Saturday, February 18, 2006

Play Loops

This post crosses from computer games to tabletops and back again.

Let's talk about how a game latches on to the mind of the player.

Last post, I explained how I think the thing which sets games apart is a weighted reward/penalty feedback loop. Example: in an FPS, the feedback loop would be the running around shooting people. The weighted reward/penalty would be health damage - both yours and theirs.

To make it a little shorter to type, I'll call these "play loops".

Here's the thing: in order for a play loop to catch a player and be fascinating, the player has to appreciate the rewards and penalties it gives him. I don't much like racing games, largely because I don't appreciate their rewards - simply being a little faster or a little slower doesn't mean much to me.

There's nothing wrong with racing games. It's simply that the play loop doesn't catch me.

There's two basic ways of getting a player to appreciate a play loop, and games commonly use both. The first is "immersion", the second is "support loops".

Using graphics, story, conversation, music, and so forth, a game designer can make the player fascinated by the world the play loop exists in. The linear progression of nearly all stories in nearly all games is a method of drawing the player into the world of the game.

What this does isn't the same as it does in the movies. Rather, it is the same, but it is applied differently: once the player is immersed in the game world, the rewards and penalties relative to that world have a much deeper hook into his mind. To give a tiny example from one corner of this broad idea, if you want to know what's going to happen next, you want to win. Suddenly, the feedback matters.

Now, it's actually a bit... off. In certain quantities, the wish to advance the game wholly replaces the wish to play the game. This happens more frequently than most developers are comfortable with. Players say, "Damn it, can't I just skip this level?"

Yow! That's bad! They wish you were a movie!

Occasionally the reason for this is having a pretty poor feedback loop - the game is too simple or buggy. However, most of the time, it's for a more subtle reason:

The play loop's rewards (health, items, levels) have been replaced in your mind with the story's rewards. The story has only one type of response: advance or not advance. This is not strong enough to support a play loop - it turns the game from a weighted reward/penalty feedback loop into a simple feedback loop. Unacceptable.

By understanding that what drives a game is a play loop instead of a simple feedback loop, you can make your immersions come in distinct reward magnitudes, just like a full play loop's.

One example of this is TimeSplitters 2: each time you beat a challenge, you get a few rewards. But the rewards you get are based on how well you beat the challenge. It's not simply "kill these zombies successfully and get a new hat", it's "kill 30 zombies, get a new hat. Kill 40 zombies, get a new hat and a new level. Kill 50 zombies, get a hat, level, and new character!" This is on top of the subtle but important feedback of bronze, silver, gold and platinum trophies - clearly displayed as a sign of the magnitude of your victory.

Actually, I'm a little irritated at the system, because I really suck at some types of challenges. Like picking up smegging bananas. I hate racing games.

It might have been a good idea to offer alternate challenges or something. It might have also been a good idea to display what you could win at each level of victory, but that's besides the point. The real point is clear: they offer rewards in which the magnitude of your victory matters. This means that the play loop remains a play loop. The reward has shifted somewhat, but still exists as a complex reward instead of just win/lose.

That is a support loop. It's another feedback loop, often a very simple one, which interacts with the main loop to increase the complexity and longevity of the main loop. This is very common. Nearly all non-casual games contain at least two support loops, if you can even distinguish which ones are support and which one is main.

I could analyze other games - or more deeply analyze TimeSplitters 2. Like all complex systems, it deserves more than I give it here. But that would be even more boring than this essay already is, so let's move quickly on:

When you create a game - no matter what type, it's important to offer this complex reward. The game cannot simply progress, because that means they'll either lose interest in the play loop or lose interest in the progression. Either way, it's totally inefficient.

So, for example, a tabletop RPG. The "main loop" is the incomprehensible morass of rules used for fighting, supported by numerous support loops for traps, treasure, religion, leveling, and everything else under the sun. (Most tabletop RPGs have at least 8 support loops.)

The basic idea of most tabletops is that they include all that crap because it means that a clueless GM can simply plug it in and let it chug. The players will theoretically latch on to these support loops and drive themselves to accomplish whatever goals are made likely.

The thing is, this is unreliable and inefficient.

Most GMs have a problem "controlling" their players. I hesitate to call it "control": a better word would be "guiding". Their players don't bother to follow the GM's nicely laid-out plot.

I don't have that problem. The reason is simple: I don't use all the default support loops. I create my own loops which follow the rules I laid out above.

Default support loops offer brutally generic feedback. Many players will have them memorized. Their rewards are too powerful while simultaneously being too simple. How many times have you heard a player say something like, "I want to get to level six!"

They've replaced the main play loop's rewards and penalties - consisting mostly of HP finagling and some resource management - with a simple "treadmill through to level six". That's bad.

Your support loops should be the guiding light of the game. To do that, you need to make other support loops subordinate to it. In my opinion, this means ditching all the crap that comes in the giant rules tome. You don't need their leveling, their item lists. You certainly don't need their whole world! Even if you're not able to think up a world on your own, cut away all (and I mean all) the pieces of the world you aren't using. Just don't mention them.

