Nemesis wins 32-5 with 251 draws

If you haven't seen our game viewer before, click here for help

Here are all the decisive games:

Game 3
Game 4
Game 13
Game 41
Game 43
Game 47
Game 65
Game 73
Game 93
Game 95
Game 103
Game 104 - Cake punishes a weak pp move
Game 112 - Nice cook from Cake
Game 121
Game 122
Game 132
Game 135
Game 137 - Nemesis uses the Octopus' tentacles to defeat Cake
Game 139 - Nemesis wins Mantell's Kelso
Game 147
Game 185
Game 195
Game 198
Game 203
Game 204 - Cake wins an admirable game, correcting a "who's who" list of Checker players.
Game 214
Game 217
Game 221
Game 229
Game 233
Game 250
Game 252 - 252, 254, 258, 260 and 284 are all the same winning ending, but are not exactly duplicates. Nemesis still has to transpose into the wins.
Game 254
Game 258
Game 260
Game 279
Game 284

If you are an experienced checker player, you will see that many of these games, Cake played a losing move, when there was a published draw. Nemesis played very classically and rarely ran into trouble. Cake threw many unpublished moves at Nemesis who was rarely phased. Cake's opening book is its Archilles heel but I know Martin is working hard on improving it.

• Cake++ is designed by Martin Fierz
• Cake++ is a very strong checkers engine that runs within the Checkerboard program, also one of Martin's creations. Both are Freeware, and can be downloaded at Martin's web site

• This is the published results of our own benchmarking labs. We made every effort to be fair and just in the benchmarking procedure, so that the results carry suitable credibility. The author of Cake++, Martin Fierz has agreed that the benchmark was conducted in a fair manner.
• Games were played December 2001 - February 2002
• 288 games were played, each program played each side of the 144 ACF 3-move ballots
• Average 60 seconds per move was allowed.
• The match was played on a PIII/600Mhz computer with 1Gb memory. When either program was thinking, it has exclusive access to the computer's resources (disk and memory.)
• Nemesis was not allowed to think on its opponents time, even though it has that capability
• Cake 1.4 had 8Mb Hash, Nemesis had 16Mb Hash.
• Nemesis used its factory settings for opening and endgame databases, i.e. Static and Dynamic databases were on, and the 8 piece database was used. Cake had the 6 piece database available.
• Cake had its opening database turned on, as per the factory settings.
• Nemesis was not adjusted during the course of the match to correct any bad or losing moves earlier in the match.
• This benchmark was probably a fair comparison of the programs' engine and database capabilities. In a match situation, Nemesis would benefit from thinking on its opponents time, and so would perform better. But to counterbalance, Cake ran with half the Hash memory here (the operator was not aware that Cake's hash memory could be increased beyond it's default 8Mb).