Problem #364 should be adjusted

#1

Problem #364 forces Red to make a much more difficult pattern of moves.

At the beginning, 11-16 is a much better move, because White either loses a man immediately or falls for a shot that wins easily for Red. However, the problem claims that the computer likes another move better, which is not as good in my opinion.

#2

Hmm, that’s a pity but you’re right. There’s a couple of problems here.

  1. I have not written the tools required to easily adjust problems after they are created. I can shorten them or disable them, but that’s about it so far.

  2. If I do adjust it by hand does the computer’s best line become an alternate? That seems weird. I don’t think the engine is wrong, just inhuman. I’m not claiming my generator is bug free but the engine line does seem pretty crushing.

  3. I don’t think I want to try to support multiple winning lines in a single problem. The generator is complicated enough already and I feel the point of a tactics puzzle is to find the best move, not just some winning line. I’d love to make endgame puzzles that have multiple paths but it seems weird for a tactics puzzle.

So I can disable it or keep it as is. I guess I’ll disable it for now. More puzzles are coming, lets just keep the good ones.

#3

One other thought that makes me think about bringing the problem back.

In the game played, Labadie played 11. 11-16. But he went on to draw that endgame, so maybe it’s not as easy as it looks. Is the computer line better after all?

Thoughts?

#4

I just looked it up on my program and it appears that the win isn’t as easy after the shot as I thought. However, at the starting position, my program does rate 11-16 has higher (10 piece database).

After about 10 seconds, it has 11-16 at about +700 and the problem suggested 5-9 at +300, so both are clear wins once the program is able to see into the endgame database.

I’m not sure the best way to fix it. From a problem perspective, you would either want to exclude this problem all together or allow both wins to be tried by the player attempting to solve the problem.

#5

Thanks for the help. I’ll leave it disabled.