I didn't include the free Rybka program because I haven't put it on the new computer. I don't see any need for it. As far as I can see there's really no reason to buy anything when you can get everything you really need free.
Thursday, September 30, 2010
Engine Tournament Test
This isn't very scientific but this morning I ran a double round 5 minute tournament and the results were:
I didn't include the free Rybka program because I haven't put it on the new computer. I don't see any need for it. As far as I can see there's really no reason to buy anything when you can get everything you really need free.
I didn't include the free Rybka program because I haven't put it on the new computer. I don't see any need for it. As far as I can see there's really no reason to buy anything when you can get everything you really need free.
Subscribe to:
Post Comments (Atom)
Very interesting experience !
ReplyDeleteOne of the most I've done was ChessMaster 11 vs. Crafty 23.01. Well, Crafty is a lightweight engine, but very strong and robust, and it was capable of beating ChessMaster even being a very small piece of code !
Truly remarkable ! I enjoy these computer chess benchmarks.
Are these games worth studying in order to learn some chess brilliancies or are they only a fight between algorithms to see which one is stronger ?
Thanks for sharing your experiences !
Interesting question and I believe the answer is that there is little value in trying to learn anything by studying computer generated games. Engines don’t plan and their analysis does not give any insight as to why they played a particular move.
ReplyDeleteI have read many annotations by strong players who have checked their analysis with an engine and it’s not unusual for them to disagree with the engine’s evaluation of a position in situations where there are no tactical threats.
I remember seeing one game by a CC GM where he defeated an opponent who was relying entirely on computer generated moves and he explained why the computer’s moves were often faulty. One CC IM, also a 2300 OTB master, stated that in the absence of tactics, he disregards engine evaluations. This makes sense if you look at CC at the top level…the top CC players are those who know enough about chess that they are able to get past those who rely entirely on engine use.
I have played for Lechenicher SchachServer in several tournaments. This site allows computer use and everybody uses one. I was allowed to start at my then official CC rating of ~2100. The results were less than spectacular: +10 -11 =10 and a 300 point drop in rating! The reasons are that I had an old, slow computer, outdated software, did not allow the engine enough time to ponder it’s moves (some players run their engines overnight!) and finally, in some cases where I disagreed with the engine’s selection, I chose other moves; sometimes successfully and sometimes not. Of course there's also the too painful to think about possibility that I'm just not good enough to compete at the higher levels of CC play!
Obviously, you can’t just buy an engine and compete at the top levels in CC unless you meet the following criteria: 1) are a very strong player 2) know how to guide the game into positions where engines do not play well 3) have a strong computer and the latest software 4) have a LOT of patience for analysis. Even in CC play below that level, on every site I’ve played on, at some level (usually around 2200), I’ve started running into about 50% engine users.
I think you are correct in that engine vs. engine play is a battle of algorithms.