Forums

1.e4 f5 : me and my oversized ego vs all of you

Sort:
johnyoudell

How do the engines evaluate the current position.  Is white still a full pawn up (or better) or have the engines eroded white's lead?

Aetheldred

btickler is right, Komodo TCEC x64 is slightly stronger than Houdini 4 pro x64. I was surprised when I saw it advertised on ICC as the strongest engine. I gave it a try and it's slightly better. The main problem with Komodo is it's incredibly slow.

johnyoudell, the engines are gaining ground, slowly but surely. White is now about +0.87.  

DiogenesDue

I did find this on the CCRL site:

http://kirill-kryukov.com/chess/discussion-board/viewtopic.php?f=7&t=7519

STOCKFISH 040514 64-bit 8CPU v HOUDINI 4 64-bit 8CPU

Intel Xeon X5430x2 Octal
ChessGUI
2048mb hash where possible
3-4-5 piece Syzygy tablebases
Ponder off
ForumUsersMix.cgb book (12 move restriction)
40 moves in 29 minutes repeating
First to win 6 games wins the match.

Re: Stockfish 040514 64-bit v Houdini 4 64-bit 8CPU (Match)

Postby Graham Banks » Mon May 05, 2014 8:01 pm

Won by Stockfish 040514 64-bit 8CPU by +6 -1 =12
 
(red highlights added by me)
 
I would call +6 -1 =12 decisive for Stockfish ;).
Komodo looks like it lost to Houdini using 40/29 (not sure why they use 29 minutes),
but as pointed out, maybe Komodo is just not as good in faster time controls.  
For 30+ ply analysis, though...
The TCEC matchups are using 40/120 with 30 second increment,
so much closer to the kind of analysis being done in this thread.
Irontiger

...Qf7 it is, then.

My answer is 14.Ne5 with the idea to follow-up by Ndf3 after some possible trades on e5.



johnyoudell

Thanks, Aetheldred.

matthew_b_rook

Chessmaster plays Nxe5

Aetheldred

Komodo @33 plyes likes Nxe5 as well.

+0.45. It seems we may have a chance after all.

Aetheldred

All engines will play Nxe5.

Masterjatin

seems to me like a reversed King's Gambit..... and to achieve that you can not play e5

Masterjatin

btickler wrote:

I'll say it again with gusto ;)...whatever Stockfish and/or Komodo pick (at 30+ plies) needs to be the default move over Houdini, because Houdini is a weaker engine.  Houdini has not been on top since sometime early last year?  All the other engines are pretty much irrelevant.

Houdini would make a nice tiebreaker if Stockfish and Komodo disagree, though.

Houdini is, at the moment, the strongest engine available, with Stockfish on the second and Komodo on the fourth with Rybka some late, if you know about latest CCRL rankings

Masterjatin

Well, if you can defeat all engines being used here, you need Titan V running Houdini 4 pro x64 and Stockfish DD x64. then, I guarantee your ego another 30 years until I make a real chess engine.

Otomun

Stockfish likes Nxe5.

Aetheldred

Masterjatin, we are using last year's engines. The ones that are competing now will be for sale in 2015 or in a couple of months (Stockfish 5).

Komodo TCEC won the championship last year, that's the engine I've got (I've got all of them, really). My computer is the same Komodo used in the 2013 championship.

It's true Houdini 4 was a beta version then, but the one in competition now is a different one.

DiogenesDue

Houdini is, at the moment, the strongest engine available, with Stockfish on the second and Komodo on the fourth with Rybka some late, if you know about latest CCRL rankings

Who cares about CCRL rankings?  CCRL only runs 40/40 and 40/4 time controls.  If you are using an engine to run analysis then you want the engines that are better with more time to spend.  That's TCEC rankings.

If you have read the previous page's posts then you've already seen the proof that Stockfish is better than Houdini both in TCEC season 6 at 40/120 (with 30 second increment), and in the latest CCRL matchup at 40/29 time controls, in which Houdini managed to win only 1 game out of 19 played against Stockfish (Stockfish beat Houdini 6 times).  Both of these are in May...this month.  Can't be anymore current than that.