Then you bring in your own support loops. These loops react to how well or how poorly your players do. Don't hook your support loop onto the back of some other support loop, either: make it the dominant loop. It touches and reshapes all the other support loops. Rewards and penalties should hit every other support loop in addition to itself: weapons, experience, information, allies, or anything else you can think of.

If the players seem to be thinking too much about one of the less important play loops, such as levels, spin the other play loops. Give them jack shit for XP: instead reward them with information. Or allies. Or items. Or glory. It shouldn't take much before the player gets interested in one of the more active loops - that's how a player's mind works.

Honestly, I can't say this will work for you - or anyone else. But it seems to work for me, and I believe it is the fundamental "gameness" of a game.


Patrick Dugan said...

I think what you're getting at with the table-top deal, whether you know it or not, is that the best support loops have to do with story dynamics. Allying with the character representing ideal blue versus ideal red, or garnering information that put the conflcit and its relevant agencies in a radical new light; these are what makes the game interesting in an almost literary way. Planescape: Torment, the Tim Schafer games, uh... I'm groping here because there really hasn't been much that fleshes this out as fully as this can be fleshed out. Thats basically what interactive storytelling is about, solving that problem of play loop distracton by making the social configuration of the character's relationships the main play loops, and the relevant forms of info and whatever major goals that characters could undertake the support loops.

Its sort of an inversion, in loop terms, you make alliances and other relationships the main feature, and things like, say questing in an Arthurian storyworld, the support. Traditional games, as you know, are all about grails and rarely about the immediacy of human relationships.

Mory said...

I don't understand what you're saying when you talk about reward/penalty loops. How can you say that this is what defines games, when the same psychological conditions occur in just about every activity in life? At a job, you are in a feedback loop of doing work. Assuming your boss is good, you will be rewarded for doing that job well, and penalized for doing it badly. So every job on Earth is actually a game, is that it?

It goes beyond that. Think of any type of performance- no, you know what, I'm just going to go ahead and take the example of playing piano, because I know it well. When I'm playing for someone else, in the loop of trying to hit the right notes and put as much nuance in as possible, I can either be rewarded (if I play well) by seeing my listener enjoy himself, or penalized by getting him bored. In fact, this type of feedback loop is there even if no one's there to listen, because I'm rewarded by hearing my own music and enjoying it, and penalized by the frustration of hearing a mistake. So all performances in the world are games, is that it?

In writing this comment, I am going through a similar reward/penalty feedback loop. The loop itself is in the writing, and there is an immediate reward/penalty from myself in thinking about what I'm writing and evaluating it. There will also be a longer-term reward/penalty when I see your response. So is all creative work a game?

When you go shopping, there are two delayed reward/penalty responses to the shopping loop. If you found a good product, then you'll be rewarded by being able to use it, and if you didn't, then you'll be penalized by not being able to use it well. If you found a good deal, you'll pat yourself on the back later for spending money well, and if you were ripped off you'll be angry at yourself later. Is shopping a game?

You know what, just to drive the point home, I'm going to go as extreme as I possibly can. Being brutally interrogated has a more clear reward/penalty feedback loop than anything I've ever heard of. If you say the right thing, you'll be rewarded by being let go, and if you say the wrong thing, you'll be penalized. It doesn't get any more clear-cut than that! Wow, I think it's a game, don't you?

Okay, that was harsh. But I think you're looking so hard for a common thread through all of games that you're finding something which isn't there. Not all games penalize the player for playing badly. Animal Crossing rewards you for doing someone's errand, but there's no accounting for, say, how quickly you do it. So there's reward, but no penalty. How is this different from any rewarding activity, such as reading a book?

Or take puzzle games. The loop is trying to think of a solution, which when you think about it isn't much of a loop at all or else it's not very effective. There is only one reward, which you get when you solve the puzzle. If you don't solve the puzzle, you think about it some more until you do. Where's the penalty?

You see my point, I think, so I'll stop here.

Craig Perko said...

Mory: Excellent!

By my definition, those things are games if and only if two things happen:

A) they have a highly varied reward/penalty system and

B) the player values that system.

So working at a job is only a game when you're trying to succeed. Too many jobs are, well, work. The amount of success hardly matters. It just turns pass/fail.

In quite the opposite way, Animal Crossing is a game. I have a friend who accidentally moved a stone and put down an arrow on the ground. He hated it! It was an aesthetic punishment, and he learned a lot about the game.

Aesthetic rewards are just as important as story or stat rewards!

There are two kinds of puzzle games. One is like Lolo, where there's only one end state: progression. The other is like Tetris. In the latter case, it's obviously a varied reward/penalty loop: how much a mismove brings you to the top of the screen and makes it more difficult to clear the lines, or how well you're clearing the lines.

