Stalemate Swindle

Frenez, one of my three readers, requested that I show a recent tournament stalemate. I was totally busted playing white against an 1800 player. The moves are not cleverly hidden like in Kubbel's studies; in fact, they're natural and obvious. White to play and draw:

69.b7! Qc7 70.b8=Q! Qxb8! stalemate! To my memory, this is the only one of my 540+ tournament games that ended in a complex stalemate. Now that Frenez has gotten the stalemate, I'm going to go in reverse chronological order like my Memento series and go over some of the crucial moments of this game. Feel free to change the channel when I get long-winded. I almost resigned this game prematurely. In the above diagram, I could see that b7 was going to be followed by Qc7, but Ka7 was going to be met by Qxa5+, so my only other option was to give up the b-pawn which was my only hope, but I hardly considered stalemate until I was one ply away and saw that it succeeded in giving me the draw. Notice that the Black King is at e6, blockading my passed pawn and preventing it from moving - we'll visit this later. The above diagram is already approximately equal. After 69.b7, Black has no real alternative to 69...Qc7 except perpetual check starting with 69...Qc6+ 70.Ka7 Qc7 {pinning the queening pawn} 71.a6 d4 72.Ka8 {unpinning} Qc6 {repinning} 73.Ka7, etc. Black blew the win one move earlier in the following diagram with Black to play:

Almost everything wins, EXCEPT 68...Qc5?, e.g. 68...Kd7 leads to mate in 8, 68...Qb4 leads to mate in 9, 68...Qb2 leads to mate in 10, the greedy 68...Qxe5 leads to mate in 10, and even the dithering 68...Qc4+ leads to mate in 10 as long as the follow-up is the approach of the Black King.

But before that, White blew a securely drawn position with some chances to swindle a win (white to play):

The moves just prior to arriving at this position were 59.f4 Kf7. I had realized that with my pawn back on f3, e5 was dangerous for me because of the way it can spring the d5 pawn free for a queening race. Unfortunately, playing 59.f4 and locking a bead on the hole at e5 made me think that e6-e5 had been prevented forever. It had not. With sober reflection on information that I definitely had access to, I could see that playing Kd7, Kd8, Kd7, and Kd6 would secure the draw for me. The e-file squares - e6, e7, and e8 - are crucial to White's winning chances and if he stays in contact with them, any variation that Black initiates with ...e5 fxe5 f4 e6 will probably end badly for Black because the e-pawn will likely check the Black King on the way to e8, giving the new White Queen time to execute the upstart Black Pawn on f2. With more sober reflection, I should have been more patient here and tried to advance my queenside pawns. 60.a4! is one move away from a winning position. However, if Black also reflects soberly, he can realize that e5 loses, and blockading my queenside pawns can secure a draw for him also. 60.a4! a5! 61.Kd7 Kf6 is drawn as is 60.b4 b5! 61.Kd7!= If somehow Black doesn't see the danger, then the variation 60.a4 Kf6? 61.a5! Kf7 62.b4 Kf6 63.Kc7 is won for White. e.g. 63...Ke7 64.Kxb7! e5 65.fxe5! f4 66.b5! axb5 67.a6! f3 68.a7 f2 69.a8=Q f1=Q 70.Qa6! Qc4 71.Qf6+! Kd7 72.Qd6+! Ke8 73.e6! initiates checkmate in 6 moves. If after 70.Qa6!, Black tries to hold with 70...Qf5 71.Qd6+ Kf7 72.Qxd5+ also keeps White's winning chances alive. Going back to 60.a4!, one last bit of interesting subtlety is that 60...b6 61.b4 Kf6! 62.b5 axb5! is still drawn albeit precariously.

At this point, my greed for pouncing on the queenside pawns made me impatient and I played 60.Kc7?? e5! 61.fxe5 Ke6. Note that Ke7 would have also prevented White from queening and also gives White a move if stalemate is a problem. Black correctly calculated that given a chance with 61...f4? White would have played 62.Kd7! and White is back from the dead and winning, e.g. 62...f3 63.e6+ Kg7 64.e7 f2 65.e8=Q f1=Q 66.Qe5+ Kf8 67.Qb8+ Kg7 68.Qxb7 and White is +4.0 in Stockfish's evaluation. For continuity, the moves that got us to the pivotal checkmate or stalemate position above were 62.Kxb7 f4 63.Kxa6 f3 64.b4 f2 65.a4 f1=Q+ 66.b5 Qc4 67.a5 Qxd4 68.b6

One more bit of shoulda-coulda. From the above diagram, rewind the game another 10 moves and we arrive here with White to make move 50:

Material is equal. Black's e-pawn is backward. White's h-pawn is passed. The crucial idea that I missed was that my h-pawn is vulnerable if it runs too far ahead, but if I centralize my White King, the h-pawn can be a valuable distraction. To that end, 50.Ke3! was a priority to prepare for a possibility of 50...Nf6? 51.Bxf6 Kxf6 52.Kf4 with winning advantage. Because of safe, time-wasting moves like Rh1-h2 and back, White can eventually zugzwang Black's king and or rook to allowing h5-h6, h6-h7, and Kf4-e5. Instead, I finally pushed back this troublesome knight on e4 and played 50.f3 Nf6. Here, 51.Bxf6 Kxf6 52.Ke3 Kg5=. The game went 51.h6 Ng8 52.Bf4 Kg6 53.Ke3 Nxh6 54.Rxh6+ Rxh6 55.Bxh6 Kxh6 56.Kf4 Kg6 57.Ke5 Kf7 58.Kd6 Kf6 59.f4 Kf7 and we meet the above "drawish" position.