On Mon, 2008-11-17 at 16:04 -0200, Mark Boon wrote: > On another note, as an experiment I have a bot running on CGOS that > is the ref-bot but instead of using a fixed number of simulations I > use a fixed amount of time that slowly diminishes towards the end of > the game. The result is it does about 200K simulations per move for > most of the game on a single processor. Its rating is currently > stuck > at 1367 or so. With 2K simulations the rating tends to be 1280 > without using the weighted formula. This one uses 100 times as many > simulations and the weighted formula, so I had expected it to rate > higher than that. Is this normal? Does the MC-AMAF combination just > not scale at all? Or could it be because there currently doesn't > seem > to be a very large population of bots playing?
After 2k playouts you are already facing serious diminishing returns - but there is still a bit more to be had. But after 5-10K playouts you get almost nothing. I'm focusing all my experiments on 1k simulations. You can test it really fast and it's a great return for the investment. If you want to go for the strongest optimized version I would suggest not going beyond 5k playouts. Up-front time loading is very powerful and you will get a good return on that too. So you might try to start out at perhaps 10K playouts and try to set it up to taper down to 100 near the end of the game. If I view strength as a "time per game" equation, I can probably get 5k results with a tapered bot that averages 2k effort or so. Another way to slice this is to have a different algorithms for playing the first 5 or 10 moves - then you can squeeze a lot more from it - assuming the new algorithm plays very fast and strong. That algorithm could be an opening book, but that won't get you into the game very far unless it's massive. - Don >
signature.asc
Description: This is a digitally signed message part
_______________________________________________ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/