In the case of things like Lolo or early King's Quests, the play loop is in figuring out the state of the universe enough to advance. For example, "ah-ha! I can make the yak block the medusa's gaze if I move him just right!"

These puzzle games contain numerous micro-puzzles which can usually be discovered in a wide variety of ways. While the play loop has a bit of a lopsided limp, it's still definitely a play loop.

As for your comment on being held prisoner: it would certainly be a game for you, and it could become one for me if your rewards and penalties varied according to some kind of rule set - some kind of feedback loop.

It's not a game I would want to play, though. You could even say it isn't a "game", but the only reason you'd say that is because these days, "game" has become bound to the word "fun".

Thanks for the comment! I love these kinds of comments, and I hope I defended myself clearly.

Patrick: No! Story dynamics are simply one way to do it. You can also do it with rule sets, or with new pictures, or whatever else.

The key is simply "dynamics", not "story".

Still, story dynamics would fit the role wonderfully, and I rather wish they were more common.

Patrick Dugan said...

My philosophy sees life as both an open game and a story told interactively by all its participants, which is every sentient mind I guess. So Mory is right in a sense, and the value of games as a medium is being able to inscribe the process that implies a certain type of experience to people, so you could make a game about going to work if you wanted, in fact a recently released anti-advergame does a nice job of making the job of McDonald's exec a game, to cutting satirical effect.

I've been over how human congnition operates with some ratio of pattern recognition and sequential logic. Using purely sequential processing, things fall into narrative formations, using purely pattern recognition things look like a graded distribution ala a purely laid out weighted feedback loop. I think since people's memory and the collective memory of the culture tends to boil down the complexity of things into retrospective stories, it makes sense for games to bring things full circle and make weighted feedback loops which implicitedly tell a story. In fact, games already have done this, its just that the "stories" produced haven't been as interestinga as the loops that produced them, for instance there isn't much dramatic naunce in mario grabbing the invinciblity star and barreling through vs. mario taking pot shots with the fire flower and carefully gaining ground.

So with interactive storytelling and social challenge, the idea is to balance that ratio, and perhaps the ratio between logic and pattern recognition as well, which would yeild more lateral/subjunctive thinking in the users. I think in that meta-genre the focus of the challenge becomes the encouragement of lateral/subjunctive though with the means of that very sort of thinking, you bait them with some information to get them thinking in one of several directions, leading to more imaginative thought and action. Whereas before discrete punishment was required, making the game follow a sort of Soviet Russia model of achievement, now we can punish people simply with a lack of reward, giving interactive storytelling more capitalistic gameplay, for the lack of a better analogy since I'm not a huge fan of capitalism in general.

Animal Crossing is a good example of this, so I guess this comment is getting at aesthetic rewards in detail.

Craig Perko said...

Hmmm. I disagree with almost everything you said after the first paragraph.

"Game" stories - such as grabbing a star vs taking potshots in Mario - are not good stories. But they are good games.

Perhaps you could make a game which told good stories, but that's incidental. A game can be an awesome game and have a really terrible story - Tetris, as an obvious example.

A game can have a great story but be a terribly shitty game. Look at about half the RPG mods out there.

A story is simply one of about ten million ways to get the player interested in the game's environment. Other, more efficient elements include characters, scenery, music, mood, customizeable outfits... and these can be had without any meaningful "story".

Even past all this, using these kind of aesthetic elements is simply one way to get players interested in the game's play loop. You can use numerous methods: complex rule loops, wagering, putting something real on the line, letting the player create the game...

But all of that is secondary. The one and only thing which makes a game a game is are weighted rewards that come out of a feedback loop.

Sure, you can distill a story out of a given play-through of a play loop. "I grabbed the star and ran for it!"

But the only reason those hold any interest is because you know what the other opportunities were. You grabbed the best reward! It has nothing to do with how good a story it is, and everything to do with clever manipulation of the play loop.

You could argue this is what all stories are about. But the idea of "story" is fundamentally different from the idea of "iterative simulation", not least because stories don't accurately represent progress through a play loop. Or any kind of feedback loop. They represent a tilted ideal of such progress.

That's a really interesting topic, and I'll write on it later.

A game without punishment still has punishment: "neutral" always moves to meet the average. This doesn't mean anything for Animal Crossing, since the day-to-day rewards and penalties have very little to do with the play loop. But it means a lot for, say, a chess player. A master and a student play at the same level one day. The student says, "wahoo! I did really well!" and the master says, "what a terrible game I just played."

Lastly, government controlled economies always fail, and they always drag down the countries they are instituted in. The more rigidly controlled an economy is, the more quickly it fails.

That's simply a basic fact. France is having a hard time - and getting harder - because it has a tightly government-controlled economy. The US economy has been getting steadily worse as the government has dipped its fingers in. Russia's economy has been revitalized by loosening government controls.

I have very strong opinions on that matter. I would be a democrat if they weren't determined to ram our economy into the dirt. Unfortunately, today's republicans are equally suicidal.