The Week in Chess

Saturday, November 30, 2013

Saros 4.1.1 x64 - Gauntlet Matches, 100 Rounds

Saros 4.1.1 x64 is a UCI chess engine by +Roberto Munter  released last November 19, 2013. It is a clone of the Ippolit family of chess engines, particularly Ivanhoe.

Saros 4.1.1  scored 39.93% with 328 wins, 751 losses and 1021 draws against the 21 strongest computer chess engines and got an ELO rating of 3016. It is the weakest among the Ippolit clones in this gauntlet matches but has a consolation of improving its ELO by 33 points from the last version tested and the distinction of belonging to the elite ELO 3000 club. The field is so strong that Saros placed last in the gauntlet rankings.

Here is the gauntlet performance of Saros 4.1.1:

Rank Engine ELO Raw Games Score% Points Win Loss Draw Chg TF% Ply
1 Stockfish 131109 x64 3211 126 100 76.50 76.5 63 10 27 0 3.00 85
2 Houdini 3 Pro x64 3177 95 100 73.50 73.5 56 9 35 0 3.00 75
3 Robodini 1.1 x64 3100 69 100 68.00 68.0 56 20 24 1 4.00 79
4 Komodo 6 x64 3137 53 100 68.00 68.0 48 12 40 0 3.00 79
5 Samsung x64 3119 45 100 66.00 66.0 49 17 34 0 0.00 0
6 Gull R600 x64 3112 36 100 65.50 65.5 45 14 41 0 0.00 0
7 Critter 1.6a x64 3098 34 100 65.00 65.0 43 13 44 0 3.00 113
8 Strelka 5.7 x64 3100 11 100 61.50 61.5 41 18 41 0 5.00 67
9 Bouquet 1.8 x64 3076 -9 100 59.00 59.0 31 13 56 0 0.00 0
10 Mars 1 x64 3067 -15 100 58.50 58.5 29 12 59 0 2.00 81
11 PanChess 00.537 x64 3058 -17 100 58.00 58.0 28 12 60 0 0.00 0
12 RyanFish 1 x64 3074 -23 100 56.50 56.5 31 18 51 -1 0.00 0
13 Firenzina 2.3.1 xTreme x64 3062 -24 100 56.50 56.5 29 16 55 0 0.00 0
14 ComStock 3 x64 3075 -24 100 56.50 56.5 30 17 53 0 0.00 0
15 Rybka 4.1 x64 3029 -25 100 56.00 56.0 32 20 48 0 2.00 106
16 Fire 2.2 xTreme x64 3059 -38 100 54.00 54.0 27 19 54 0 0.00 0
17 RobboLito 0.21Q x64 3055 -41 100 53.50 53.5 22 15 63 0 4.00 22
18 Igorrit 0.086v9 x64 3056 -43 100 53.00 53.0 24 18 58 -1 1.00 119
19 LEOpard 0.7c x64 3034 -44 100 53.50 53.5 21 14 65 0 0.00 0
20 Tactico Power 2011 x64 3036 -50 100 52.00 52.0 25 21 54 0 3.00 60
21 Ivanhoe 46h x64 3042 -56 100 50.50 50.5 21 20 59 0 5.00 22
22 Saros 4.1.1 x64 3016 -61 2100 39.93 838.5 328 751 1021 3016 1.81 79
.
Download the computer chess engines tournament games here.

Owl Computer Chess Engines Rating List #77

The 77th Owl Computer Chess Engines Rating List #77 released, 11/30/2013.

View the full rating list here.

Owl Computer Chess Engines Rating List #76

The 76th Owl Computer Chess Engines Rating List #76 release, 11/29/2013.

View the full rating list here.

Friday, November 29, 2013

Junior 13.08.4 x64 inflated ELO rating 3200

Right after I posted the last Junior 13.08.4 gauntlet matches yesterday, November 11, 2013, I noticed that the ELO rating of Junior in its website is 3200+.

I was intrigued of such very high rating so I looked at the reliable and trusted rating list sites and found that it was really overstated. The ELO rating of Junior at CCRL is 3026 while at CEGT it is 2841. The OWL rating list site has it at 2807.  The bloated ELO rating of Junior is almost equal to the rating of number one Stockfish 131109 x64 at 3211.  To check whether I am correct in my rating list statistics, I decided to have a one-on-one match between Junior 13.08.4 and Stockfish 131109.

And here is the result:

Junior 13.08.4 x64 vs. Stockfish 131109 x64 - Match 100R 1M1S
RankEngineScoreStJuS-B
1Stockfish 131109 x6488.5/100· ·· ·· ·83-6-11 1017.75 
2Junior 13.8.04 x64 11.5/1006-83-11· ·· ·· · 1017.75 


100 games played / Tournament finished

Tournament start: 2013.11.29, 01:28:19
Latest update: 2013.11.29, 08:08:40
Level: Blitz 1/1
Hardware: AMD Phenom(tm) II X4 945 Processor with 2.0 GB Memory
Operating system: Windows 7 Ultimate Professional Service Pack 1 (Build 7601) 64 bit
Table created with: Arena 3.0


ELO Rating estimate with BayesElo:
 Rank Name                               Elo    +    - games score oppo. draws
   1      Stockfish 131109 x64    166   80   80   100   89%  -166   11%
   2      Junior 13.8.04 x64        -166   80   80   100   12%   166   11%

Junior's score and the equivalent BayesElo rating shows that indeed it's published ELO rating estimate is very high.  There is no relative data shown to support it in its website, so it can be assumed that Junior's 3200+ ELO rating is for commercial purposes.

