DrMAL: Interesting opening, transposes to a KI attack until 5.a3 seems H-bomb wanted something unusual to go way out of book and simply play a basically equal but unknown middlegame (perhaps even slightly in black's favor).As noted above, 19.Nb8 may have been better but I don't think black has any real advantage there either. In that line, maybe 24...e4 was best for some possible advantage, 24...Bxb4 is met by 25.axb4 and if 25...a6 the knight on b8 is quite silly (tiny advantage white). Otherwise, both 25...Na6 26.b5 Nb4 27.Bf1 cxb5 28.Nxb5 and 25...Rd2 26.e4 f4 27.gxf4 exf4 28.e5 lead to (also insignificant) advantage for white.
Taking the knight with 21...Bxc3 was not particularly great (e.g., 21...Rxd5 was probably better). After 24.Qxa7 Rd2 25.c6 black is busy trying to get a pawn back while white can simply get his rook active via Rd1 (e.g., 25...Bc4 26.Rd1 Rxb2 27.Qd4 Nxc6 28.Qc5) with slight advantage.
Naka played 24.Nb5 instead for B vs. N and a pawn up but black's position compensated (equality). Similarly, 27...Nd3 looks nice at first but simply 28.Rf1 back and the pawn is protected black has no real attack (but white is still up a pawn). Perhaps 27...Kh7 was best to just get off the back rank and prepare something better (simply accept the game is probably drawn).
The follow-up 28...h5 wastes a tempo, 29...g6 more so and now white has an advantage after 30.h4. Black's pieces are tied up with a do-nothing attack while the pawn on a6 is now tasty. Nakamura must have missed 31.Qa5! here to gobble it.
After 31.Qb8 the position looks double-edged but still drawish until 32...Ra2 (instead of 32...Qe7 or 32...Rb2) a tactical mistake. After 33.b5 the only decent move to survive was 33...Rb2 (34.Qb7 Kg7 and a likely draw).
Kibitzing
.
unclear).
