A psychologist's thoughts on how and why we play games

Saturday, March 10, 2012

Flow Theory case study: Tetris

In the previous article I was wondering about whether losing periodically is necessary for the experience of flow, or if ever losing means immediately losing your flow.  A friend mentioned Tetris to me, which will make for a great case study.

A lot of the classic video games are very good at providing flow because of their linearly increasing difficulty and the inevitability of losing.  Consider Tetris. Tetris has all the basics: good controls, simple and concise rules, and constant feedback in the form of your score.  You can tell when you're doing well, because the lines disappear and your score increases.  You can tell when you're doing very well because the screen also flashes when you line up a Tetris.

 "[The] only video game I can think of that left cultural marks for 
decades & will last 100 [years] is Tetris" - Raph Koster

The key to Tetris (and many classics like it, such as Missile Command and Breakout) is that the difficulty steadily increases over time.  The player may start the difficulty at any level.  Then, after every regular interval (10 lines) the game becomes more difficult as the pieces begin to fall faster.  The difficulty, then, will inevitably reach optimal challenge.  Sooner or later, the player will be optimally challenged, just barely slotting the pieces where they're needed, or making small mistakes and having to be clever and quick to correct them.  After a while longer, the pieces will fall too quickly, the player will become too challenged, and soon after, the game will end.

Tetris is very good at spending the player's time optimally challenged. The player doesn't have to spend any time bored, because he/she can set the starting difficulty at any level.  The particularly brilliant part is that the player spends very little time frustrated, because the game ends shortly after the game becomes too difficult.  When the difficulty has gone too high, the player loses and starts a fresh game, beginning again at a lower difficulty.  The player never has to play through extended periods which are too easy or try to drag him/herself through an exceedingly difficult part of the game.

Graph of difficulty (the speed at which blocks fall) over time within a single game of Tetris.  Sometime between the start and end of the game, Tetris will inevitably reach optimum challenge.  The player can reduce the time spent in "Ho-hum" territory by starting at a higher difficulty.

You can almost think of Tetris as searching for the player's skill level. The player punches in the best guess, since starting at a higher level means earning more points before losing.  Tetris then ratchets up the difficulty until the player can no longer play, assuming that someplace in between it has to run into the perfect challenge.  I have to wonder - if we could stop Tetris from growing more difficult when it hit that perfect balance, would we be happy to continue playing for half an hour at a time?  I suspect that the game around the experience of flow is still important - chasing after that high score means you can also track your skill as a player.

Games which are intended to be beaten or finished don't have the luxury of this particular search strategy.  If a game like Halo or Mass Effect 2 steadily ratcheted up the difficulty until the player died, the player would probably become pretty annoyed at the inevitability of having to die and lose progress before the game made it easier again.  Developers of these games have to try to make the whole game optimally challenging, meeting the player's skill and then growing in difficulty as the player grows in skill.  It's almost an impossible task, which is why probably a better design decision for these games is finely-grained difficulty settings.

 The best difficulty system in a long, long time.

I'm not talking about "easy, normal, hard", where the difference between settings is enemies having +/- 50% health and the player taking +/- 50% damage - that's not nearly granular enough.  Instead, I'm thinking of something like Bastion's "idols" system, where the player can change particular parameters about enemies, making them a little tougher for a little extra reward (ie enemies move faster; enemies recover health over time; enemies drop a grenade when they die; enemies drop fewer power-ups).

Helping your players to optimal challenge is easy when they're destined to lose - just keep cranking it up and they're bound to hit it eventually!  Keeping them there for hours of a game that takes away their progress when they die, though, that's the really tricky part.


  1. In a rather unrelated note, I think that one great way to torture a human being would be to force them to watch Tetris played poorly. It would be something like the awkward moments on The Office or Meet the Parents, but at an extremely visceral level.

    1. http://www.youtube.com/watch?v=Alw5hs0chj0 the first few minutes of this video always make me cringe, ditto the scene in Office Space where Peter is playing Tetris (poorly) in his cubicle.

    2. The first few minutes of that 2 minute video

  2. The graph confuses me. Does the player's skill never increase?

    1. Over several games of Tetris, sure. The graph represents a single game of Tetris, though.

    2. If this is your graph, you may want to consider re-labeling the x-axis as "Duration of Play" or something similar, just to avoid confusions like this.

  3. As for the tricky part, I assume that's where the high score system comes into place?

  4. Id nominate limbo as a good example of a game where failure isnt flow breaking at all its the culmination of the tension and horror of the world expressed in a gorey sequence and the mistakes that lead to it.

    Id see dying in limbo as an integral part of the game experience as give how shocking a lot of the deaths can be and how unceremonious it is forms part of the overall story and thematic elements and ties nicely into things like using other children to activate traps or the bodies of suicidal children as makeshift rafts.