Stockfish is open source, so as soon as the Stockfish team checks in the code for a new version, you should be able to go get it and compile it if you are so inclined.  You shouldn't have to wait for an "official" release.  Houdini is a commercial product that does not make their current code public, and so lags far behind in release cycles.

Masterjatin

What I see:

14...Nxe5 15.dxe5 Bd7 16.Nf3 h6 17.Bh4

No engine though I could use Stockfish.....

Masterjatin
btickler wrote:

Who cares about CCRL rankings?  CCRL only runs 40/40 and 40/4 time controls.  If you are using an engine to run analysis then you want the engines that are better with more time to spend.  That's TCEC rankings.

If you have read the previous page's posts then you've already seen the proof that Stockfish is better than Houdini both in TCEC season 6 at 40/120 (with 30 second increment), and in the latest CCRL matchup at 40/29 time controls, in which Houdini managed to win only 1 game out of 19 played against Stockfish (Stockfish beat Houdini 6 times).  Both of these are in May...this month.  Can't be anymore current than that.

Stockfish is open source, so as soon as the Stockfish team checks in the code for a new version, you should be able to go get it and compile it if you are so inclined.  You shouldn't have to wait for an "official" release.  Houdini is a commercial product that does not make their current code public, and so lags far behind in release cycles.

The problem is that I don't know about much rankings. On Stockfish page I could find CCRL only, and there in 40/40 Houdini was at top. That too I could not get pgn because the file size never came in managers.

I will design an engine that will play so aggressively that noone will get enough space and it will strike weakest point on the board.

Well, can you suggest me any engine rankings which gives commentated or non commentated pgn, though?

Aetheldred
btickler wrote:

Houdini is, at the moment, the strongest engine available, with Stockfish on the second and Komodo on the fourth with Rybka some late, if you know about latest CCRL rankings

Who cares about CCRL rankings?  CCRL only runs 40/40 and 40/4 time controls.  If you are using an engine to run analysis then you want the engines that are better with more time to spend.  That's TCEC rankings.

If you have read the previous page's posts then you've already seen the proof that Stockfish is better than Houdini both in TCEC season 6 at 40/120 (with 30 second increment), and in the latest CCRL matchup at 40/29 time controls, in which Houdini managed to win only 1 game out of 19 played against Stockfish (Stockfish beat Houdini 6 times).  Both of these are in May...this month.  Can't be anymore current than that.

Stockfish is open source, so as soon as the Stockfish team checks in the code for a new version, you should be able to go get it and compile it if you are so inclined.  You shouldn't have to wait for an "official" release.  Houdini is a commercial product that does not make their current code public, and so lags far behind in release cycles.

Good info, thanks!

DiogenesDue
Masterjatin wrote:

The problem is that I don't know about much rankings. On Stockfish page I could find CCRL only, and there in 40/40 Houdini was at top. That too I could not get pgn because the file size never came in managers.

I will design an engine that will play so aggressively that noone will get enough space and it will strike weakest point on the board.

Well, can you suggest me any engine rankings which gives commentated or non commentated pgn, though?

Heh, you can probably get the Stockfish source to use as your starting point and if you can improve on their code, your engine will surely be #1 ;).

Try the TCEC page.  For example, you can watch the final going on as we speak between Stockfish and Komodo, and from the File menu you save a live PGN as they make their moves.

For example, this is round 23's current state (it's move 14) in PGN form:

