Moby Dam's games at the 2015 Olympiad

[Rein Halbersma]

Moby Dam is a very nice program (clean source base, compiles very fast, runs fast). Its game against Scan is analyzed in this other thread. Below some other interesting games.


Slagzet.com - Moby Dam after 32... 12-18

Several suboptimal moves happened here according to Kingsrow. Here Slagzet.com ignored the exchange 33. 34-29 to clean up its right wing. Instead it attacked with 33. 30-24? 11-17! 34. 47-42 8-12? 35. 33-28 23-29 36. 24x33 14-20 37. 25x14 13-19 38. 14x23 18x40



with a rather unclear position. White is a piece up, and black cannot directly get a king (40-45? 49-44 and the threat 44-40 and 33-29). There are a lot of delaying moves (horizon effect!) that prevent programs without 8 piece databases from resolving the position accurately. Kingsrow takes a minute or so to find a database draw in this position. Moby Dam managed to convert this to a nice win.

Instead of 34... 8-12?, black could have played 34... 6-11! (prevents 35. 33-28? through 14-20! and 4-9 and 9x40!) and now 35. 49-43 14-19 36. 34-30 4-9 37. 33-29 23x34 38. 30x39 19x30 39. 25x34 17-22 with a deadly lock around white's left wing.

[Rein Halbersma]

Moby Dam - Sjende Blyn after 23... 15x24

Another remarkable game (similar to JDraughts - Scan): white let his opponent make a king in exchange for three men after 24. 39-34!? 18-23 25. 37-32 24-29 26. 33x24 19x50 27. 28x6



The black king would quickly be exchanged for a drawn position. I would be curious to how far both programs searched before allowing / taking the combination to king. Did they see the that the king could be forced to be taken back?

UPDATE: Sjende Blyn's author Jelle Wiersma remarks on Toernooibase that a recent change to his program made it judge the position with the king somewhat too rosy.