Finite State Jiu-Jitsu

on January 20, 2012

Today completes my first week of BJJ under Rodrigo Cabral. I admit that I have spent more time on YouTube BJJ than real life BJJ, but being the huge nerd that I am, I have already started to intellectualize what jiu-jitsu is.

Now, I want to make it clear that I am a total n00b, so I probably have no clue what I’m talking about. Take what I’m saying here and then armbar it to make it tap out. I want to record this thought so that I can check back in in the distant future to see if I got it right. This is a distant future in which I am the most lethal nerd. More on that in a later post.

So what is “Finite State Jiu-Jitsu”? It is inspired by my CS background. I understand the depth of one’s jiu-jitsu as a quantifiable property of a finite-state machine. The nodes are basically positions like guard, half-guard, mount, etc and the techniques are essentially transitions (think guard passes etc). A black-belt like my instructor probably has hundreds and hundreds of nodes and thousands of transitions melded into his central nervous system, just a few electric pulses away from deadly submission. He has nodes and transitions for things that I do not know exist.

A total n00b like me, on the other hand, has a very finite-state machine, consisting of a few guards, mount, combat position, and uh… yeah that might be it for now. I have literally one transition between all of those, except maybe from Mount -> Submission, between which I have two arrows: the kimura and the armbar.

In this model, I view each class as adding a node or a transition, growing my repertoire. I see drilling and practice as making the lookup for those nodes and transitions much faster. During my roll today, I literally felt some O(n) lookup time as I had to sit under the mount, idly trying to control my opponent’s arms, and conjure up the individual steps I needed to sweep. By the time I figured it out, she guillotined me. Nice.