16-07-2013, 03:34 AM
From the recent Scottish Championships
Andy informs me this was the 35th game between us. Somewhere around the fifteenth game we realized that neither of us had ever been more than plus one, so we compared it to a set in a tennis match. Finally the ‘first set’ was won 12.5-10.5 by Andy (after I missed a few ‘set points’), who then walked through the second set 6-3 which I think was 3-0 in wins. However my interest is just kept from being extinguished by being one up in the third set…
Before the last round, 12 players could (in theory) win the Scottish Championship title outright, including both of us - but this latter outcome would require all five Scots on 5 to lose to their (mostly GM) opponents, and the others on 4.5 all to draw. If I could beat Andy I reckoned my chances of being the unworthy winner were about 2000 to 1 - not worth biting my nails for, I thought.
To the game
Andy describes this game as an epic; this is probably not because of the length of our little series, but due to the possibilities that arose near the end of the game. Our opinions in the post mortem changed constantly, and in fact our final consensus was also wrong. However, Houdini doesn’t do much better.
[pos]6k1/p4p1p/1p1q2p1/2pr4/8/4P1P1/PPQ2P1P/2R3K1 w - - 0 25[/pos]
Black had a comfortable opening and is clearly a bit better here. I played 25 Qa4 in order to see a bit of daylight and Andy replied Rd2. There followed 26 Rc2 Rd1+ 27 Kg2.
[pos]6k1/p4p1p/1p4p1/2pq4/Q7/4P1P1/PPR2PKP/3r4 w - - 6 28[/pos]
Black is at the crossroads; he can gradually improve his position with moves like h5 or Kf8 in anticipation of a rook ending, and keeping a plus – especially with me being a bit short of time (there must have been something wrong with the clock). Or he can consider the more direct Qd5+ which forces White’s hand because f3 loses a pawn after Rd2+, or if instead Kh3?? Rg1 penetrates; so W has to play e4 after which Black could play Qd3 (but watch that check on e8). On the other hand, e4 forces Black’s hand in turn because if then Qd4 the rook ending isn’t dangerous enough yet and if Qe6 or Qe5 White has the trick Rxc5.
Andy perhaps rather hastily went for the committal 27 … Qd5+ 28 e4 Qd3 (now the only way to keep trying to win) 29 Qe8+ Kg7 30 Qe5+
Now f6 is met most simply by Qc7+.
30 … Kf8 31 Qb8+
[pos]1Q3k2/p4p1p/1p4p1/2p5/4P3/3q2P1/PPR2PKP/3r4 b - - 6 31[/pos]
31 … Ke7!
A cunning move, as if Qc7+ Kf6! Qf4+ Ke6 the checks run out: Qg4+ Kd6. White can bail out with Qe2 but Black has made progress in the rook ending – the King is now handily placed thanks to my checks, and he can start with c4 and Kc5, though the draw can probably still be held.
32 Qxa7+
Now – the crunch.
[pos]8/Q3kp1p/1p4p1/2p5/4P3/3q2P1/PPR2PKP/3r4 b - - 0 32[/pos]
Black’s last move was played quite quickly, as Black can still safely go back to f8 and g7 with an easy draw.
What do YOU play?
32 … Kf6!
Andy took a while over this. The (!) is for the calculated gamble - which turns out to be correct - and the guile. I didn’t believe Andy would go for this, and was quite happy that he was presenting me with his queenside. In fact, I was willing him. Come on, Andy! (as they say…) Firstly I didn’t believe there would be a forced mate; secondly I was threatening perpetual check, and lastly, I would have three long, lovely minutes to look for a flight square or other defence. Was I on a break point!?
33 Qxb6+ Kg7 34 Qxc5 Qf1+ 35 Kf3 Rd3+
[pos]8/5pkp/6p1/2Q5/4P3/3r1KP1/PPR2P1P/5q2 w - - 3 36[/pos]
What do YOU play?
