Author: Don Dailey
Date: 18:04:00 01/26/98
Go up one level in this thread
I'll try to post all the findings in the next 2 or 3 days. I'm running Cilkchess at 1 hour per position on several alphas, so I should have some data to contribute. I think it takes more that just long runs though, I think each problem has to be manually inspected to get a true sense of what is going on, which you guy's are doing a good job of. The most useful format for me is for each one to break down their findings into simple lists, 1 per line. This would help me a whole lot if you guy's would make the effort, compiling these is a lot of work. If each of you can give me a SEPARATE list of: 1) easy. 2) Only 1 solution, completely correct. 3) Multiple solutions completely correct. 4) Bad problems. 5) Flaky problems. Here is an example of how I would prefer to see your lists: (I'm making up the data) Put any remarks you want after a semicolon. 1 3 4 6 ; remarks 7 ; remarks For giving solutions do this: 1 exd5 2 Nf3 Bc4 ; for multiple solutions If you do not give me this information I will pick it tediously out of the postings, but this will help me a lot. Any solutions you give are understood to be the ones you believe are correct, not neccesarily the ones given in the set. Also you do not necessarily have to give me each of the 5 categories I listed, I'll take any information you supply to me. I will compile this information and cross check against everyones results looking for disagreement. If any two of you send disagreeing results I'll find them and I will post everything back to you. For the first pass we are not pruning anything except incorrectness. Since there is a lot of disagreement on what the set should be I want all this information so I can easily commpile any variation we agree upon. I would prefer this info emailed to be instead of posted since it might get a little messy and this gives me a better record. - Don On January 26, 1998 at 17:47:48, Amir Ban wrote: >On January 26, 1998 at 10:52:04, Ed Schröder wrote: > >>>Posted by Amir Ban on January 26, 1998 at 04:14:19: >> >>>I've no problem with what's going on. This will take 2-3 weeks more at >>>this pace, but don't take this as a commitment. I now realize that I >>>should really have the book to do this, but having you there with Rebel >>>is nearly as good, thanks. I am leaving the task of gathering the >>>results and editting the suite to others. I think Don took the >>>assignment. >> >>Manfred Rosenboom is planning to release a new version of EPD2diag >>soon. I have the beta version which allows me to automatically analyze >>any EPD set on any preferred time level using Rebel9. >> >>If this is of any help let me know. I can run the set overnight with >>the new EPD2diag and email you the results. See for yourself. >> >>- Ed - >> >> >>>Amir > >Thanks Ed. I don't think I need it. I have my own 60-seconds run that I >use as a base, and that leaves me with about 230 problems to look at. >Maybe another program test run would cut this by 20-30, but that still >doesn't change much for what I'm doing. Don and Bruce have everyone's >results, maybe they can use yours too. > >I think it would be useful if someone posts the intermediate conclusions >here and let everyone take a look and contest the finding if they wish. >I did that for the first 200 a while back. > >Amir
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.