Spliffjiffer wrote:Ka6 draws, but almost no engine can find it...
On my system Critter finds it in well under a minute. But I gave up waiting for Crafty, Firenzina, Komodo CCT, Shredder, various versions/derivatives of Stockfish...
Sting SF 3 solves it in 4 minutes on my dual core and beleive that all the others clones too.
In this game Ditrty seemed to exhibit a search bug. It was announcing mate in 10 while Spartacus was saying mated in 11. As the game progressed the engines kept disagreeing 1 ply. After 32... Kg7 Dirty's DTM suddenly jumped up. It obviously had not included that move in its tree as a valid defence. It recalculated a mate, and started to execute it, but again on the assumption that Kf8-g7 would not be a valid move. At that time Spartacus had lost the mate from view altogether (its score rising back to -3) because the only mates within its horizon would have to go over an already reached position, which it scored as 0. At some point Spartacus' score even fell back to 0 because actual repeats occured.
After Dirty was surprised by the Kg7 defence for the second time, it also lost the imagined too-fast mate from view, because it would involve repeats. Hence it advanced a Pawn to create a clean slate. There it finally did find a mate.
Spliffjiffer wrote:Ka6 draws, but almost no engine can find it...
On my system Critter finds it in well under a minute. But I gave up waiting for Crafty, Firenzina, Komodo CCT, Shredder, various versions/derivatives of Stockfish...
Sting SF 3 solves it in 4 minutes on my dual core and beleive that all the others clones too. :lol:
Your patience was greater than mine.
On my system after a re-run:
Critter 45s
Sting 403s
Meanwhile stockfish_13060116_x64 is still looking after 1100+s (depth 39).
Hi,Interesting,but more interesting is that Dirty saw M2 at move 32 if black would have played
32 Kf8 and normally it wouldn't have a problem mating in 2 after kg7.
Maybe the laptop overheated(turboboost) or it's bug but I can't reproduce it.
Best
Fonzy
Hi,Interesting,but more interesting is that Dirty saw M2 at move 32 if black would have played
32 Kf8 and normally it wouldn't have a problem mating in 2 after kg7.
Maybe the laptop overheated(turboboost) or it's bug but I can't reproduce it.
See also hgm's post above.
Best
Fonzy
Spliffjiffer wrote:Ka6 draws, but almost no engine can find it...
On my system Critter finds it in well under a minute. But I gave up waiting for Crafty, Firenzina, Komodo CCT, Shredder, various versions/derivatives of Stockfish...
Sting SF 3 solves it in 4 minutes on my dual core and beleive that all the others clones too.
Your patience was greater than mine.
On my system after a re-run:
Critter 45s
Sting 403s
Meanwhile stockfish_13060116_x64 is still looking after 1100+s (depth 39).
Marek, I think in the near future the 1100+s is not any diference.
Monarch is so weak, with hardly any clever extensions, it finds mate in 3 instantly. It only needs 719 nodes. I'm not sure what the problem is with other engines.
[D]1r3k1r/1bp5/1p1bR1Q1/3p4/3P4/2P5/P4PPP/6K1 w - -
Oops, my mistake. I was analyzing the first occurrence. On the second occurrence Monarch makes the same mistake as everyone.
Monarch returns a draw when the position is encountered for the first time. I suppose an improvement would be return a draw on the first occurrence if both positions are part of the search and on the third occurrence if part of the game history.