[Event "TCEC Season 6 - Superfinal"] [Site "http://tcec.chessdom.com"] [Date "2014.05.23"] [Round "23"] [White "Stockfish 170514"] [Black "Komodo 7x"] [Result "*"] [BlackElo "3155"] [ECO "D11"] [EventDate "2014.05.19"] [Opening "Slav: 4.Qc2"] [Termination "unterminated"] [TimeControl "7200+30"] [WhiteElo "3164"] { WhiteEngineOptions: Protocol=UCI; Hash=16384; Threads=16; SyzygyPath=C:/EGTB/Syzygy/; Emergency Base Time=1000; OwnBook=false; Ponder=false;, BlackEngineOptions: Protocol=UCI; Hash=16384; Table Memory=256; Threads=16; SyzygyPath=C:/EGTB/Syzygy/; Drawscore=0; OwnBook=false; Ponder=false; } 1. d4 { book } d5 { book } 2. c4 { book } c6 { book } 3. Nf3 { book } Nf6 { book } 4. Qc2 { book } dxc4 { book } 5. e4 { book } b5 { book } 6. b3 { book } cxb3 { book } 7. axb3 { book } e6 { book } 8. Bd2 { book } Bb7 { book } 9. Ba5 { d=35, pd=Qc8, mt=00:03:04, tl=02:01:25, s=16948 kN/s, n=3120223590, pv=Ba5 Qc8 Be2 Na6 Nbd2 Be7 Qc3 O-O b4 c5 bxc5 Nxe4 Nxe4 Bxe4 Bxb5 Nc7 Bc4 Nd5 Qd2 Bxf3 gxf3 Qd7 O-O Rab8 Rfc1 Bf6 Rd1 g6 Rdb1 Rfc8, tb=0, R50=49, wv=0.15, } Qd7 { d=27, pd=Nbd2, mt=00:04:07, tl=02:00:22, s=15532 kN/s, n=3849388280, pv=Qd7 Nbd2 Na6 Qc3 Bd6 Ne5 Qc8 b4 O-O Be2 Nd7 Nxd7 Qxd7 e5 Be7 O-O f6 exf6 Rxf6 Ne4 Rf4 f3 Rf5 Rfd1 Rd5 g3 Rf5 f4 Qd5 Bd3 Bf6 Bc2, tb=4, R50=49, wv=0.23, } 10. Nbd2 { d=38, pd=Na6, mt=00:02:59, tl=01:58:56, s=17049 kN/s, n=3066372659, pv=Nbd2 Na6 Qc3 Bd6 Ne5 Qc8 Be2 O-O b4 Nd7 Nd3 Qb8 e5 Be7 O-O Nc7 Bxc7 Qxc7 Ne4 Nb6 Ndc5 Nd5 Qd2 Rfd8 Rfc1 a6 g3 Qb8 f4 Qc7 Ra3 h6 Raa1 Rac8 Bf3 Kh8 Rc2 Kg8 Rcc1, tb=0, R50=48, wv=0.18, } Be7 { d=29, pd=b4, mt=00:04:41, tl=01:56:10, s=17154 kN/s, n=4835855814, pv=Be7 b4 O-O Bd3 Na6 Qc3 Bd8 O-O Bb6 Nb3 h6 Ne5 Qc7 g3 Rfd8 Rfc1 Rac8 Ra2 Ra8 Ra3 Nd7 Nf3 Nf8 Nc5 Nxc5 dxc5 Bxa5 Rxa5 f6 e5 Qd7 Bc2, tb=61, R50=48, wv=0.31, } 11. b4 { d=38, pd=O-O, mt=00:07:07, tl=01:52:18, s=18023 kN/s, n=7709979659, pv=b4 O-O Be2 Na6 Qc3 Bd6 Qb2 Ne8 Ne5 Qe7 Nd3 f5 e5 Bc7 Nb3 Bb6 Ndc5 Nxc5 Nxc5 Bxc5 bxc5 Nc7 O-O a6 Bf3 Rad8 Rfb1 Rd7 Ra3 Nd5 Qd2 g6 Rbb3 Qg7 h3 Qe7 Bxd5 Rxd5, tb=0, R50=47, wv=0.19, } Na6 { d=29, pd=Qc3, mt=00:03:48, tl=01:52:51, s=18168 