Search found 637 matches

by Joerg Oster
Mon Feb 18, 2019 5:02 pm
Forum: Computer Chess Club: General Topics
Topic: EG Cook
Replies: 14
Views: 524

Re: Cooking the Cook (not)

Suggested repair is moving the Knight one square down :roll: Now it is become a simple Mate in 10 problem... Won't get a prize. It turned out not to be easy for Kaissa though it had the first move wrong. At least that is what it thinks. So, for engines: who finds the right move fastest and sees Mat...
by Joerg Oster
Sat Feb 09, 2019 9:00 am
Forum: Computer Chess Club: General Topics
Topic: McCain X1 Source Available
Replies: 25
Views: 1196

Re: McCain X1 Source Available

We now have some Windows and macOS exe's. The Linux flavors have not compile properly and the reason is unknown at this time. The makefile contains the bench for McCain - if you compile the Linux flavor and the the the total node count matches the makefile (4914655) , please let me know. Also you c...
by Joerg Oster
Sun Feb 03, 2019 11:30 am
Forum: Computer Chess Club: General Topics
Topic: start your engines...
Replies: 14
Views: 919

Re: start your engines...

q4bk1/pb4p1/1p2pp2/3n3p/2NP4/1P3BP1/PB3P1P/R1R3K1 b - - 0 1 ; best move Nf4 from a 2017 ICCF game, found on the HIARCS ICCF forum ... It takes Zander more than 7 minutes to find Nf4, 3 Threads, 1 GB Hash, 6-syzygy: info depth 36 seldepth 64 multipv 1 score cp 121 nodes 1727920009 nps 4303039 hashfu...
by Joerg Oster
Thu Jan 24, 2019 3:14 pm
Forum: Computer Chess Club: General Topics
Topic: a test position that I composed
Replies: 13
Views: 915

Re: a test position that I composed

This one looks really tough. If you force Kb2 Kd7 then SF gets it quickly, but even SF struggles otherwise. EDIT: I spoke too soon, SF manages in the original position. SF10 from Lichess got it right. Not single-threaded and without TBs. 128 MB Hash, 1 Thread, no TBs, 5 min: info depth 77 seldepth ...
by Joerg Oster
Sat Jan 05, 2019 3:14 pm
Forum: Computer Chess Club: General Topics
Topic: Ancient Crafty Flexes His Muscles
Replies: 12
Views: 1129

Re: Ancient Crafty Flexes His Muscles

Here are a couple of positions from a test set I am currently working on: r7/1b1r4/k1p1p1p1/1p1pPpPp/p1PP1P1P/PP1K4/8/4Q3 w - - bm Qa5+; ce -1496; c5 "dorszcz.0525"; 8/3k1b1r/2p1p1p1/rpPpPpPp/p2P1P1P/P2K4/8/4B3 w - - bm Bb4; ce -1201; c5 "dorszcz.0258"; Try it with Stockfish and you will feel great...
by Joerg Oster
Sat Dec 22, 2018 1:52 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: UCI2WB 4.0
Replies: 47
Views: 2136

Re: UCI2WB 4.0

I see one 'isready' command which is answered by Stockfish with 'readyok'. I don't see a second 'isready' command. I admit that the logging by UCI2WB is a bit messy, as both commands sent to the engine as remarks about the adapter's internal state appear prefixed by just the same '#' symbol. But th...
by Joerg Oster
Sat Dec 22, 2018 12:10 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: UCI2WB 4.0
Replies: 47
Views: 2136

Re: UCI2WB 4.0

I now have made a test version of UCI2WB that reports what it is doing on an independent log file. This gives me the following when starting Stockfish 10, pasting a game, and then switching on analysis mode: # queue 'protover', searching=0 # command protover # uci >> Stockfish 10 64 by T. Romstad, ...
by Joerg Oster
Tue Dec 18, 2018 10:19 am
Forum: Computer Chess Club: General Topics
Topic: Komodo 12.3 is out
Replies: 76
Views: 7064

Re: Komodo 12.3 is out

Also please note that although 12.3 MCTS is still nearly a hundred elo below normal 12.3, it can already defeat 12.3 if both are in MultiPV = 2 (or more) mode. For this reason, I have switched to using the MCTS version myself when analyzing games or opening s. "Version 12.3 will probably be the las...
by Joerg Oster
Fri Nov 09, 2018 8:48 pm
Forum: Computer Chess Club: General Topics
Topic: Komodo MCTS scales worse with TC than Komodo A/B?
Replies: 22
Views: 2559

Re: Komodo MCTS scales worse with TC than Komodo A/B?

First, I checked that K12.2 AB and K12.2 MCTS behave very similarly in all sorts of variants and phases of the game, so Leela weird behavior in variants and endgames (weaker than regular chess performance) is due mostly to its eval and not to its MCTS search. K MCTS is simply about 100-150 Elo poin...
by Joerg Oster
Tue Nov 06, 2018 10:07 pm
Forum: Computer Chess Club: General Topics
Topic: Komodo 12.2 Announcement
Replies: 54
Views: 6541

Re: Komodo 12.2 Announcement

That does look odd. I will take a look. I have not seen that before. Was this from the start position? Mark, I also just tried the Windows version. Same issue. uciok setoption name Use MCTS value true ucinewgame isready readyok go infinite info depth 10 seldepth 5 time 2002 score cp 33 nodes 157 np...