What's The Point of Positive Feedback In A Game?

So, assuming you know all about positive and negative feedback in games. (If you don't, this is a nice visual.)

Mark Nau (@marknau), Tomo Moriwaki (@tomomoriwaki), and I were discussing sixty second shooter's game mechanics last GDC, and Mark said something that gave me (and I think Tomo) pause - "Why have the chain multipliers? It just increases the separation between the good players and the bad players." (Have you noticed so many of my game design posts begin with 'Mark Nau says'?)

I think I opened my mouth and nothing came out. Tomo said something along the lines of because then you're getting crazy amounts of points and it's awesome but didn't really put his finger on why it was awesome.

My instinct is that it's important, but maybe that's just me-too-ism. Games are rife with positive feedback systems.

• In Rockband, if you get a lot of notes right in succession, you get a bonus multiplier; if you get notes right in the right places, you get powerups. Thus a good player scores remarkably higher than a bad player. But why does that matter? Why not make it so you get one point per note, done? The good players will still beat the bad players.
• Similarly, chain multipliers in any fighters and shooters.
• Winner's Outs in basketball - you score a basket, you get to try again.
• Wargames often have you conquer territory, which gets you resources, which lets you build bigger armies, which lets you conquer more territory. At a certain point your victory becomes a foregone conclusion, unless there are also negative feedback mechanisms in place. If we simplified the wargames so it was just about war, we wouldn't need to balance with the negative feedback mechanisms. I don't see how you could remove positive feedback from wargames completely, though - look at chess.

So, I don't know. Maybe it's as simple as 'it makes the player feel like a hero' - and it seems less cheap than simply tacking a few more zeroes onto the end of the score. But I don't know. What do you think?

(I will say that the chain multiplier in sixty second shooter does create an interesting decision - it encourages you to save your targets for when you can take them out in rapid succession, and because it goes into Chain Reaction mode when you hit 50, there's something to be said for timing your shots to keep yourself in that mode - you can even use the spiral enemies as batteries to keep yourself in chain reaction mode rather than nuking them outright. I don't even know if that's the optimal strategy or not - it doesn't get much more interesting than that, when the guy who made the game isn't sure of the right answer.)

Comments

[Cy]

It seems like the best way to match the scoring to the actual learning curve. If people get to 90% accuracy pretty quick, you want a scoring system where the spread between 89% accuracy and 91% accuracy is larger and more consistent than 2%. If it's linear, your skill level gets lost in the noise and it's harder to get a sense of your progress.

[Greg]

Chain multipliers are pure Skinner-box reinforcement; it's exciting to rack up the score. Tomo is right. And who cares that it "increases the division between good and bad players"? How many people really care?

Your point about wargames is more salient. Positive-reinforcement games, where production resources grow with conquest, are a problem, because the end-game becomes boring; it's clear who's going to win, the rest is mopping up. Note, however, that most classic wargames are NOT this way; they're negative sum games, in which my victory over you also causes me some casualties. Consider the Eastern Front; the Germans romp at first, but get worn down, and eventually get trounced by the Russians, who have more manpower resources, and eventually learn the ropes of armored warfare and start out-producing the Germans.

Positive reinforcement in scoring can increase player excitement; positive reinforcement in capability almost needs to be counterbalanced by some negative reinforcement loop, or you wind up with a tedious end-game.

[Grilljanne]

I think that both have their place. Positive feedback loops in a competitive game amplifies differences in skill. In a game with only PFLs, the most skillful (or lucky) player will usually win, and do it quickly. In a game with only NFL, the risk of a stalemate is very high.

In addition, I think that games with lots of NFL have a problem with that actions in the early game are meaningless. In Mario Kart, only the last half of the final lap really matters. While games with PFL have the opposite problem (late-game is already decided), players can surrender. No one plays a Starcraft game to its actual end.

Basically, both positive and negative feedback loops have the same problems, only opposite. Both are present in most games, so I guess the challenge is balancing them and applying them in the correct situation.