strategy beats another strategy, because players will rarely be committing that hard, for that long, to a specific strategy anyway.
A way to summarize what I’m saying here is just this: players shouldn’t be “going rushdown” or “going defense”, really. If such a thing is noticable at all, they should be post-hoc analysis of general trends. Instead, every game is some mixture of all three. All games will end in either the early game, the mid-game, or the late game, and these strategies are loose attempts to target those areas for a win. Players can plan somewhat around a somewhat localized rush/defense/econ, but they must adapt frequently, and take advantage of opportunities as they arise.
Why it’s useful
This design pattern is super-useful to me. Actually, I find it hard to get away from the concept of “early win, mid game win, late game win” as a basic strategy-design model. In fact, I think these ideas, the idea of “investing”, to be fundamental to strategy games. And I find that the Eurogame-style of “race to VP” tends to cause a lot of problems. These problems are somewhat beyond the scope of this article, but to put it concisely, if you have a game that ends at turn 200, or something similar, there is just always a point that all players need to switch from investing to cashing-in. This can be seen clearly in Eurogames and particularly in a card game like Dominion (in fact, you can learn a lot about this from comparing Dominion, which is a race-to-VP, to Puzzle Strike, which has a dynamic match length).
Using this model also makes designing a single-player strategy game make a lot more sense. It doesn’t really make a lot of sense to implement something like the StarCraft-style “rushdown” into a Rogue-like, or even a 4X strategy game, really. For me, it helps so much to think of things in these terms, though. Here’s a recent (yet still desperately in need of an update) spreadsheet for Push the Lane. I encourage you to ignore the details because they’re all already wrong!
This shows a loose way you can define the computer’s powers (in this case, the spawn-types and their abilities and behaviors) around general rush/econ/defense parameters. In PTL, “rushing” for the computer is simply trying to kill the player, or kill the player’s central base itself. Defense would be disrupting the player (simply slowing him down, either by stunning, creating a lot of obstacles, or being hard but necessary to catch). And econ would be destroying towers (this one, I’m the least sure about).
Anyway, the point isn’t exactly the details of how I’m implementing this in Push the Lane, but rather just to demonstrate how it could be implemented. Here’s a couple more quick examples, if you’re interested.
- In Puzzle Strike, econ is buying money, rushdown is combining and adding more gems to the gempile, and defense is counter-crashing and disruption.
- In Dinofarm Games’ upcoming game, Jelly Bomber (a Dr.Mario-like), econ is getting gratuitous combos (combos that don’t get you closer to the goal) to gain super-power meter. You can also make combos with Apricots which give you super-power meter, but create unbreakable PITS in that spot.
Defense is comboing out anything near the top (general Tetris-style defensive play), as well as comboing Blackberries, which causes the jam-line (a line pushing all pieces up towards the top of the jar) to settle a bit faster.
Rushdown is going straight for fruits (which are Viruses in Dr. Mario), especially Strawberries, because Strawberries give you special sugar-pills which make it much easier to do more fruit combos. However, Strawberries also create more jam, increasing the jamline faster.
Hopefully these examples demonstrates how this way of thinking is a useful tool in the toolbox of a game designer. If you use something like this, tell me about it in the comments below. Thanks for reading!