Frets On Fire Metrics
Aug 27, 10:35 AM
In addition to our Quake III work, we’ve been doing some metrics integration with Frets on Fire, which is an open source Guitar Hero clone written in Python. The production values are actually really good, I recommend you give it a shot. It can be played with a keyboard, although I’d recommend grabbing a DualShock-to-USB converter and playing with an actual Guitar Hero guitar.
Anyway, we’ve already found some interesting data from our metrics. Here are two pictures of Jeff’s performance on a song. The big green dots are the notes that Jeff was supposed to hit, and the small green dots connected by a line are how far off Jeff was when he hit the note. Jeff actually hit every note in these screenshots, they just show how far on or off target he was.
We can see that the left-hand portion of the song is significantly easier than the right-hand portion: there are fewer notes and they’re spaced further apart. Yet Jeff is actually more accurate when the notes are closer together! It’s like he knows he has to buckle down and concentrate when the notes come fast, but doesn’t really need to pay attention when the notes are sparse.
Maybe this has something to do with Csikszentmihalyi’s theory of flow? As applied to video games, it would mean that Jeff was bored and not “in the zone” when he was being under-challenged, but the fast notes were just about right for him. Here’s a good excerpt from Jenova Chen’s thesis that talks about flow theory in games.
Great post, that last paragraph blew my mind. Csikszentmihalyi’s theory makes a lot of sense here. I guess good adaptability to Flow makes a good game great.
— Ben W · Aug 27, 09:37 PM · #
Thanks for the interesting post and insights to game design.
The flow theory makes sense, but also from my experience I’ve found it harder to stay in rhythm if notes are far apart. It’s not like botch them up every time, but a steady fast rhythm is much easier to keep than slow one.
— Perti Purho · Sep 2, 06:03 AM · #
Obviously, GH and the like provide excellent case studies for Flow-related analysis.
The answer here may be much more obvious, however, which is this: The closer-together notes may simply have followed more easily recognizable patterns than the sparse ones, and therefore he was able to perform pattern-recognition and adapt to them more easily.
I’m pretty certain, at the very least, that this is true for me.
— Darren Torpey · Oct 6, 09:29 PM · #