A conversation about Maxis’s upcoming Spore yesterday made me think harder about a specific design feature it will employ: Dynamic Difficulty Adjustment. DDA is an appealing idea at first blush, but the more I think about it, the less appealing I find it.
The idea is to adjust the difficulty of the game to the player’s skill. A game which is too easy is dull; a game which is too hard is frustrating. Fun lies in a sweet spot between the two. Unfortunately for game designers, that sweet spot is different for different players, with varying degrees of talent and skill.
Enter the concept of dynamic difficulty. Instead of taking a guess at what level of challenge most people will enjoy, possibly getting it wrong, and certainly alienating consumers too far towards either end of the spectrum, designers hope with DDA to adjust the play experience to each individual. As you play, the computer or console tracks your progress. If you take too long on certain levels, miss too many scoring opportunities, or die too often, the next level will be made easier—fewer enemies, perhaps, or stupider ones, or power-ups hidden in more obvious locations, or faster resource production. Conversely, if you’re breezing along, the game raises the difficulty until you have to start pushing yourself a bit. Done right, everyone from 4-year-old with no motor control to pinball wizard gets the thrill of victory against what feels like stiff opposition.
This technique has been around a while, actually. I first noticed it myself in Area 51: after many plays, the first couple minutes became routine, until I suddenly began to face aliens that had never appeared before; the game was adding enemies to speed my inevitable death. If, on the other hand, I screwed up and died early, the game took it easy on me for a while, and aliens that should have been there shooting at me weren’t. Because I am not an arcade junkie, I suppose other games had already employed DDA, and this was the first I’d noticed it.
There’s a difference between arcade games and computer/console games, however, when it comes to DDA. Arcade games set themselves to a level to suck your quarters at maximum efficiency, and DDA works just fine for this. Games you buy at an outlet seek to employ DDA to make your experience as fun as possible, and maybe it works. But I can see a huge pitfall which may be impossible to avoid.
The ideal level of challenge depends not only on a player’s skill, but upon his tolerance for losing once in a while, and the game has no way to measure this. I like to lose occasionally—occasionally!—just to demonstrate that playing well matters, and to feel like the usual victories took skill. Because this is the majority attitude, DDA will tend to aim for this level of difficulty. But not everyone feels that way. My friend Tim enjoys a hard-fought loss, including full nuclear exchanges, in a Civilization game more than a sweeping victory. Other gamers go out and look up cheat codes before they even start the game. Effective DDA will disenfranchise such players as thoroughly as fixed difficulty levels disenfranchise the very skilled or very unskilled. Under DDA, Tim’s opponents would slowly handicap themselves to make sure he wins the arms race, and the cheat code kids will soon find themselves facing so many enemies that immunity to bullets only makes fights eternal, rather than guaranteeing victory. In this way, DDA fails to serve its purpose of matching everyone’s play style.
Less dramatic than winning and losing, but possibly more important to overall game enjoyment, is the question of the difficulty of moment-to-moment play. Play improves with increased attention and effort, but we are, after all, talking about games. Some players want deep involvement, others just to goof around, and the game has no way to measure this, either. Suppose, for example, that you’re playing a Space Invaders-style shooting game, and doing fairly well. So the game ramps up the difficulty. You can still handle the new difficulty, but you have to pay close attention, maybe pause the game occasionally, and push your thumbs harder than you like. You’re still winning, but the play is no longer so fun. In theory, you could commit suicide to tell the game to ease up, but few players will have the willpower to do that. 4X players experience a similar phenomenon in automated functions. 4X games often offer to automate certain tedious tasks, like laying roads or fighting tactical battles, but because the automated functions are inferior to an intelligent person, the player feels compelled to perform the tedious task himself, in order to preserve a possibly vital advantage. DDA merely shifts the design question from “How hard should this be?” to “How much focus should this require?”
In a similar vein, DDA takes some of the fun out of experimentation. A well-designed strategy game has room for several distinct and viable strategies. Sampling these, and figuring out how to make each work well, and discovering which works best is fun. It is also fun to invent oddball strategies and see whether they have merit or fail spectacularly. But if the game adjusts its own difficulty while the player explores them, he has no way to judge which strategies work because they are sound and which ones work only because of the built-in handicap. Traditionally, a player wanting to fool around with oddball strategies could set the difficulty levels himself at the start of the game, choosing an appropriate handicap; DDA denies this control.
To cite a specific example, and to return to Spore, Maxis favors games like the Sims and SimEarth which are about exploring the game itself, seeing what it can do, and not about victory of one kind or another. They do very well with the segment of gamers who like to explore. Gamers as a whole, however, are a goal-oriented bunch, and often pooh-pooh Maxis games because they are too easy to win. This is unavoidable; if players are to be encouraged to pursue lots of interesting new ideas, they can’t be made to lose every time they do. The core of Maxis fans will want to see what happens if their creatures grow too many legs, or if they deliberately hunt dangerous beasts, or if they pick a fight with a technologically advanced species, and must be allowed to pick up the pieces and continue when these actions don’t work so well. Players who stick to optimal strategies, obviously, perform much better against some scale of progress.
I believe Spore embraces DDA in an attempt to reach both audiences, but expect it will fail. DDA aims largely to ramp up environmental resistance to players seeking to reach interstellar empires as efficiently as possible, so that such players can feel they’re beating meaningful opponents, or overcoming significant environmental obstacles. Done well, this can work, but only at the expense of making every play of the game equally effective. Players who like to win will rapidly discover they reach the same endpoint in the same amount of time no matter what they do, and will grow bored almost as soon as they would with a fixed difficulty set too low. Done poorly, Maxis will turn out another SimEarth or SimAnt, both of which could be won without a single mouse click. These games played themselves, and, while the player was welcome to participate, he wasn’t necessary, or even particularly helpful.
Almost all of us like to win, almost all the time. We like the illusion of a challenge, but not, generally, a true challenge. Although hard-core players insist on the importance of an objective challenge, and sneer at sissies who play at “medium” difficulty or dislike PvP, this is because the hardcore players can easily master harder tasks. Typically, those same players quietly alter the difficulty of games they aren’t so good at, or, more often, won’t play those games at all. DDA cunningly aims to satisfy what we really want, and not what we pretend we want. Skillfully aimed, DDA can give just the right amount of resistance to player victory to give most of us a fun time. But I don’t think it can give any more of us any more fun than a skillfully aimed fixed difficulty level.