Thanks to Bernhard Bauer for sharing this position.
Thanks to you and Bernhard for bringing up this position.
Actually it's quite easy for Humans. With feeding the first 2 plies programs find it it fast too.
The reason I replied here is that Andscacs 0.93 crashes here in analysis mode before I can make the last move for mating!
This is the no-popcount version. As I have never tried mate riddles before with Andscacs it might be not related to this position,
but to making the last mating move. Can anyone confirm this behaviour?
(could be a design decision)
tpoppins wrote:
Back to the present. Has Stockfish emerged with 24.Qxe5 yet? Clean hash, no source mods, no MultiPV? I bet it's still 24.Qa3 0.00, three years and +200 Elo later. How many more years are you prepared to wait, Louis?
Hang on, let me check my SPARC workstation. It has been searching Gusev-Auerbach since 2002. Of course, Glaurung, not Stockfish.
Humans are incredible creatures. Once in a while, they find moves that engines cannot. Yet.
Just to put things in perspective, you and Dann will always be among my favorite posters, regardless of any difference of opinion on SF or any other engine.
Just to put things in perspective, you and Dann will always be among my favorite posters, regardless of any difference of opinion on SF or any other engine.
Just to put things in perspective, you and Dann will always be among my favorite posters, regardless of any difference of opinion on SF or any other engine.
Thanks, very kind of you to say.
I consider both of you as very interesting and valuable contributors.
I do not think that my opinion is better than that of anyone else.
I do not think SF to be infallible on quiet positions. But I do think that on average it is better than any of the alternatives that are available to me.
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.
tpoppins wrote:
In the past two years there have been numerous posts about an alarming number of positions SF evaluates as 0.00. Someone even coined the term "Drawfish". I personally have seen hundreds of such positions on Let's Check, ranging from quiet to turbo-charged. There is a current thread discussing another such case and SF's "tunnel vision".
About "tunnel vision", I think I largely agree with Dann. Given enough time/depth, Stockfish will almost always find a winning move if one exists. Though sometimes the wait can be excruciating.
[D] rk6/p1r3p1/P3B1Kp/1p2B3/8/8/8/8 w - - 0 1
tpoppins wrote:
In the past two years there have been numerous posts about an alarming number of positions SF evaluates as 0.00. Someone even coined the term "Drawfish". I personally have seen hundreds of such positions on Let's Check, ranging from quiet to turbo-charged. There is a current thread discussing another such case and SF's "tunnel vision".
About "tunnel vision", I think I largely agree with Dann. Given enough time/depth, Stockfish will almost always find a winning move if one exists. Though sometimes the wait can be excruciating.
[D] rk6/p1r3p1/P3B1Kp/1p2B3/8/8/8/8 w - - 0 1
And half a day in time.
I guess it is the price for sophisticated pruning.
Sacrifices can fool it so that the correct move choices are very hard to find. But there is no technique that simply abandons the search. They only reduce it. But that can make it take a very long time to solve some problems, like this one.
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.
tpoppins wrote:
In the past two years there have been numerous posts about an alarming number of positions SF evaluates as 0.00. Someone even coined the term "Drawfish". I personally have seen hundreds of such positions on Let's Check, ranging from quiet to turbo-charged. There is a current thread discussing another such case and SF's "tunnel vision".
About "tunnel vision", I think I largely agree with Dann. Given enough time/depth, Stockfish will almost always find a winning move if one exists. Though sometimes the wait can be excruciating.
[D] rk6/p1r3p1/P3B1Kp/1p2B3/8/8/8/8 w - - 0 1
And half a day in time.
I guess it is the price for sophisticated pruning.
Sacrifices can fool it so that the correct move choices are very hard to find. But there is no technique that simply abandons the search. They only reduce it. But that can make it take a very long time to solve some problems, like this one.
It is simple to fix it without a significant change in playing strength.
Simply do not use null move pruning when the number of the legal moves of the opponent is small and calculate number of legal moves for the opponent only when remaining depth is at least 10 so it is not too expensive.
tpoppins wrote:
In the past two years there have been numerous posts about an alarming number of positions SF evaluates as 0.00. Someone even coined the term "Drawfish". I personally have seen hundreds of such positions on Let's Check, ranging from quiet to turbo-charged. There is a current thread discussing another such case and SF's "tunnel vision".
About "tunnel vision", I think I largely agree with Dann. Given enough time/depth, Stockfish will almost always find a winning move if one exists. Though sometimes the wait can be excruciating.
[D] rk6/p1r3p1/P3B1Kp/1p2B3/8/8/8/8 w - - 0 1
And half a day in time.
I guess it is the price for sophisticated pruning.
Sacrifices can fool it so that the correct move choices are very hard to find. But there is no technique that simply abandons the search. They only reduce it. But that can make it take a very long time to solve some problems, like this one.
It is simple to fix it without a significant change in playing strength.
Simply do not use null move pruning when the number of the legal moves of the opponent is small and calculate number of legal moves for the opponent only when remaining depth is at least 10 so it is not too expensive.
Instead of a special calculation for the legal move count, why not detect it on the fly?
At every ply in the search, we must have a move count number.
So, for either side, if the legal moves is at a certain number, disable null move pruning forward from there.
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.
