Author: Komputer Korner
Date: 21:51:13 06/02/98
Go up one level in this thread
On June 02, 1998 at 21:42:31, Robert Hyatt wrote: >On June 02, 1998 at 21:40:42, Robert Hyatt wrote: > >>On June 02, 1998 at 17:15:04, Komputer Korner wrote: >> >>>On June 02, 1998 at 12:58:48, Robert Hyatt wrote: >>> >>>>On June 02, 1998 at 10:47:02, Ulrich Tuerke wrote: >>>> >>>>>On June 02, 1998 at 10:12:30, Robert Hyatt wrote: >>>>> >>>>>>On June 02, 1998 at 02:01:43, Komputer Korner wrote: >>>>>> >>>>>>>On June 02, 1998 at 00:51:08, Havergal Brian wrote: >>>>>>> >>>>>>>>This is part of my Winboard.ini file. So far I've had no >>>>>>>>crashes/problems. Just put this in the appropriate spot with >>>>>>>>the changes you want. >>>>>>>> >>>>>>>>/cp >>>>>>>>/mg 50 >>>>>>>>/sgf wcraftyvscomet >>>>>>>>/fcp="comet-wb.exe" >>>>>>>>/scp="wcrafty-12_6 xboard" >>>>>>> >>>>>>>Since I turned pondering off, no more crashes but I still have the >>>>>>>problem of Crafty corrupting it's book after a long natch of 40 games. I >>>>>>>have to rebuild it's book each time. There is definitely something >>>>>>>broken with Crafty's book code because in my latest match Crafty was out >>>>>>>of book after 1.e4e5 2.Nf3 !!!!!! >>>>>>>-- >>>>>>>Kpmputer Korner >>>>>> >>>>>> >>>>>>lets back this trolley up a bit. So far as I can tell, there is >>>>>>*nothing* >>>>>>wrong with crafty's book code. It is *impossible* for it to corrupt the >>>>>>book, because it never adds anything to the book at all. If it were >>>>>>doing >>>>>>this, we'd be seeing *lots* of problems on the servers where we play 500 >>>>>>games >>>>>>a day on each program... The problem most likely lies somewhere in your >>>>>>setup, whether you have flakey memory, or something else, is unknown. >>>>>>But >>>>>>I played one 330 game match yesterday crafty vs crafty with *zero* >>>>>>problems >>>>>>of any kind... >>>>>> >>>>>>Be careful of "there is definitely something broken with crafty's book >>>>>>code." as the evidence doesn't support that when everyone else seems to >>>>>>be able to play crafty vs xyz with no problems... >>>>> >>>>>I have had a similar (the same ?) effect when playing Crafty with >>>>>winboard. I think that there is nothing wrong with its book and book >>>>>code. May be, KK has installed the small Crafty book as I did. With the >>>>>default choice for "book random", Crafty only plays lines the frequency >>>>>of which are higher than some threshold value (Bob knows better). >>>>>Apparently, the desired frequencies can hardly be reached in the small >>>>>book. However if you set "book random 5" instead of the default >>>>>randomness for instance, then the threshold is chosen lower and >>>>>variations will be found. >>>>>May be KK has the same effect. >>>> >>>> >>>>I had explained this danger to KK... "small book" means no useful >>>>frequency >>>>data to choose moves. It is more than possible that the book is so >>>>narrow >>>>that after a few games it has learned that moves are bad. >>>> >>>>This can be verified by adding "show book" to the crafty.rc file, and >>>>then >>>>starting crafty in a dos window (in the right directory) and then trying >>>>the opening moves in question. It will display *every* book move it >>>>finds, >>>>and will show which are flagged as unplayable "?" due to bad learned >>>>results, no wins, or not played enough to "trust"... >>> >>>What this means then that no matter how large the book is, eventually ( >>>it may take a very long time for this to happen) Crafty will not trust >>>any book moves and will think on it's own from the beginning. I don't >>>like this way of learning. Much better would be learning by adding to >>>its book the way that M-chess and 99% of humans do it. >> >> >>not at all, it learns "good" and "bad" stuff... the problem is >>different. >>If you use a small book, it has no "frequency" data on which to base >>it's >>move selection. It will *never* play a book move that has no wins, or >>which >>was not played more than once. So you could have 100,000 positions in >>the >>file, with only a few hundred that were playable... that's why I never >>recommend the "small" book. use the 10mb KKUP book.bin and you won't >>have >>any odd book problems... > > >one thing I forgot. If you want to play with book learning, please feel >free to do so. But if you are going to play comet vs crafty and post >the >results, please use a decent book. Otherwise the match results can be >skewed badly, which sends me off looking for bugs on a wild-goose chase. > >use a decent book for public matches, use oddball books for private >matches >and this confusion doesn't happen. Okay I will use the KK Kup book from now on, but when I used the small book that was created from a 400 game Morphy pgn file from M-Chess 7.1, I used the n parameter as well. BTW there is a typo in your documentation on this. The n parameter specifies that unless the move is played in at least n number of games that move won't make it into the book. So I created the book this way so that there were more than 1 result for each move. It pruned the book down to a small amount but there should definitely have been wins in the book after 1.e4e5 2.Nf3!!!!!! I saw Crafty out of book after this sequence. later on in the match it was't out of book until the sequence 1.e4e5 2.Nf3Nc6 3.Bb5Bc5 4.c3 So the book wasn't corrupted but just that Crafty refused to use some of it early on in the match. The previous series of Comet- Crafty matches started out okay from a fairly large book of Kasparov games but eventually after 50 or so games, Crafty refused to use it's book any more. That was the series where both programs were pondering and I had crashes all the time in Winboard. The way you have implemented book learning in Crafty forces users to rebuild books each time to make Crafty's book larger. This is against some lofty AI principles of having the machine learn by itself. It would truly be impressive if Crafty could make it's book larger by itself based on results of its games (MChess 7.1 can do this) and even more impressive if it could start out from no book at all and build it's book this way. . I realize that to implement this could mean different versions for servers and others because of the fast time constraints on servers, but there are eons more Craftys out in the rest of the cyber world than there are on servers. I guess it depends on your goals with Crafty. Based on your results, how many ELO points is Crafty suffering from playing on a Pentium 166 as compared to a Pentium Pro or beyond. I don't need the point comparison in relation to speed of machine but just the extra points that Crafty gets when the 64 bit code is optimized. In other words Comet against Crafty on a pentium pro as compared to Comet against Crafty on a Pentium 166Mhz. How many more ELO does Crafty get this way after factoring out the speed increase from a 166 to a Pentium Pro 200? -- Komputer Korner
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.