I like Pelletier-Nakamura from 2015 (http://www.chessgames.com/perl/chessgame?gid=1806311). In an age old theoretical line, Pelletier sacs suddenly a bishop with 13.axb6.Kempelen wrote:Hello,
I am going to make a speech at my club about curiosities of computer chess and about making the people to know the pros and cons of analizing with an engine. I need material for the speech. Would anyone here like to share with me :
- Positions where engine does not evaluate correctly the position
- Fortress
- Sac that are incorrect acording to engine, but could be a very practical choice in OTB human games
- or any position that serves to a teaching porpose for the speech.
Thank you very much in advance.
Fermin
[d]r1bq1rk1/2p1npbp/1p1p2p1/P2Pp3/2P1Pn2/B1N5/P2NBPPP/R2Q1RK1 w - - 3 13
If you look only at stockfish pv's, you will never see that move, since stockfish thinks it throws away its positional advantage. This would be true for correspondence chess, but over the board amongst humans the sac turned out to absolutely deadly on the spot. Nakamura, despite thinking for a considerable time, made the move 14...Ra5, which seemingly already looses or is very close to that. Even with best defence, the sacrifice seems to be sound, though.
This demonstrates, that it is still posssible to capitalize on its creativity, even in the presence of the beast, if you know how.