[Event "Koelle 2017"]
[Site "Mechanics Institute"]
[Date "2017-11-23"]
[White "Rodney Jacobs"]
[Black "Jamie Brotheridge"]
[WhiteElo "1704"]
[BlackElo "1566"]
[Result "1-0"]

1.e4 c6 2.d4 d5 3.Nc3 {(RJ) Jamie & I have a standing argument over the bayonet attack (White's g4 after Black's Bf5 in the advance variation of the Caro-Kann.) I was tempted to play it again, particularly after Jamie recently assured me it was rubbish. But decided to play the classical line, particularly after having spent a lot of time on it preparing for a game with Heath} dxe4 4.Nxe4 Nf6 {(RJ) I had a bit of an idea Jamie might play this.} 5.Nxf6+ exf6 {(RJ) The conservative approach. The downside for Black is that, in an endgame, White can often create a passed pawn using his queenside majority, however Black can't do the same on the Kingside. The alternative is the more dynamic gf, which gives Black the open g file to play with. It also has the advantage of not allowing White a safe queenside majority.} 6.Nf3 Na6 {(RJ) Setting a little trap.} 7.Bxa6? {(RJ) And falling headlong into it. Strangely, this weak move is the third most popular given on Chessbase, the best moves being given as a3 or c3, both of which keep Black's bishop and knight out of b4.} Qa5+ 8.Qd2 {(RJ) Chessbase gives c3 or Bd2. I don't think my Qd2 achieves anything.} Qxa6 {(RJ) Black has relieved White of his strong white bishop, at the expense of an inactive knight. Also the position of the Black queen makes it very awkward for Black to complete his development naturally. The engines put it it around - .4} 9.Qe3+ Be6 10.Bd2 {(RJ) Continuing to play inaccurately. b3 is better.} Bd6 11.b3 O-O {(RJ) It's now - .1. White has to play very carefully here, with the looming Re8.} 12.Qe2 {(RJ) Best.} Qxe2+ 13.Kxe2 Bg4 14.Be3 Rfe8 15.h3 Bxf3+ {(RJ) Now it's Black who starts to play inaccurately. After the game Jamie said he should not have given up this bishop. The engines agree, giving Bf5 as best.} 16.Kxf3 Rad8 17.Rad1 {(RJ) Best.} a6 {(RJ) The engines indicate that the tide has turned, with White .5 up.} 18.g3 Rd7 {(RJ) Planning to double rooks on the e file, but a mistake. Black's rook and bishop on the d file are vulnerable.} 19.d5! cxd5 {(RJ) A further innacuracy. As we felt after the game, c5, keeping the d file closed, is necessary, even though that leaves White with a strong passed pawn. After this, White's game plays itself.} 20.Rxd5 Kf8? {(RJ) This is, though, the big one. The engines point out that Rdd8!, admitting the error on move 18, and planning to meet Rhd1 with Bc7, holds, although White still has the initiative. Neither of of us noticed this during the game or afterwards.} 21.Rhd1 Ke7 22.c4 Red8 23.Bb6 {#r}