On Wed, Jun 2, 2010 at 5:25 PM, Jeff Nowakowski <[email protected]> wrote: > On 06/02/2010 01:14 PM, Don Dailey wrote: >> >> Why are you comparing humans to computers? It's ridiculous to measure >> progress by comparing to the top human players. What we care about is >> how >> much progress we can make from year to year. > > Come on Don, you know that the top players are the gold standard that the > programs are trying to beat. That's why lots of programmers have been moving > to Go from chess. > > The original question was why is Go harder than chess for computers, as it > clearly is (are you disputing this?). I think Steve answered the question > very well in his original reply. > > I understand your dispute with Dave's simplification, but I think you can > agree that alpha-beta for chess was a strong foundation to build on, that it > didn't work for Go, and that we now have monte-carlo tree search as the > foundation. Nobody disputes Go has made progress and will continue to do so.
I am with Don on this one. Computer chess has been a hugely successful endeavor through the efforts of lots of very smart people over many years. To deny credit to the people involved by attributing that success to hardware speedups alone is offensive and is a manipulation of reality. As someone who spent his weekends for several years writing a chess program, I can tell you that there isn't anything easy about it. You "count the wood", you "give a little consideration to position", you use alpha-beta and you get a program that plays like crap. If you add quiescence search, you'll do better, but the tactics will be weak and it will systematically blunder any position that involves a king-side attack or a passed pawn. You can get your program to 2000 ELO in a couple of months of refinements (hash tables, check externsions, reasonable move ordering, iterative deepening, king safety, passed-pawn evaluation...), and then you have another 1200 ELO to go before you can compete with the big boys. _______________________________________________ Computer-go mailing list [email protected] http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
