Unseen Mate-in-3

Hello,

I wanted to ask how the stockfish engine used by chess.com analyzes the position captured in the provided image? The board position is a mate-in-three problem. The only lines recommended by the engine are mating sequences of length 4. However, if the first move is correctly played (that being 1. Rh3 b3), the engine will then recommend the rest of the line as a mating sequence of length 2. 

I have the latest version of stockfish (12) installed on my own machine and have used the stockfish engine in [Deleted by Mod: PeacemaKing; Please do not discuss/advertise competitive chess websites]. Both suggest the correct mating sequence of length 3 starting with 1. Rh3.

What makes the stockfish engine employed by chess.com different from these other applications of the engine? I would like to better understand the features provided on this site.

Sincerely,

Ian

Can you detail what the mate in 3 is ?

I am struggling to see it.  After 1. Rh3 b3 white must move the Q off the 8th rank or it will be stalemate. Wherever the white Q goes (excepting d5) black can move the B with check.

Aha....I just seen it. Very neat mate in 3.

I suspect it will be down to how deep the analysis looks. Chess engines don't look for elegant mates, they simply crunch numbers..lots of them. The move 2. Rh1 will be put further down the candidate-move scale because it will be given a -6 score. A longer and deeper analysis of the position will find the move eventually.

It seems that you found out, but I'll detail the sequence anyway. The correct solution is 1. Rh3 b3 2. Rh1 gxh1=Q+ 3. Qxh1#

Also, I apologize to the moderators for mentioning competing sites. I'm not trying to promote other sites or criticize chess.com. I wanted to provide different use cases that all involve Stockfish.

I certainly agree that chess engines don't look for elegant mates. The main thing I'm trying to understand is: How is the Stockfish engine used by chess.com different than the current version of Stockfish that I've downloaded to my personal computer?

Changing to max depth on self analysis finds it immediately. As does changing to Komodo without depth change.

isn't just to play Ra3 or Rc3 when bx, Qc8 and then either cx or ax or even a2 or c2 Qb7# ??

nvm he can play b3

but I find it quite uncommon because in a losing position, most players will try to get material but who knows

Te reto 

Thanks Martin_Stahl. I changed to max depth, and it did find it for me as well. So why does max depth find it when depth 15 doesn't?

La partida

Answer is that changing the depth doesn't just change the depth, it also changes the pruning.  Depth is a meaningless number by itself, as it is just the deapest an engine gets in certain lines, the lines it finds most worth looking at.  Other lines are "pruned" away.  Even at depth 15, the engine does not look at every line to depth of 15.  Basically, depth 15 is code for "crappy superficial analysis" so its natural that there would be more pruning as well as the depth limitation.  All of that said, it is remarkable that it prunes so much it doesn't find a mate in 3.  

Isn't black just up a pawn here?

Thanks for posting the game. I actually like 9. ... Ndxe5 from your opponent, which seems to just win a central pawn. However, it is really strange then that the computer suggests that the recapture is the 'brilliant' move. Seems like the prior move was the better move and the latter was just forced, as you said.

Hey EdanLeung, that's not correct. I posted the M3 sequence. It's a forced line for Black. They only have one move during each of their turns.

Epiloque escribió:

Virtus_ian wrote:

Isn't black just up a pawn here?

yes, retaking was forced. I got a miniature against him in this game bc it was 10 mins. I just wanted to point out how poor the computer is. It should be noted that his blitz rating is 300 points lower than his rapid, and he is ~1550 rapid, but this was before the update. So this is not me beating on a lower rated player for fun, he was better than me.

 

for mine it says it is the best move but not brilliant soo what?

and black has only a 0.13 positional adavantage (bc the central pawns) but that is not enough to be considered "BRILLIANT MOVE"

Entonces la partida