WBEC Predictions with bayeselo

by Rémi Coulom.

This page is a step-by-step guide to use bayeselo to produce predictions in the WBEC Ridderkerk computer-chess tournament.

Step-by-Step Instructions

These instructions generate predictions for the second division. It should be easy to generalize them to other divisions.

• Download all the pgn files of the WBEC history into a directory, and uncompress them. Also download the pgn file(s) of the division where predictions have to be made. At the time of writing of this document, the file list is:

  -rw-r--r--  1 coulom Aucun 16890041 Jan 14 12:34 Edition11.pgn
  -rw-r--r--  1 coulom Aucun    52499 Feb  7 15:34 WBEC12_2ndDiv.pgn
  -rw-r--r--  1 coulom Aucun   391116 Feb  5 15:07 WBEC12_2ndDiv_Tot.pgn
  -rw-r--r--  1 coulom Aucun   607874 Aug 25 18:08 WBEC12_NE_Tot.pgn
  -rw-r--r--  1 coulom Aucun 56397822 Jan 14 12:29 WBEC1to10.pgn
  -rw-r--r--  1 coulom Aucun  2611731 Jan 14 12:38 WBEC_NE_Tot.pgn

• Two versions of Zarkov participate in this division (Zarkov 4.85, and Zarkov 4.86). bayeselo does not handle this automatically. Rename "Zarkov 4.85" to "Zarkov 4.86" in WBEC12_2ndDiv_Tot.pgn so that the two versions are considered identical.
• Copy bayeselo.exe, and prediction.cui to the directory with all the pgn files. You may have to edit prediction.cui if you use other pgn files than those listed above. You may also edit it to change the number of simulations.
• Open a command-line console and cd to the directory, and run bayeselo.exe <prediction.cui there. It should produce an output that looks like:

  version 0054, Copyright (C) 1997-2005 Remi Coulom.
  compiled Feb  9 2006 14:28:44.
  This program comes with ABSOLUTELY NO WARRANTY.
  This is free software, and you are welcome to redistribute it
  under the terms and conditions of the GNU General Public License.
  See http://www.gnu.org/copyleft/gpl.html for details.
  ResultSet>ResultSet>ResultSet>ResultSet>off
  21 game(s) loaded, 0 game(s) with unknown result ignored.
  165 game(s) loaded, 0 game(s) with unknown result ignored.
  454 game(s) loaded, 0 game(s) with unknown result ignored.
  44835 game(s) loaded, 1 game(s) with unknown result ignored.
  47144 game(s) loaded, 0 game(s) with unknown result ignored.
  Iteration 100: 0.0061693 
  Iteration 200: 0.00157669 
  Iteration 300: 0.000429331 
  Iteration 400: 0.000121862 
  Iteration 500: 3.49192e-05 
  Iteration 600: 1.00304e-05 
  00:00:01,04
  144 game(s) loaded, 0 game(s) with unknown result ignored.
  165 game(s) loaded, 0 game(s) with unknown result ignored.
  2
  10000

• One this is done, two files should have been produced in the current directory: players.txt that contains the list of players (with their ratings and standard deviations), and predictions.txt that contains this table of predictions:

