lkaufman wrote:I thought the previous poster got it right and just doesn't care about anything except blitz chess. Perhaps I am wrong; only he can clarify this.
No,I thought it was 40 moves in 2 minutes time control....I hate blitz time controls and for that I think that Komodo's performance is even more impressive
Dr.D
_No one can hit as hard as life.But it ain’t about how hard you can hit.It’s about how hard you can get hit and keep moving forward.How much you can take and keep moving forward….
I didn't mean to mislead anyone by the caption. I would much rather have Komodo be considered the best program at 40/2 hours than the best program at 40/2 minutes, if we can only have one of the two options! Really, the main reason to care about blitz ratings is that it takes so long to get an accurate rating at a long time control that by the time it is done, the program being rated is no longer the current version.
Thanks. The real sense of accomplishment is having achieved this level without starting from someone else's codebase. When we started on this adventure (with DOCH) I think the most we could even hope for was to be second behind Rybka 3 someday. I would never have guessed we would reach today's level by now, especially given that we never made any single advance that was worth more than twenty elo points. I can't guess where the next real gains will be found, but I have no doubt that we will continue to make progress.
lkaufman wrote:Thanks. The real sense of accomplishment is having achieved this level without starting from someone else's codebase. When we started on this adventure (with DOCH) I think the most we could even hope for was to be second behind Rybka 3 someday. I would never have guessed we would reach today's level by now, especially given that we never made any single advance that was worth more than twenty elo points. I can't guess where the next real gains will be found, but I have no doubt that we will continue to make progress.
So for post game analysis of an OTB game is it good to let Komodo look at each position for about 3-4 minutes? Is that productive enough?
There's no magic amount of time that's best; the longer the better the quality. I used an average of 15 minutes for the analysis I did for my new book. What this result suggests (subject to sample error of course) is that if you have at least 3 minutes available for the computer to think, Komodo might be your best choice (assuming equal number of cores). If you want an answer in a few seconds Houdini might be a better choice until we catch it in the blitz lists too.
No, Critter 1.4 and SF 2.2.2 came out after Komodo 4. Also note that Komodo 3 was skipped. As for Houdini 2.0, yes it was before Komodo 4, but it got a lower rating on their 40/20 list than Houdini 1.5, so they had no compelling reason to test it at 40/2. I would say Critter 1.4 is the most deserving of being tested next, considering its large gain over Critter 0.9 based on 40/20 list. I think they are rightly following the policy of only testing new versions at 40/2 when the version has shown a substantial gain at 40/20. Otherwise they'll never have decent samples of any engine.