36 Kf4
36 Kg4 was possible – but why give him two or three checks instead of none? Didn't seem logical.
This also didn’t seem the logical moment to have my think, better to reserve the time for when I can see his actual threat. Besides, at the moment I’ve got a draw with Qe5+ as f6?? gets mated: Qe7+ Kh6 Qf8+ Kg5 g4+ and mates.
36 … Qg2!
[pos]8/5pkp/6p1/2Q5/4PK2/3r2P1/PPR2PqP/8 w - - 5 37[/pos]
This threatens a quick mate with Qf3+. It also scuppers my perpetual with the queen (with the BK on h5, the black queen stops my g4+) so the first thing I did was look for another one. What about the rook? Qe5+ f6 Rc7+ Kh6 is menacing but Rxh7+ Kxh7 Qe7+ does the trick, as with the h-pawn gone Black can’t run to h5.
I can’t run with the King as f6+ mates; and Kg4 invites pawn checks. Not seeing anything obvious, I took the draw with Qe5+ etc.
Result: 1/2 - 1/2.
Did I have better? In the post-mortem we thought White had 37 Qe5+ f6 38 Qe7+ Kh6 39 Kg4! now that the BK blocks that dangerous h-pawn check. True Black now has a perpetual with Qf3+ Kh3 Qh5+ and Qf3+ etc, but the question arose whether Black had a clear win or with the tournament situation perhaps inviting risks (only two results short of the required miracle), the game might happily continue along the precipice. Black has a decent winning/losing try with the other pawn check 39 … f5+! 40 Kf4 (evidently best…) Qf3+ Ke6 Qxe4+ Kf6 Qd4+
[pos]8/4Q2p/5Kpk/5p2/3q4/3r2P1/PPR2P1P/8 w - - 2 43[/pos]
43 Kf7 (Keep running! Qe5 loses for White)
43 … Qd5+ (Now Qg7+ loses for Black as after Ke8 Qg8+ the queen interposes with check)
44 Kf6 (not Qe6? the quiet move Qd8 wins).
And now it’s Black who has to repeat moves!
The most interesting line is that ‘illogical’ Kg4 I declined earlier, back at move 36:
[pos]8/5pkp/6p1/2Q5/4P1K1/3r2P1/PPR2P1P/5q2 b - - 4 36[/pos]
36 Kg4 is however logical to Houdini, which after 3 minutes evaluates as pretty much a White win (+0.90) despite the checks. For example 36 .. h5+ (which also stops my perpetual check) 37 Kh4 the most obvious move is Qd1 but h3 stops the immediate threats and White is 3 pawns up. We got that far in the PM.
It seems I missed a good opportunity to beat an IM. But as usual in this ending, ‘seems’ isn’t enough – the players and kibitzers were wrong. The computer is wrong! There is no win! Though, I need the computer to prove it…running it over longer than the time it would have in a game.
After 36 Kg4 Black plays Qd1+ 37 Kh4 h6:
[pos]8/5pk1/6pp/2Q5/4P2K/3r2P1/PPR2P1P/3q4 w - - 0 38[/pos]
Despite the near-decisive negative computer evaluation, Black is able to plug away and the advantage dissipates as White keeps shedding pawns to stop mate threats. In one line B gets his pawn to h3, in another he gets the e4 pawn then W has to do Qf4 and B trades queens, doubling the f-pawns then with the R on the seventh, takes on b2 - with a probably drawn ending!
I’ll just give two sample computer lines arising from this line (with genuine apologies to the Barefoot Bronstein in Edinburgh):
1) 38 f4 Kh7 39. Kh3 h5 40. Qf2 Qg4+ 41. Kg2 h4 42. Qe2 h3+ 43. Kf1 Qd7 and White has to play 44 f5 Kg7 45 Rc1 Rd2 46 Qc4 Rxh2
[pos]8/3q1pk1/6p1/5P2/2Q1P3/6Pp/PP5r/2R2K2 w - - 0 47[/pos]
2) 38 Rc1 Qe2 39 Kh3 Qxe4 40 Qc4 Qf3 41 a4 Kh7 42 Re1 Rd2 43Qf4 Qxf4 44 gxf4 Rxb2
[pos]8/5p1k/6pp/8/P4P2/7K/1r3P1P/4R3 w - - 0 45[/pos]
This heavy piece endgame would have been fun to play in time trouble, especially if the championship had happened to depend on the outcome! So here is my suggestion for championship tie breaks - all the players involved play preset, computer-generated ‘precipice’ positions, with five minutes on the clock...
