I have recently entered a computer chess tournament. My engine, Spike 1.1, is pretty strong, but not nearly as strong as some of the other competitors. So my question is this: What would be the best method to upgrade a UCI engine, resulting in a drastic ELO increase?
I know it's possible, because one of my friends firends on FICS, his name is EJD i think, updgraded his version of Spike 1.1 so that it went from 2400 to 2700, at the top of the rating list!
Originally posted by LanndonKaneTry Prodeo.
I have recently entered a computer chess tournament. My engine, Spike 1.1, is pretty strong, but not nearly as strong as some of the other competitors. So my question is this: What would be the best method to upgrade a UCI engine, resulting in a drastic ELO increase?
I know it's possible, because one of my friends firends on FICS, his name is EJD i th ...[text shortened]... aded his version of Spike 1.1 so that it went from 2400 to 2700, at the top of the rating list!
It will woop the tar outa spike. It's also free. Great for arena chess.
http://members.home.nl/matador/prodeo.htm
- Joined
- 21 Apr 06
- Moves
- 4211
Originally posted by MctaytoI guess they enjoy fine tuning the engines to get that little bit extra out of them. Another reason might be (and one iv considered) is setting the opening book to play just one line mainly little know ones like. 1.e4 e6 2.b3 d5 3.Bb2 and then leaving the engine to play it against other peoples engines for say a month and seeing how the games went with win/loss% of moves. Also a lot of the engines can learn to some extent while this might not change their overall elo by say more than one or two points, you could improve its performance down the lines you get it to play often.
I don't see the point of a game that is engine v engine ??
Am I missing something, what is the point ??
Originally posted by BedlamSetting the engine priority to highest will do nothing unless you have other things going on in the background. And you really shouldn't if you want good play out of your engine.
Get Rybka, wicked little engine, tends to duff almost anything.
Use XP to set the engines priority to the highest.....get a faster processor....use an engine opening book.....tablebases.
Originally posted by LanndonKaneDo you have tablebases installed and associated with the engine?
I have recently entered a computer chess tournament. My engine, Spike 1.1, is pretty strong, but not nearly as strong as some of the other competitors. So my question is this: What would be the best method to upgrade a UCI engine, resulting in a drastic ELO increase?
I know it's possible, because one of my friends firends on FICS, his name is EJD i th ...[text shortened]... aded his version of Spike 1.1 so that it went from 2400 to 2700, at the top of the rating list!
Bedlam is right, if you have the money, get Rybka. On average, it will beat all the other engines, all other conditions being equal. (This isn't just my opinion, the CEGT engine testing site confirms this.) If you don't want to spend money on a new engine, then your best bet is probably the newest version of Toga (I think it's Toga II 1.2.1a). Toga is probably the strongest free engine. Spike 1.2 (yes a new version is out!) and Gambit Fruit are almost as good as Toga. I've used Toga, and it's relatively bug free. I haven't used Rybka, but I've heard others complain at times about various bugs, at least on the Playchess server (I'm not sure whether the bugs are caused by the engine or the Playchess GUI).
And yes, tablebases will help a little. (But don't expect a huge improvement.) Don't bother trying to generate them, it will take you forever. (Trust me, I actually tried.) There are a few web sites where you can download them for free. But be warned, a 5 piece tablebase takes a little over 7 gigabytes of hard drive space. (4 piece is only about 30 megabytes.) And don't forget, the 2 kings and any pawns also count as pieces.
And a decent opening book helps. I've used the free Sedat and Harry Schnapp books, and they're not bad. I'm sure the books that come with commercial engines like Fritz, CM, etc aren't bad either. Some people claim that custom made books can help a rating greatly, but it is very labor intensive, and I don't have the time or desire to do that.
Also, the hardware makes a difference. A faster CPU translates to a higher rating. Not that you would run right out and buy a new PC just for chess, but just be aware that hardware also matters.
If you need me to give you any links to any of the stuff I've mentioned, just ask. (I just didn't want to spend time looking them all up if you could easily find them yourself.)