  Rank Player name           Points EPoints  StdDev   ERank    1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
     1 Amyan 1.59.7            13.5   70.36    6.84    1.36   78 16  3  1  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0
     2 Zarkov 4.86             11.0   61.53    7.19    2.98   18 48 14  6  4  3  2  1  1  1  1  1  0  0  0  0  0  0  0  0  0  0  0  0
     3 Nejmet 3.07             10.5   53.47    3.95    5.34    0  6 19 21 16 11  8  6  4  3  2  1  1  1  1  0  0  0  0  0  0  0  0  0
     4 Frenzee 200             11.5   48.63    4.60    9.41    0  1  5  7  8  9  9  8  8  7  7  6  5  5  4  3  3  2  1  1  0  0  0  0
     5 Abrok 5.0                7.5   48.62    4.04    9.30    0  0  4  6  8  9 10 10 10  8  7  6  5  4  4  3  2  1  1  0  0  0  0  0
     6 Queen 3.09               7.0   48.58    8.17   10.17    1  7 10  9  7  6  5  4  4  4  4  3  4  3  4  4  3  3  3  3  2  3  2  1
     7 PostModernist 1010a      9.0   48.50    4.14    9.41    0  1  3  6  8  9 10 10  8  8  7  6  5  5  4  3  2  2  1  1  0  0  0  0
     8 Matacz 1.02              7.0   48.33    7.84   10.26    1  6 10  8  7  6  6  5  4  4  4  4  3  3  4  4  4  3  3  3  3  2  2  1
     9 Snitch 1.24              8.5   48.30    7.88   10.31    1  6 10  8  7  6  5  5  5  4  4  4  4  4  4  3  4  4  4  3  2  2  2  1
    10 Dragon 4.7.5             7.0   47.23    4.62   10.75    0  0  2  5  6  7  7  8  8  7  7  7  6  6  5  5  4  3  2  1  1  1  0  0
    11 CEng 2.53.6b             7.5   47.13    7.96   11.24    1  4  9  7  7  5  5  5  5  4  4  4  4  4  4  4  4  4  4  3  3  3  3  2
    12 Arasan 8.4               9.5   45.72    4.60   12.18    0  0  1  3  4  5  6  6  7  7  7  8  7  7  7  6  6  5  4  3  2  1  0  0
    13 Phalanx XXII             6.0   45.40    4.12   12.51    0  0  1  2  3  4  5  6  7  7  7  8  9  8  7  7  6  5  4  2  2  1  0  0
    14 KnightDreamer 3.3        7.5   45.13    4.18   12.81    0  0  1  2  2  4  5  5  6  7  8  8  8  8  8  7  6  5  4  3  2  1  0  0
    15 Leila 0.53h              7.0   44.81    4.08   13.14    0  0  0  1  2  3  4  5  6  7  7  8  8  9  8  8  7  6  5  3  2  1  0  0
    16 Gosu 0.11                9.0   44.65    8.13   13.18    0  3  5  4  5  4  4  4  4  4  4  4  4  4  4  4  5  5  5  5  5  5  5  3
    17 Amateur 2.86             5.0   44.37    4.84   13.55    0  0  1  2  3  3  4  4  5  6  6  6  7  7  7  8  7  7  6  4  3  2  1  0
    18 TheCrazyBishop 0052     10.0   43.98    4.18   13.99    0  0  0  1  2  2  3  4  5  6  6  7  8  8  8  8  8  8  6  4  3  2  1  0
    19 Bruja 1.9                4.5   38.51    4.84   18.73    0  0  0  0  0  0  1  1  1  1  2  2  3  3  4  5  7  9 11 12 12 12  9  5
    20 Muse 0.899b uci          3.5   37.66    4.28   19.50    0  0  0  0  0  0  0  0  0  1  1  1  2  3  3  5  6  9 11 14 15 14 11  5
    21 Capture R01              4.5   37.60    6.14   18.99    0  0  0  0  1  1  1  1  1  2  2  2  2  3  3  5  5  7  8 10 10 12 13 11
    22 TRACE 1.35               7.0   37.11    7.79   18.76    0  0  1  1  1  1  1  1  2  2  2  2  2  2  3  4  4  5  6  7  8 11 15 17
    23 Chezzz 1.0.3             5.0   35.69    4.00   20.87    0  0  0  0  0  0  0  0  0  0  0  0  0  1  2  2  4  5  8 12 16 19 18 11
    24 FrancescaMAD 0.10        5.5   32.71    7.34   21.26    0  0  0  0  0  0  0  1  1  1  1  1  1  1  2  2  3  3  4  5  7 10 17 40

Caveats

bayeselo considers that different versions of the same program are different players. It might be possible to improve prediction accuracy by assuming that successive versions are likely to be close in strength, or that new versions are stronger than previous versions.