Three Best Chess Short Stories
To round out discussing Best Chess, here are a couple of other stray thoughts I’ve thought while making it and playing it over the last week or so.
Does it work? An importantly weird element of writing a chess-solving AI like this is that you can’t actually test it. Given that it’s objective is to solve chess, and given that’s impossible (time-wise), you’ll never really know whether it works or not. That was a constant strange feeling that led me to repeatedly ‘prove’ to myself that the approach I was taking (to implement the depth-first search) was legitimate and would solve chess, given enough time. (Since release various people keep querying whether it works, whether it would run out of space, etc., constantly shaking me to my core.)
I solved chess! One embarrassing story from development was when I managed to think (for about 2 or 3 minutes) that the game might solve chess randomly after running for only a short time because it would just stumble into the solution. Rilla and I then postulated an issue of Science with me on the cover and the headline ‘Idiot Solves Chess’ (me being the idiot). But then of course I recalled that this algorithm (or any) can’t actually solve chess during a partial run like that – the whole point is you have to examined all the options, which takes almost forever. (The follow-up Science cover was going to be ‘Idiot is Just an Idiot After All’.)
That’s it. I’m moving on to my next thing, called… BREAKSOUT!