I tested it against v2.4 and it scored +70 elo after 200 games in 30"+0.5", but when tested in a rating list against a varied population of engine (v2.4 and v2.5 under samt conditions), it didn't add much elo (approx +15-20 elo)
Morality
1/ i should play more games in testing
2/ self testing sucks!
Anyway, here's the list with the addidtion of DC 2.5, and I also removed all the games of 2.4 to not create any rating distortion (cf explanation on CCRL page)
Code: Select all
Rank Name Elo + - games score oppo. draws
1 Critter 1.4 3242 33 32 350 74% 3044 34%
2 IvanHoe 999946h 3185 31 31 350 64% 3061 37%
3 Stockfish 2.2.1 3166 31 30 400 66% 3015 32%
4 Protector 1.4 2913 34 34 350 46% 2949 22%
5 Umko 1.2 2859 28 28 500 50% 2870 24%
6 Toga 1.4.1 2836 26 26 600 55% 2813 22%
7 Daydreamer 1.75 2734 27 27 450 60% 2660 29%
8 Fruit 2.1 2700 25 24 550 52% 2683 27%
9 Crafty 23.4 2653 27 27 500 38% 2760 24%
10 Cheng3 1.07 2643 66 62 80 70% 2508 30%
11 GNU Chess 5.07.173b 2642 26 26 500 47% 2666 27%
12 Arasan 13.4 2633 29 29 400 44% 2675 23%
13 Rodent 0.10 2624 29 28 400 55% 2589 25%
14 Pepito 1.59 2588 26 26 500 53% 2561 21%
15 Sloppy 0.2.2 2538 23 23 680 44% 2579 23%
16 Greko 9.0 2491 23 23 700 44% 2542 21%
17 Pawny 0.3.1 2469 26 26 500 50% 2469 19%
18 DoubleCheck 2.5 2421 30 30 400 48% 2436 15%
19 Olithink 5.3.0 2376 33 33 300 46% 2401 20%
20 EXchess 6.0.2 2353 31 32 350 41% 2424 19%
21 Sungorus 1.4 2343 29 30 400 37% 2445 22%
22 Jazz 501 2328 33 34 300 40% 2409 20%