Forrest Curo wrote: >> I'd like to know how well MoGo would have played if you let it think >> for a week for every move. > > Probably diminishing returns. Once a series of random playouts has > given it a selection of the more significant points to consider, I'd > expect move-order, forcing moves, the need to follow a sequence to a > stable conclusion to become so critical that any number of pure-random > playouts would fall short of giving a proper evaluation of a position. Diminishing returns is a certainty, the only question is how severe. Many other games have proved that you are not likely to see much diminishing returns if the game is complex. You may be able to observe it in simpler games. Certainly, you can see it in games which have now been solved - you have the ultimate diminshing returns because more effort doesn't improve on perfection.
You can expect roughly linear ELO improvement per doubling with a well written MC program that builds a tree and doesn't use non-scalable techniques. You are not likely to see this diminish severely in the next 50 years. The current top programs may use techniques that affect their scalability, but empirically are a good choice for modern day hardware. Lazarus has this for instance. I prune some moves on a permanent basis that are not admissible (there is a chance that I am pruning the best moves.) So a version of Lazarus running on a much faster computer may be weakened by this. (Not just weakened, but permanently crippled. Since I have artificially limited the strength you might actually be able to observe a point of diminishing returns in the not too distant future with this version of Lazarus. I'm pretty sure most people have this backwards. You will see the largest diminishing returns at very high strength levels, not at the low end where we are now. There is no evidence whatsoever that severe diminishing returns is "just around the corner." It's ok to believe it is and try to explain why you think this is the case, but the idea of sudden diminishing returns has been debunked by all kinds of experiments with other games. At least I would like some kind of evidence presented before I reconsider whether there is something truly odd about GO that makes it difference. > -------- > > Another matter: > > I'm very fond of C, because my love-hate relation with computers goes > back to the days when it was essential to know where your program was > and what it was up to; I'm still happier with systems where I know > they're doing pretty much what I asked for. I think I have the same bias. I don't really like C but I certainly enjoy the power and control. I think that is also why I migrated away from DOS and Windows. > > But for go I think I'll need a complex design with multiply-linked > lists, which I can do in C, but not without my mind turning to > sphagetti-knots for the duration. > > So Scheme is one of the languages I've been considering, and in the > process I stumbled upon a list of programs it was used to write. One > of them: GIMP (Graphic Images Manipulation Program). > Relevance?--Graphic images of any detail are enormous chunks of data; > doing even a simple computation on one of these files has got to > require a lot of bit-crunching, which used to be pretty time-consuming > even when I was processing low-resolution grayscale photos for a > monthly tabloid. I haven't run a direct comparison with GIMP's > commercial rivals, but it's impressively fast by my standards... > I intend to look at scheme in a serious way too. But I don't see myself writing a chess program in scheme. > Forrest Curo > > > > ---------------------------------------------------------------- > This message was sent using IMP, the Internet Messaging Program. > > _______________________________________________ > computer-go mailing list > computer-go@computer-go.org > http://www.computer-go.org/mailman/listinfo/computer-go/ > _______________________________________________ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
