DDT3000 – In-lightning, 15 minutes per game, ICC, 2011
White to move. Would you, or would you not sacrifice an exchange here?
It depends, and while this is a question of judgement but I would argue that 20. Rxf6!? is not really a sacrifice, but instead a fair trade off. I took on f6 without much thought, and went on to lose a complicated game by dropping a bishop in a probably drawn endgame. But given the opportunity – I would take on f6 again and have a fun game. On to the actual positional justification of the move:
I was not too happy with my position, since my bishop was passive, rook on b1 - awkwardly placed, and while the knight had some prospects on f5 and d5 - Black has ability to cover those squares. So I took an opportunity, which, given that both White and Black had about 10 minutes left on the clock (no increment) - I would take again. In a standard time controls game - I would likely not take such a decision very lightly, but in rapid chess - this move is almost standard.
After 20. Rxf6 gxf6 21. Ng3 In return for exchange, White gets the following advantages:
1) The knight gets a permanent control over f5
2) The bishop - permanent control over d5. Opposite coloured bishops definitely contribute to compensation.
3) Black's king is somewhat weakened.
Sure an exchange is worth something, sure White does not have any immediate threats; but the long term control over the light squares - is enough to create threats for your opponent, which given limited time - make things equally hard for White and Black. White has no bad pieces and a lot of positional trumps, so even if Black has a chance to try to convert an advantage, he would first have to defend for 20 moves. The game was not without mistakes, but here is a position which we arrived at after a few more moves
Even though I went wrong with 32. Nd6? – I still maintained a lot of chance until the very end. The risk White takes upon with Rxf6 in my view is equal to the risk that Black takes by allowing such sacrifices.