CEGT - rating lists April 18th 2010

[Werner]

Hi all ,
our rating lists are now online and can be found under the attached links.

40 / 20:
Today we have a little anniversary:
- we cracked the 400.000 mark and
- we have a new nr. 2 in our list!
- we added 1383 games made with 30 different engines to our list. See more in our list "Games of the week". In total our 40/20 list is based now on 401.282 games.

New engines:

New in our list is Stockfish 1.7.1 x64. With 4CPUs we have a rating of 3156 in 737 games. This is +39 to v. 1.6 and 25 points behind Rybka 3 now: Our new nr. 2!
The 1CPU version scored with 3080 elos in 525 games: +50 to version 1.6 (only 18 points behind Rybka 3).

Updated engines:

Code: Select all

1 Rybka 3 x64 4CPU 3181 +11 -11 3019 games (+1)
4 Naum 4.2 x64 4CPU 3139 +15 -15 1146 games (+3)
52 Komodo 1.0 x64 2972 +13 -13 1475 games (+-0)
354 Daydreamer 1.75 x64 2707 +21 -21 614 games (+2)

Blitz Update from last Friday
New games: 8000
All games: 623.428

New engines

We had a main update with Stockfish 1.7.1 and Spark 0.4:

13 Stockfish 1.7.1 x64 1CPU 3104 19 19 900 games: +65 to v. 1.6 and 14 points behind Rybka 3!
14 Stockfish 1.7.1 w32 2CPU 3100 27 27 400 games: +59 to v. 1.6
33 Stockfish 1.7.1 w32 1CPU 3032 21 21 800 games: +56 to v. 1.6
129 Spark 0.4 x64 1CPU 2924 19 19 800 games: +53 to v. 0.3
189 Spark 0.4 w32 1CPU 2865 20 20 800 games: +30 to v. 0.3
641 DanaSah 4.37 2535 26 26 500 games: -10 to v. 4.24
694 Philou 3.14.1 2482 26 26 500 games: +10 to v. 3.10
698 Arasan 9.5 2477 27 27 500 games: +15 to v. 9.0

Updated engines:

Code: Select all

7 Naum 4.2 x64 4CPU 3154 +12 -12 2000 games (-6)
21 Naum 4.2 w32 2CPU 3064 +16 -16 1200 games (-14)
25 Naum 4.2 x64 1CPU 3047 +15 -15 1200 games (+8)
48 Naum 4.2 w32 1CPU 3007 +13 -13 2200 games (-10)

A big „Thank you“ to all testers as usual!

links:
40/20: http://www.husvankempen.de/nunn/rating.htm
Blitz: http://www.husvankempen.de/nunn/blitz.htm
40/120: http://www.husvankempen.de/nunn/rating120.htm
Tester: http://www.husvankempen.de/nunn/testers/testers.htm
Games of the week: http://www.husvankempen.de/nunn/40_40%2 ... on/gow.JPG
Elo-comparison: http://www.husvankempen.de/nunn/Replay/ ... arison.htm

Werner
CEGT-Team

[Leto]

For the 40/20 list I recommend adding more games between Stockfish 1.7.1 x64 4CPU vs Rybka 3 x64 2CPU. Here Stockfish performed 70% but only in 17 games earning a 3297 elo performance, and I think this might be artificially increasing Stockfish's rating in the list. With more games I expect the ratio in this matchup to fall closer to 50%.

Stockfish 1.7.1 x64 4CPU's score of 48% against Naum 4.2 x64 4CPU, and the issue I just mentioned above tells me that Stockfish might not really be the second strongest engine.

[Uri Blass]

For the 40/20 list I recommend adding more games between Stockfish 1.7.1 x64 4CPU vs Rybka 3 x64 2CPU. Here Stockfish performed 70% but only in 17 games earning a 3297 elo performance, and I think this might be artificially increasing Stockfish's rating in the list. With more games I expect the ratio in this matchup to fall closer to 50%.

Stockfish 1.7.1 x64 4CPU's score of 48% against Naum 4.2 x64 4CPU, and the issue I just mentioned above tells me that Stockfish might not really be the second strongest engine.

The result against Rybka3 2 cpu has little influence and I am sure that stockfish is second place even if you do not include the games against Rybka3 2 cpu

I also have no reason to think that more games against Rybka3 2 cpu are going to reduce Stockfish's rating.

I think that stockfish suffered from playing too many games against Naum and it is not that stockfish enjoyed from small number of games against Rybka3 2 cpu.

Here are Naum4.2 results against the top engines.
Naum4.2 4 cpu-Rybka 3 x64 4CPU + 31 = 73 - 30 50.4 % 3184
Naum4.2 4 cpu-Stockfish 1.7.1 x64 4CPU - 3156 88 + 21 = 48 - 19 51.1 % 3163
Naum4.2 4 cpu-Rybka 3 x64 2CPU - 3145 183 + 37 = 86 - 60 43.7 % 3102

Uri

[Werner]

Hi Leto,
of course we will add more games against other opponents. The match against R3 2CPU will be continued. If Naum 4.2 gets a lot of games with open positions it can beat every engine. So only a lot of games can show which one is nr. 2.
But I think the score of Stockfish 4CPU will rise with more games: compare the results of our Blitz list with v. 1.6 - its +50 points or more!

Werner