persistent stalemate threat

[Vinvin]

Here's one played 2 weeks ago


and 4 exercises (black to play) : http://www.olibris.fr/produits/22/fiche ... _folle.pdf

[lucasart]

and 4 exercises (black to play) : http://www.olibris.fr/produits/22/fiche ... _folle.pdf

Ces positions sont assez evidentes. En tous cas, j'aime beaucoup les illustrations. Ce qui serait plus interessant est un fichier epd avec des positions de ce genre, ou une menace de pat persistente existe. Parfois la situation est bien plus complique, et fait intervenir des pieces clouees par exemple.

[Vinvin]

Désolé, je n'en ai pas mais voici un exemple de "fou fou" (="crazy bishop")

http://en.wikipedia.org/wiki/Desperado_ ... elikiewicz
Black resigned (???)

Other examples from this page are easy for computers.

and 4 exercises (black to play) : http://www.olibris.fr/produits/22/fiche ... _folle.pdf

Ces positions sont assez evidentes. En tous cas, j'aime beaucoup les illustrations. Ce qui serait plus interessant est un fichier epd avec des positions de ce genre, ou une menace de pat persistente existe. Parfois la situation est bien plus complique, et fait intervenir des pieces clouees par exemple.

[JVMerlino]

[d]8/pR5p/4k3/2P1B3/5P1r/4P3/P2r4/2R3K1 w - - 0 34

This may be helpful for you, since it is on the same theme. This is a position I use regularly for making sure I didn't break my draw detection.

Rb2 or Rg7 are the two best moves, and trying to win material by Rxa7 leads to a draw involving lots of checks -- meaning, not a quick 3-fold draw. So if your draw detection is broken, your engine might play it, since it is the only move that wins material (although gives up the win).

All other moves are inferior to the first two, although a few still win, but most of them also allow Black to draw.

Hope it helps,
jm

[Uri Blass]

No need to detect draw by repetition to find Rg7 and not Rxa7

Rg7 simply wins more material than Rxa7 if you search deep enough
because black is losing the rook for the c pawn that want to be a queen.

The real test for chess engines is if they can see the draw after Rxa7
Most engines can see it and it is clearly different than the position in the beginning of the thread when engines cannot see it.

[lucasart]

[d]8/pR5p/4k3/2P1B3/5P1r/4P3/P2r4/2R3K1 w - - 0 34

This may be helpful for you, since it is on the same theme. This is a position I use regularly for making sure I didn't break my draw detection.

Rb2 or Rg7 are the two best moves, and trying to win material by Rxa7 leads to a draw involving lots of checks -- meaning, not a quick 3-fold draw. So if your draw detection is broken, your engine might play it, since it is the only move that wins material (although gives up the win).

All other moves are inferior to the first two, although a few still win, but most of them also allow Black to draw.

Hope it helps,
jm

Oddly enough, there's nothing smart in my engine to detect stalemate threats, unless when a stalemate occurs in a search node (search, not qsearch), yet it doesn't play Rxa7 here. DoubleCheck prefers Rxa7 only until depth 5. From depth 6 onwards it chooses different moves: a3 (depth 6) a4 (depth 7), and finally at depth 8 it plays Rg7 which is arguably the best move.

[Vinvin]

Bad link, it's : http://en.wikipedia.org/wiki/Desperado_ ... s_Palatnik

Désolé, je n'en ai pas mais voici un exemple de "fou fou" (="crazy bishop")

http://en.wikipedia.org/wiki/Desperado_ ... elikiewicz
Black resigned (???)

Other examples from this page are easy for computers.

and 4 exercises (black to play) : http://www.olibris.fr/produits/22/fiche ... _folle.pdf

Ces positions sont assez evidentes. En tous cas, j'aime beaucoup les illustrations. Ce qui serait plus interessant est un fichier epd avec des positions de ce genre, ou une menace de pat persistente existe. Parfois la situation est bien plus complique, et fait intervenir des pieces clouees par exemple.