Here's a scenario that is not uncommon in open tournaments. Your latest opponent is a complete unknown who is rated perhaps 150-200 points below you, and right in the opening he plays something slightly unusual. I don't mean a respectable side variation, I mean an odd move order that gives you an opportunity to play for an immediate advantage. The chance will disappear if you don't take it now. What should you do?
I for one like to spend some time trying to refute such moves. Just imagine, this could be the big reason why no one plays this way (or will not play this way after your brilliant win makes the rounds!) If you pass up this chance, memories of the game will always be tinged with regret, especially if you go on to lose.
After the opening moves 1.c4 c5 2.Nf3 Nf6 3.Nc3 Nc6 4.d4 cxd4 5.Nxd4 a recent opponent of mine played the unusual 5...Nxd4!? 6.Qxd4 g6. The position is shown in the first diagram.
There is no compelling reason for Black to exchange knights on the fifth move unless he is worried that White will avoid simplification by playing 6.Nc2 or 6.Nb3 in reply to 5...g6. Of course White should not consider exchanging on c6 himself because it only strengthens Black's pawn position.
After 6...g6 White has a decision to make. Should he transpose back to regular lines by playing 7.e4 or 7.g3, or should he try to inflict damage on Black's pawn structure with the aggressive 7.Nd5!? After the further moves 7...Bg7 8.Bg5 (not 8.Bf4?! d6! intending 9...e5) 8...0-0 White has another choice to make. He must capture on f6, or else Black will play 9...Nxd5 and break the bind. There are two ways to do this: 9.Nxf6+ or 9.Bxf6+. Lots to think about!
During the game I concentrated my efforts on 9.Nxf6+. Because it creates a significant material imbalance (centralised knight vs unopposed dark-squared bishop), I don't think White should take with the bishop unless he is ready with prepared analysis or some previous experience.
After 9.Nxf6+ exf6 10.Bf4 f5 11.Qd2 Black has the obvious move 11...Qb6. White has to protect the b-pawn since 12.Bd6 Qxb2 13.Qxb2 Bxb2 14.Rb1 Bc3+ 15.Kd1 Re8 16.c5!? (White is a pawn down and must try for compensation) 16...Be5! is starting to look good for Black. So White's next move has to be 12.Rb1. Castling queenside instead drops a pawn after 12...Qxf2.
After 12.Rb1 it was time to pause and take stock of the situation (remember that I'm analysing on my mental chessboard!) White has managed to give Black a backward d-pawn on an open file, but in order to accomplish this he has had to cash in his lead in development. If White were given the free moves g3, Bg2, 0-0 and Rfd1 (or even Bd6 and c5) then one could speak of a serious positional advantage for White. Unfortunately, Black can throw a large spanner into the works by playing 12...d5!? (see the second diagram.)
This is a highly thematic idea. At one stroke Black eliminates his backward pawn and opens up the centre in order to put pressure on White's position. White cannot play the quiet 13.e3 because of the variation 13...dxc4 14.Bxc4 Rd8 15.Qc2 Qb4+, which strands his king in the centre. And 13.Qxd5 is no better in view of 13...Be6 14.Qb5 Qxb5 15.cxb5 Bxa2 with a clear advantage for Black. The only reply to 12...d5 is 13.cxd5. I got this far in my analysis and decided I didn't like White's position. He is a pawn ahead, and it is a centre pawn, but he is at least three moves away from completing his development. I remembered Capablanca's guideline that a pawn sacrifice in the opening is usually sound if the opponent's development is delayed by at least three moves. That is certainly the case here, and in fact Black can keep up the pressure with 13...Rd8.
So it turns out that 7.Nd5!? followed in due course by 9.Nxf6+ is a dubious idea. After spending 5 minutes or so on this analysis I made the practical decision to reenter the main lines, in this case by playing 7.g3.
Later on I had a look in my database and found the old game Capablanca-Colle, Barcelona 1929. Capablanca had the same opportunity to play 7.Nd5, but he too transposed to the main lines, this time with 7.e4. That settled the matter, at least in my mind. If 7.Nd5 gave White an advantage greater than the main lines, Capablanca would certainly have played it!