OpenChess

mdtr

i have several games where Stockfish 17/04/16 (default parameters, single cpu) drew with 3-fold winning games or at least with a good advantage, the last one against rybka 2.3:

1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Ba4 Nf6 5. O-O Be7 6. Re1 b5 7. Bb3 d6 8.
c3 O-O 9. h3 Bb7 10. d4 Re8 11. Nbd2 {+0.06/13 ...

mdtr

both 1CPU 64 bit, no tablebases

Code: Select all

Rank Name                  Elo    +    - games score oppo. draws
   1 Houdini 3 x64           4   19   19   200   52%    -4   54%
   2 Stockfish 221013 64    -4   19   19   200   49%     4   54%

houdini 3 scored 51.5% (+49 =108 -43) +10 elo

mdtr

both 1CPU 64 bit, no tablebases

Code: Select all

Rank Name                  Elo    +    - games score oppo. draws
   1 Stockfish 221013 64     1   16   16   300   50%    -1   52%
   2 Houdini 3 x64          -1   16   16   300   50%     1   52%

stockfish won by a nose
50.1% (+72 =157 -71) +1 elo

mdtr

i'm not talking about hardware, i'm wondering how much stronger (let's compare it with houdini 3) a chess engine could be (staying with the actual chess engine "paradigm", nothing too strange)
i'd go for +500 elo :)

edit: i'm asking about this because i remember i could not imagine an engine ...

mdtr

both 1CPU 64 bit, no tablebases

Code: Select all

Rank Name                  Elo    +    - games score oppo. draws
   1 Houdini 3 x64           5   16   16   300   52%    -5   49%
   2 Stockfish 181013 64    -5   16   16   300   48%     5   49%

-10 elo, i was expecting a better result for SF this time :/

mdtr

Magdalina wrote:Thanks for posting. Marco's team would see this soon. You did not change the contempt value etc? that was done with the fish's default parameter setting?

yes both engines were in their default parameter setting, except houdini Threads=1, obviously

mdtr

stockfish is gaining elo because of tuning and tweaks, its development will be slower and slower since it's difficult to tune a extremely tuned engine... SF needs new ideas and those idea are not showing up

also you don't know how strong is the houdini dev version since it's private....

http ...

mdtr

stockfish is gaining elo because of tuning and tweaks, its development will be slower and slower since it's difficult to tune a extremely tuned engine... SF needs new ideas and those idea are not showing up

also you don't know how strong is the houdini dev version since it's private....

mdtr

that's not a pgn so we can't help you

mdtr

it looks easy to program your chess engine not to do this thing...

[Event "1 Minutes/Game"]
[Site "Engine Match"]
[Date "2013.10.16"]
[Round "2"]
[White "Stockfish 141013 64"]
[Black "Houdini 3 x64"]
[Result "0-1"]

1. e4 {book 0s} e6 {book 0s} 2. d3 {book 0s} d5 {book 0s}
3. Nd2 {book 0s} Nf6 ...

OpenChess

OpenChess

Search found 18 matches

stockfish 3-fold rep bug

2 sec/move: Stockfish 221013 64 - Houdini 3 x64

1 sec/move: Stockfish 221013 64 - Houdini 3 x64

how strong a chess engine could be?

1 sec/move: Stockfish 181013 64 - Houdini 3 x64

Re: SF lost on time because didn't want to accept a draw

Re: The Difference Between Stockfish & Houdini.

Re: The Difference Between Stockfish & Houdini.

Re: What engines see this line?

SF lost on time because didn't want to accept a draw