Polugaevsky Variation, 10.exf6, 16...Qc7, 16...Bb7

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Rebel, chrisw

Peter Berger
Posts: 653
Joined: Thu Mar 09, 2006 2:56 pm

Re: Polugaevsky Variation, 10.exf6, 16...Qc7, 16...Bb7

Post by Peter Berger »

As I seem to be the only one who fed this to Stockfish for some time:

Neue Partie Line
1n2k2r/2qr1p2/p3p3/6Np/Pp2Q3/2b5/2P1B1PP/1R3R1K w k - 0 1
Analysis by Brainfish 221119 64 BMI2:
23.h3 Th6
Weiß steht deutlich besser: +/- (1.55 ++) Tiefe: 60/86 11:07:07 132896MN, tb=32482205

I don't have the slightest idea what this means though. h3 is certainly a candidate here even for a mediocre player, but whether this means this position is really won?! who knows ...
BeyondCritics
Posts: 396
Joined: Sat May 05, 2012 2:48 pm
Full name: Oliver Roese

Re: Polugaevsky Variation, 10.exf6, 16...Qc7, 16...Bb7

Post by BeyondCritics »

This position was specifically posted by me, because the alleged key move 26.h3 is not a "candidate move" for me, by far not. I understand that not everyone bothers that, so i must apologize for having asked for your time.
When stockfish presented 23.h3... to me..., i was shocked, so i choose to investigate this position a little bit closer, since in my experience these dark holes are the best opportunities to learn something new. And with the help of stockfish it turned out, that in this specific postion black is seemingly so neatly tied up already, that all what is needed is some luft for the king and whites attack is crushing, at least that is what i read out of the stockfish pvs. In order to draw such a conclusion OTB, you would have to investigate and understand thouroughly every(!) sensible option for black, since 26.h3 makes a loophole for the king but does otherwise absolutely nothing for white, execept for a irrevocably weakening of the g3 square (Is e.g. h4, Sc6-e7-f5 an option in the long run?).
As a side note: "Candidate moves" are those you consider in the very first stop of the thinking process, e.g. by "pattern matching" (https://en.wikipedia.org/wiki/Candidate_move). The next step could be, to investigate these moves very, very carefully and try to detect problems with them. You would then go back and try to fix this problems and that may lead you to other moves.
E.g. in this position
[d]1n2k2r/2qr1p2/p3p3/6Np/Pp2Q3/2b5/2P1B1PP/1R3R1K w k - 0 23
you may notice first that you are a pawn down, blacks king is in the centre and his pieces are not very active. All this cries out for energetic action.
You might want to consider 23.Nxf7, 23.Rf3 and in a long game maybe 23.Qh4(ideas Ne4,Bxh5), 23.Qa8!?, 23.Qe3(idea Ne4), 23.Qf3. Leela finds 24.Bd3. These are candidate moves, you get them by pure intuition.
Next step could be to weed out big problems and sort them for plausibility. You would e.g. notice, that 23.Qe3?? looses on the spot, because of 23...Bd2. Now you have learned the counter blow 23...Bd2. Next after 23.Qh4, there is 23...Qe5 24.Ne4 Rd4, so you could have learned Qe5, Rd4 ideas.
Next step could be to investigate some lines deeper.
Then you would investigate 23.Nxf7 deeper, seeing that it not works. 23.Rf3 gives an advantage but is not decisive. Maybe now you would notice, that your weak backrank is a problem in this line and therefore you would go back trying 23.h3, with the "threat" 24.Rf3.
But frankly my assessment is, that all this would require near super human performance, ymmv.