When fighting fair doesn't mean being equals: asymmetry in game design.

Well, Jacob got me thinking about meaningful choices in games, and then I asked him about weapon placement within a level--and wouldn't ya know it, but then I got to thinking about other things!
Amazing world we live in, where things make us think about things.
Namely, I started to think about meaningful choices before you start playing a game. As Jacob explained, meaningful choices make games more fun--but sometimes, the choices you want to give to players don't work well together. Say that in UT99 having both a rocket launcher and a flak cannon at the same time is too powerful--do you remove the rocket launcher? Maybe the flak cannon? But they're both so friggin' awesome--what do you do?
This is where asymmetry comes in. In asymmetrical games, players are not given the same choices during the game because of a choice they make before the game begins. For example, I might choose to play as the Protoss in Starcraft II, which will give me access to the most powerful, armored, and costly troops in the game--but I'm trading those advantages for the sheer numbers of the Zerg Swarm and the long-range artillery and highly defensible positions of the Terran Dominion. Or consider Magic: The Gathering. Because players are allowed to create their decks before playing, there are a near-infinite number of decks available, which means no two MTG games are likely to ever play out the same way.
Now, let's clear up a misconception--asymmetry is not the same thing as chance. In Scrabble, when players first fill their letter racks, they will likely draw different letter tiles from the pile, which will give them different options for which words they can make. However, the player did not choose these tiles; it was chance that they drew these specific tiles. In asymmetrical games, players choose to have certain options available to them, which also means some options will be made unavailable to them. Let's imagine a Scrabble variation in which players can choose to have sole access to a handful of consonants. For example, I might choose to have access to the R, L, and N tiles, but that would keep me from ever getting an S or T tile. Makes Scrabble a little more fun, doesn't it?
Well, maybe. It'll be different, and probably interesting--but fun? "Fun" competitive games are balanced. I chose those letters because Wheel of Fortune told me to, but I'm not sure that having access to R, L, and N, but never getting an S or T is balanced. This is one of the major problems with asymmetrical games; balancing the advantages and disadvantages can be immensely difficult. By their asymmetrical nature, we're comparing apples and oranges. We need to determine how much winning potential each fruit has, and then add an apple to the scale, or make the oranges sweeter, or put the apples up on higher branches so they're harder to get. Supposing we get R, L, and N tiles, and our opponent gets S and T tiles (remember: we can't get our opponent's "claimed" tiles), and all other tiles are otherwise available, we might try to balance the game by determining how many words use R, L, or N in English, and then compare that to how many words use S or T. This would give us a sense of how many options each player gets, and if the number of words is equal, then the game should be balanced. Right?
Well, maybe maybe. In our list of R, L, and N words, we should probably exclude any words that use an S or a T, because our player cannot make those words.
Well, maybe maybe maybe. They can make those words supposing our opponent has already played an S or T on the board. However, most asymmetrical games do not allow a player to ever get the options they chose to remove from their arsenal, so Scrabble's a bit wonky in this regard. Instead, let's say you can never even use a letter that's been removed from your options, so I can't ever draw an S or a T, nor can I play any word with an S or a T in it, even if my opponent's played one on the board already. Lain, yes. Slain? STFU.
So, now that we squashed that loophole, the sides are each unique and balanced, right?
Well, maybe maybe ah hell--they might be, except for point values. Yes, all 5 letters that players have special access to are worth only 1 point, but this may actually hurt the RLN player. That player is now more likely to draw low-scoring tiles than his opponent, which is a small detail--but it could end up consistently giving the advantage to the ST player. Of course, letter distribution is also varied in Scrabble, so maybe they're already even (R+L+N tile total = S+T tile total). Unfortunately, nope--looks like it's (6+4+6 /= 4+6). Sad! The first player has 90 tiles available, and the ST player has 84 tiles available. To get the average tile value for each player, we can do a little harmless math. Let's take the total number of points of all tiles in the Scrabble set (187) minus the values each player does not get:
RLN: 187-10 = 177
ST: 187-16 = 171
Then divide those point totals by the total tiles available to that player:
177 / 90 = 1.96667
171 / 84 = 2.03571
Look at them decimal points! It's not a huge difference, but it is a difference. I don't know if it's a statistically significant difference, and so would have to either ask someone who's better at math to work on it, or simply play a few hundred games and see if there's a noticeable difference in point totals for ST players vs. RLN players.
And then we can even toss in another variable: the ST player will be able to add an S to the end of some of his higher-scoring words, allowing him to pick up some essentially freebie points. It's much less likely that the RLN player will be able to do that. Again, it may not break the game by itself--but combined with the 0.06904 difference in tile value, we could see the advantage sway to the ST player.
That, frankly, is the tip of the asymmetrical iceberg--there are many more variables to consider, and the game could use some added flair (one player gets double and triple letter spaces, but the other gets double and triple word spaces, for example)--and then we have to tweak all of those values and likelihoods. It's daunting, and you're probably still amazed by how those decimals could matter at all, so why bother?
I mean, really. Why? Most of the games we've played for generations aren't asymmetrical--chess, checkers, Monopoly--and sports are never asymmetrical. (I'm fat, so I may be unaware of a weird one somewhere--but even when the offense and defense have different numbers of players on the field, the entire game becomes symmetrical when they flip roles.) Not only do you have to create more content for your game, and you need to make sure that each starting option works and is fun to play, but you need to make sure it's balanced against the other three, five, nine options your opponent might pick.
Which, in fact, is your first reason for making your game asymmetrical: players have more choice when playing your game. Team Fortress 2 features 9 unique classes, ranging from a guy who only shoots explosives to a medic to a teleporter-building engineer to a sniper. And a pyromaniac. Plus a scout who runs wicked fast and can double-jump. There's something for everyone, which gives your game a wider appeal than a game where everyone's an explosive-throwing medic. Hey...
No. That diversity is still more interesting.
And of course, in the same way that it's much more difficult for the game designers to balance an asymmetrical game, it's also more difficult for players to break an asymmetric game. Players will always look for a dominant strategy, but with the variety of options available, it becomes less clear which option or strategy is the absolute best. Brand-new concept time: the metagame is all of the out-of-game information players consider while playing the game. Sounds weird, right--what out-of-game information could be important in the game? In Starcraft, the Zerg race has a lot of cheap, but fast and expendable troops. This means they can make a small but deadly force very early in the game, which they can use to attack opponents before they have defenses up. This strategy is commonly known as rushing, but there's nothing in the rules of Starcraft that say the Zerg have to rush or the other races can't rush--players have determined that rushing as Zerg is effective, so it's common knowledge that you should be on the lookout for a rush if you're playing against a Zerg opponent. That's the metagame--you understand how other players play the game, which affects how you decide to play the game.
The metagame for asymmetric games, in general, is much deeper than symmetrical games. Because players have so many options, and players all value these options differently (and can use them to different extents because of their particular skills), there are more choices to be made in a particular game--not only are you making choices, but your opponent is making choices, which you must respond to (by making choices), and then they make choices...
You get it. It all goes back to meaningful choices--pack more of those into a game, and it will remain more interesting (also known as "fun") for players for much longer. Asymmetric games, for all of the challenge they present to designers, are great at providing players with meaningful choices.
Now, as to whether or not we should make an asymmetric game... well, I'll remain mum on that.
Reader Comments