kN/s, n=4154964580, pv=Na6 Qc3 O-O Bd3 Bd8 O-O Bb6 Nb3 h6 Nc5 Nxc5 dxc5 Qc7 g3 Rfd8 Bc2 Bxa5 Rxa5 a6 Rfa1 Qd7 Kg2 Qc7 Re1 Rac8 h3 Rb8 Rf1 Qe7 Ne5 Qc7, tb=62, R50=50, wv=0.30, } 12. Qc3 { d=39, pd=O-O, mt=00:01:59, tl=01:50:48, s=17587 kN/s, n=2108019758, pv=Qc3 O-O Be2 Bd6 Qb2 Ne8 Ne5 Qe7 Nd3 f5 e5 Bc7 Nb3 Bb6 Ndc5 Nxc5 Nxc5 Bxc5 bxc5 Nc7 O-O a6 Bf3 Rad8 Rfb1 Rd7 Ra3 Nd5 Qd2 g6 Rab3 Rf7 Ra1 Rf8 Qd3, tb=74, R50=50, wv=0.19, } O-O { d=28, pd=Bd3, mt=00:03:45, tl=01:49:35, s=18632 kN/s, n=4208860283, pv=O-O Bd3 Bd8 O-O Qe7 Rfb1 Bb6 Nb3 Rfd8 Bc2 h6 Nc5 Qc7 Ne5 Nxc5 dxc5 Re8 g3 Rad8 f4 Bxa5 Rxa5 a6 Rf1 Qe7 Bd3 Qc7 Rf2 Qe7 Rc2 Nd7 Nf3, tb=31, R50=49, wv=0.29, } 13. Be2 { d=38, pd=Bd6, mt=00:02:00, tl=01:49:17, s=18217 kN/s, n=2199067155, pv=Be2 Bd6 Qb2 Ne8 Ne5 Qe7 Nd3 f5 e5 Bc7 Nb3 Bb6 Ndc5 Nxc5 Nxc5 Bxc5 bxc5 Nc7 O-O a6 Bf3 Rad8 Rfb1 Rd7 Ra3 Nd5 Qd2 g6 Rab3 Rf7 Ra1 Rf8 Qd3, tb=0, R50=50, wv=0.19, } Bd8 { d=26, pd=Bxd8, mt=00:03:08, tl=01:46:57, s=17652 kN/s, n=3326897385, pv=Bd8 Bxd8 Rfxd8 O-O Qe7 Rfb1 h6 Bd3 Nd7 Nb3 Rac8 Nc5 Ndxc5 bxc5 Ra8 Ra5 Nc7 Ne5 a6 Re1 Ne8 Bc2 Nf6 Rea1 Rac8 Nf3 Rc7 Qd3 b4 e5, tb=6, R50=50, wv=0.28, } 14. Bxd8 { d=36, pd=Rfxd8, mt=00:05:04, tl=01:44:43, s=18352 kN/s, n=5583194441, pv=Bxd8 Rfxd8 O-O Qe7 Rfb1 Nd7 g3 Kh8 Bf1 Nb6 Ra5 Na4 Qa3 Nb8 e5 Nd7 Ne4 Ndb6 Bd3 Nd5 Nd6 a6 Nd2 Rxd6 exd6 Qxd6 Ne4 Qe7 Nc5 Nxc5 dxc5 Qf6 Qb2 Nc3 Re1 Rd8 Re3 g6 Ra1, tb=2, R50=49, wv=0.22, } Rfxd8 { d=25, pd=O-O, mt=00:01:23, tl=01:46:03, s=18178 kN/s, n=1515462887, pv=Rfxd8 O-O h6 Rfb1 Nc7 Qc5 Na6 Qc2 Qe7 Qc3 Nd7 Ra5 Nab8 Rba1 Nb6 Nb3 Na4 R5xa4 bxa4 Nc5 a5 bxa5 Na6 Bxa6 Bxa6 Rxa4 Be2 Ne5 Bb5 Ra2 Qg5, tb=2, R50=50, wv=0.26, } *

Under the File menu you can also choose Archive Mode and bring up all the completed games.  If you use the directory tree and look at Season 6 - Stage 4, for example, you can download it as one giant PGN, or download a zip file of the games.  

DiogenesDue

P.S. Stockfish is beating Komodo pretty handily right now with 22 games complete...

+4 -0 =18

I'm not sure how many rounds the final is.

xakean-ski

Spark agrees with Nxe5