Not precise, but fun.
Collins (1915) – Slater (2129)
USATE 2010 rd3
White doesn’t mind Black pushing the pawn to f4, as it takes the pressure off of e4 and allows White to pursue his queenside ambitions – namely preparing and playing c5.
If he’d played 14.Bd2, he would now have time to play 15. f3 and be good to go. However, he didn’t.
Denys Shmelov, in a beer-round postmortem: “After f3 appeared I expected the game to be over very soon.”
Black wants White to take on f3. White does not want to do so. The pawns would be both a target on the f-file and a wall keeping White’s pieces from helping save the king. As it so often happens, though, that might have been less dangerous than leaving the pawn on f3 as in the game.
18.Rae1 h5 So that after 19.Re3 Bh6 20.Rxf3 Rxf3 White would have to double the pawns after all, because his queen is busy guarding the Bd2.
Wins in a cool manner, though afterwards Denys pointed out 20…Qd7 is instantly 0-1!
Funny – I played that very move 4 moves ago, but here it never crossed my mind.
I guess I was preoccupied with calculations around 20…Qg5 (21.Kh2 Bxh3! 22.Kxh3 Rh4+!!). Then I decided 20…Bxh3 was more forcing than …Qg5.
The key move. Down a whole rook, Black is simply going to go Qh4-h3-g2 mate. It takes three tempi, but White just can’t stop it without giving up his queen as follows.
Sneaking the Queen to the kingside – but after Black’s reply she doesn’t have any safe squares there.
No rush. It’s trapped. 25.Kh2 h4! makes it even worse — because he’s forced to play Qxe5 instead of making me take on g3, when White can recapture fxg3 allowing a rook to cover g2 on the second rank).
White has some material but it’s wildly uncoordinated (Nh1 anybody?) and he’s going to lose his center pawns.