Tough choice

[EdTrice]

White to move

The question is: Do you play 23-19 and hope your opponent can't see a saving 2x2 pitch leading to a draw, or do you play 32-28, which stops the 2x2, but could lead to a draw that might be longer, but easier?

The actual game went ...23-19, 1-6 30-25!, 12-16* 19x12, 15-18* 22x15, 10x19x28 25-22 {with some more moves}

[Alex_Moiseyev]

Do you play 23-19 ?

Ed, this position comes from 9-13 22-18 11-15 opening, 24-20 attack and is well known to published play. In this position I would play 30-25, not 23-19 ... and then hope for 1-6? 32-28 and red can resign.

I won the game against grandmaster Paul Davis in 1998 VA Open by using this trick.

Alex

[EdTrice]

Ed, this position comes from 9-13 22-18 11-15 opening, 24-20 attack and is well known to published play. In this position I would play 30-25, not 23-19 ... and then hope for 1-6? 32-28 and red can resign.

I won the game against grandmaster Paul Davis in 1998 VA Open by using this trick.

Alex

Hi Alex,

This positions outlines the "problem" that I am trying to circumvent with the Aggressive Draw Heurisitc.

23-19 leads to the best score for white, because the draw is "hardest" if you count the drawn nodes and compare them with the total nodes.

32-28 comes in as the second best move with a draw that has more chances for red to survive.

30-25 looks like an "easy draw" when you enumerate the draws and compare them with the total ratio of losing paths along the way. While there are pitfalls, the program can "see" them and avoid them, but a human might not.

The underlying assumption of the A.D.H. is that all draws are "known", so the path that must be most treacherous is the narrowest one through the mix.

Perhaps I need to recompute the scores factoring into something other than the pure ratios.

Thanks for your input.