AylerKupp: <<whiteshark> Timothy Taylor 's analysis is flawed. His line with <22.Bxf7+ Kh8 23.Rxb7> is busted by <23... Nc6! < >> with the idea/threat of ...Nd4 and ...Bxf3>Well, you're sort of right. After 22.Bxd7+ Kh8 23.Rxb7 White <is> busted by 23...Nc6, but if instead 22.Rxf7 then it is Black who has the worst of it, at least according to the latest Komodo 5.1r1 which gives the following lines at d=21:
1. [+0.68]: 23...Rxf7 24.Bxf7+ Kxf7 25.Qc4+ d5 26.Qxb4 Bxf3 27.Qb7+ Kg8 28.gxf3 d4 29.f4 Rb8 30.Qd5+ Qxd5 31.exd5 Rd8 32.fxe5 Rxd5 33.Kg2 Rxe5 34.Rxd4 a5 35.b3 Kf7 36.Kf3 Rc5 37.h4 Rc2 38.Ra4 Rc5 39.Kf4 Kf6 40.Rc4 Rf5+ 41.Kg3 Rd5
2. [+1.34]: 23...Kh8 24.Rxf8+ Qxf8 25.h3 Bxf3 26.Qxf3 Qxf3 27.gxf3 Rd8 28.f4 Nc6 29.fxe5 Nxe5 30.h4 Ng6 31.Kf1 Nf4 32.h5 Nxh5 33.e5 Nf4 34.Rd4 g5 35.Rxd6 Re8 36.Rxh6+ Kg7 37.Rxa6 Rxe5 38.Ra7+ Kf6 39.Rf7+ Kg6 40.Rc7 Kf6 41.Bc4 Re7 42.Rxe7
It is possible that better moves for Black would be found at higher search plies but I doubt it since Komodo's evals were increasing for PV=1, even if not by much at each search ply. And Houdini 1.5a agrees with Komodo, giving the following lines at d=25, also with its evals increasing at each search ply for PV=1:
1. [+0.50]: 23...Rxf7 24.Bxf7+ Kxf7 25.Qc4+ d5 26.Qxb4 Bxf3 27.Qb7+ Kg8 28.gxf3 d4 29.f4 Rb8 30.Qd5+ Qxd5 31.exd5 Rd8 32.fxe5 Rxd5 33.Kg2 Rxe5 34.Rxd4 a5 35.b3 Kf7 36.Kf3 Rc5 37.Rc4 Rd5 38.h4 g6 39.Ke3 Re5+ 40.Kd3
2. [+1.43]: 23...Kh8 24.Rxf8+ Qxf8 25.h3 Bxf3 26.Qxf3 Qxf3 27.gxf3 Rd8 28.f4 Nc6 29.fxe5 Nxe5 30.f4 Ng6 31.f5 Ne5 32.Rd5 Kh7 33.Ra5 Nd3 34.Bf7 Ra8 35.Bg6+ Kg8 36.Rd5 Nxb2 37.Rxd6 Nc4 38.Re6 a5 39.Kg2 Kf8 40.Kf3 Rb8 41.Kf4 Rd8 42.Rc6
True, the position for PV=1 after 36...Rc5 (the last move for which Komodo and Houdini agree is the best line) may well be a draw, although Rybka 4.1 thinks that White may have reasonable winning chances, evaluating the position a clear pawn up at [+1.00], d=24. But certainly it is White that has all the winning chances, so it probably represents the best practical try:
So Timothy Taylor was also sort of right. His comment after 22.Rb7: "Yet, after just glancing at this position, I saw – as any master should – the possibility of a combinatorial blow on f7, based on the hanging BN on b4". He just didn't pick the proper of the two moves, 22.Bxf7+ or 22.Rxf7 – as any patzer like me would have after (a) having been pointed to the idea and (b) shown the refutation of one of the two moves. Or, as Sherlock Holmes once said, "Once you eliminate the impossible, whatever remains, no matter how improbable, must be the truth." :-)