There I wanted to resign already, but at that moment I was not prepared to start new game - to continue the game with few more moves...

The immediate push of the pawn means that it is one square away from Queening. This is a very strong latent threat. This would limit where the Black Queen could travel regarding the guarding of the b8 square. In addition keeping the White Queen at d3 is more efficient in preventing nuisance checks from Black. That fact and the immediate threat on b8 combined with that ensuing check makes 38.b7 more efficient in ending the game more quickly. Assuming 38. Qxd5, Qa1 , 39. Kg2,Qb2 , 40. Kh3,Qf3 another check looming at f1 depending on Whites next move etc etc. Anyway we are back where we started, and all White has done is just move his King around really. Black was lost and I wanted to finish the game asap. I think 38. b7 does it much quicker. A matter of taste as you say.

johnnyawesome is right: 38.b7! is the better move than 38.Qd5, though both seem to win. My first thought here that 38.Qxd5(!?) would help support the pawns (by denying the Black Queen the e5-square - point I omitted to mention in my original comment), places the Queen even more centrally than it is at d3, and covers the long diagonal. It is true that 38...Qa1ch 39.Kg2 Qb2ch 40.Kh3? is inconvenient for White after 40...Qf2, but White is not forced to play to h3. Now, at d3 the WQ puts a crimp in Black's counteraction against the WK. With the Q at d5, can White escape the checks? 38.Qxd5!? Qa1ch 39.Kf2 Qb2ch 40.Ke3 Qc3ch 41.Ke4 Qe1ch 42.Kd4 Qf2ch 43.Ke5 Qxb6 44.Kf6 Qc7 45.Qd7 ... OK, that's just a sampling, and Black doesn't have to take on b6, but White will (I think - I could be proved wrong!) escape the checks. All that takes time, however, and, on reflection, 38.b7 is a quicker road to the win. E.g. 38.b7! Qe5 39.Qc2! Qc7 40.Qc5... Driven back to a passive position Black's Q has no counterattack available. Maybe this is why loreta rejected ...Qe5 as a defence here and in subsequent moves.

An instructive game, showing the dangers of too quiet play early on. Maybe that's one of the hazards of the Caro-Kann - as Black's early play is restrained anyhow, he can too easily allow White to build a big lead in active development.

I have used the Caro-Kann defence exclusively for many years now. I have just about played against most of the lines. I believe it is one of the strongest defences Black has against an e4 opening. I just love it. The fact that I dismantled it in this game does not detract from its strength. What makes it good to play, is the fact that in every line, only a few tyes of structures can be formed. Definitely, positional understanding and strategy is far more relevant than tactics. I have found strong end game skills are important. Basically, Blacks goals are to contest the centre and develop without creating major weaknesses. Black missed out in this game. Pawn structure is of paramount importance, as is piece placement depending on the structure. Hence my aversion to an early c5 on move 7 in this game. The Botvinnik- Carls gambit is an exception with c5 on move 3. This my favourite reply to 'The Advance Variaton'. Black gets to develop quickly but must be careful not to try to recover the pawn to quickly. Patience again is the key. The Caro-Kann is for patient strategic players.