Download the match games here.

Thursday, November 28, 2013

Junior 13.08.4 x64 - Gauntlet Matches, 100 Rounds

Junior 13.08.4 x64 (Yokohama) is a UCI chess engine by Amir Ban and Shay Bushinsky released last November 1, 2013.

Junior scored 43.2% with 280 wins, 416 losses and 304 draws against the 10 strong chess engines selection.  It garnered 2807 ELO ratings points and #40 in the Top 100 Strongest Computer Chess Engines list.
.
Rank Engine ELO Raw Games Score% Points Win Loss Draw Chg TF% Ply
1 Tinapa 1.01 x64 2903 75 100 67.50 67.5 52 17 31 -7 3.00 88
2 Crab PGO x64 2911 61 100 65.00 65.0 51 21 28 -11 5.00 148
3 Shredder 12 x64 2916 61 100 64.50 64.5 53 24 23 -4 3.00 160
4 OpenCritter 1.1.37 x64 2969 44 100 63.00 63.0 48 22 30 -25 3.00 40
5 Hiarcs 14 2848 37 100 61.50 61.5 49 26 25 1 2.00 113
6 Protector 1.5.0 x64 2905 14 100 58.50 58.5 40 23 37 -20 7.00 120
7 Thinker 54D Inert x64 2772 -39 100 50.50 50.5 33 32 35 -4 2.00 82
8 Junior 13.08.4 x64 2807 -42 1000 43.20 432.0 280 416 304 2807 2.70 78
9 Onno 1.2.70 x64 2810 -45 100 49.50 49.5 34 35 31 -14 2.00 160
10 Doch 1.3.4 x64 2785 -67 100 46.00 46.0 27 35 38 -14 0.00 0
11 Quazar 0.4 x64 2767 -98 100 42.00 42.0 29 45 26 -17 3.00 63
.
Download computer chess engines tournament games here.

Note:  TF% = Time forfeit percentage
          Ply = Average moves of games lost by time forfeit

Wednesday, November 27, 2013

Samsung x64 - Gauntlet Matches, 100 Rounds

Samsung x64 is a UCI chess engine by +Felix Gulden  released last October 24, 2013. Samsung x64 is probably a clone of Gull as  indicated by close similarity of its file size, sparse UCI parameters, search depth, nodes per second speed and others.  Felix Gulden is the one who released the latest development version of Gull R600 without proper permission from the original author,  Vadim Demichev.

Samsung x64 scored 55.1% with 704 wins, 500 losses and 796 draws against the top 20 strongest chess engines. It earned 3123 ELO rating points and the 4th place in the Top 100 Strongest Chess Engines list, 1 rank higher than Gull R600.

Here is the gauntlet performance of Samsung x64:

Rank Engine ELO Raw Games Score% Points Win Loss Draw Chg TF% Ply
1 Stockfish 131109 x64 3216 138 100 66.50 66.5 49 16 35 2 0.00 0
2 Houdini 3 Pro x64 3181 106 100 62.00 62.0 41 17 42 0 0.00 0
3 Komodo 6 x64 3141 100 100 60.00 60.0 46 26 28 3 3.00 99
4 Robodini 1.1 x64 3103 69 100 56.00 56.0 35 23 42 1 3.00 78
5 Samsung x64 3123 31 2000 55.10 1102.0 704 500 796 3123 1.65 60
6 Critter 1.6a x64 3102 27 100 49.50 49.5 23 24 53 0 0.00 0
7 Gull R600 x64 3116 24 100 49.00 49.0 31 33 36 0 1.00 117
8 RyanFish 1 x64 3079 12 100 47.00 47.0 32 38 30 0 1.00 27
9 Bouquet 1.8 x64 3081 4 100 45.50 45.5 23 32 45 1 2.00 131
10 ComStock 3 x64 3080 2 100 45.50 45.5 26 35 39 0 0.00 0
11 Strelka 5.7 x64 3104 2 100 45.00 45.0 22 32 46 -1 3.00 65
12 Firenzina 2.3.1 xTreme x64 3066 -8 100 43.50 43.5 21 34 45 0 0.00 0
13 RobboLito 0.21Q x64 3060 -23 100 42.00 42.0 24 40 36 0 2.00 17
14 Fire 2.2 xTreme x64 3064 -24 100 41.50 41.5 20 37 43 0 0.00 0
15 Mars 1 x64 3072 -31 100 39.50 39.5 15 36 49 0 1.00 211
16 PanChess 00.537 x64 3062 -47 100 37.50 37.5 17 42 41 -1 0.00 0
17 Rybka 4.1 x64 3033 -52 100 37.00 37.0 16 42 42 0 0.00 0
18 Igorrit 0.086v9 x64 3061 -53 100 37.00 37.0 18 44 38 -1 1.00 61
19 Tactico Power 2011 x64 3040 -79 100 32.50 32.5 12 47 41 -1 2.00 99
20 Ivanhoe 46h x64 3047 -87 100 32.00 32.0 14 50 36 -1 2.00 181
21 LEOpard 0.7c x64 3038 -111 100 29.50 29.5 15 56 29 -1 0.00 0
.
Download the computer chess engines tournament games here.

Note:  TF% = Time forfeit percentage
            Ply = Average moves of games lost by time forfeit

Chessdom News