But I had compensation in the shape of the bishop-pair, and in the diagram I have hit back in the centre with 19...c5.
 |
| The critical moment - how should Black respond? |
*****
*****
*****
*****
Clearly, if king safety were the only important factor in chess, Black is in big trouble.
But the white queen cannot mate by itself, and meanwhile White has a knight en prise.
Stockfish15 and Komodo13.02 give White at least a slight edge, but it is dependant on finding the best continuation.
The game saw 20.Nf3?, which may look reasonable at first glance as the knight attacks g5 and may be able to support the white queen.
But the engines reckon Black is winning after 20...g4, and even more so after the game's 20...c4.
The point is the white bishop has no good move. 21.Ba4 b5 does not help the white cause, so the only bishop move is the retrograde 21.Bd1, but then 21...Nf5 22.Qh3 (best, according to the engines) Qd3 (22...Bc5+ and 22...Qb6+ are also strong) is devastating, thanks to the threat of ...Bc5+ followed by ...Qxe1#. White's best is probably 23.Nbd2, but 23...Ne3 24.Re1 Qxc3 25.Ba4 gxf4 leaves Black two pawns up and with the better coordinated pieces. Meanwhile the black king, although with very little pawn-cover, seems safe.
Instead White played 21.fxg5!?, but after 21...Nf5 the engines agree White's best is the game's 22.Qxh7+!? Kxh7 23.Bc2. However it is all-but-hopeless as White gets just two pieces for the queen.
Going back to the diagram, it should not be hard to see 20.Bc2 is a better try, but after 20...Bxc2 21.Nxc2 the position favours Black, according to the engines, although in practice it would be tricky to play since the black king is exposed.
When an enemy king is shorn of most of its pawn-cover, it is natural to look for mating attacks, even if these cost material.
With that attitude in mind, the engines' 20.Rf3! should be looked at, especially as 20...cxd4?? 21.Rh3
gives White a mate-in-two.
gives White a mate-in-two.
Black has to find 20...Nf5, although after 21.Nxf5 Bxf5 22.Na3 the engines reckon White is at least slightly better.
I am not claiming everyone should be able to find 20.Rf3!, but the move played in the game strikes me as an example of blind optimism over concrete analysis, an impression reinforced in the game when my opponent slammed down the follow-up Qxh7+.
The situation reminds me of grandmaster David Bronstein's warning to the effect that "losing your objectivity nearly always means losing the game."
I'd have thought most half-decent players would consider 20. Rf3. You can't mate with Queen alone, so you need to think how to get a second piece into the attack and, more importantly, what mating patterns exist. Getting a Knight to F5 (to mate on G7) isn't going to happen, so the only other obvious alternative is a rook lift to the H file ...
ReplyDeleteYou may well be right. Certainly Rf3 is the most direct way of supporting the queen, and seems obvious once the reply ...cxd4 is ruled out. On the other hand PB is rated 2038, and I very much regard that as half-decent. Perhaps the problem was he somehow had it fixed in his mind that the attacked knight had to move.
Delete