Andy informs me this was the 35th game between us. Somewhere around the fifteenth game we realized that neither of us had ever been more than plus one, so we compared it to a set in a tennis match. Finally the ‘first set’ was won 12.5-10.5 by Andy (after I missed a few ‘set points’), who then walked through the second set 6-3 which I think was 3-0 in wins. However my interest is just kept from being extinguished by being one up in the third set…
Before the last round, 12 players could (in theory) win the Scottish Championship title outright, including both of us - but this latter outcome would require all five Scots on 5 to lose to their (mostly GM) opponents, and the others on 4.5 all to draw. If I could beat Andy I reckoned my chances of being the unworthy winner were about 2000 to 1 - not worth biting my nails for, I thought.
To the game
Andy describes this game as an epic; this is probably not because of the length of our little series, but due to the possibilities that arose near the end of the game. Our opinions in the post mortem changed constantly, and in fact our final consensus was also wrong. However, Houdini doesn’t do much better.
[pos]6k1/p4p1p/1p1q2p1/2pr4/8/4P1P1/PPQ2P1P/2R3K1 w - - 0 25[/pos]
Black had a comfortable opening and is clearly a bit better here. I played 25 Qa4 in order to see a bit of daylight and Andy replied Rd2. There followed 26 Rc2 Rd1+ 27 Kg2.
[pos]6k1/p4p1p/1p4p1/2pq4/Q7/4P1P1/PPR2PKP/3r4 w - - 6 28[/pos]
Black is at the crossroads; he can gradually improve his position with moves like h5 or Kf8 in anticipation of a rook ending, and keeping a plus – especially with me being a bit short of time (there must have been something wrong with the clock). Or he can consider the more direct Qd5+ which forces White’s hand because f3 loses a pawn after Rd2+, or if instead Kh3?? Rg1 penetrates; so W has to play e4 after which Black could play Qd3 (but watch that check on e8). On the other hand, e4 forces Black’s hand in turn because if then Qd4 the rook ending isn’t dangerous enough yet and if Qe6 or Qe5 White has the trick Rxc5.
Andy perhaps rather hastily went for the committal 27 … Qd5+ 28 e4 Qd3 (now the only way to keep trying to win) 29 Qe8+ Kg7 30 Qe5+
Now f6 is met most simply by Qc7+.
30 … Kf8 31 Qb8+
[pos]1Q3k2/p4p1p/1p4p1/2p5/4P3/3q2P1/PPR2PKP/3r4 b - - 6 31[/pos]
31 … Ke7!
A cunning move, as if Qc7+ Kf6! Qf4+ Ke6 the checks run out: Qg4+ Kd6. White can bail out with Qe2 but Black has made progress in the rook ending – the King is now handily placed thanks to my checks, and he can start with c4 and Kc5, though the draw can probably still be held.
32 Qxa7+
Now – the crunch.
[pos]8/Q3kp1p/1p4p1/2p5/4P3/3q2P1/PPR2PKP/3r4 b - - 0 32[/pos]
Black’s last move was played quite quickly, as Black can still safely go back to f8 and g7 with an easy draw.
What do YOU play?
32 … Kf6!
Andy took a while over this. The (!) is for the calculated gamble - which turns out to be correct - and the guile. I didn’t believe Andy would go for this, and was quite happy that he was presenting me with his queenside. In fact, I was willing him. Come on, Andy! (as they say…) Firstly I didn’t believe there would be a forced mate; secondly I was threatening perpetual check, and lastly, I would have three long, lovely minutes to look for a flight square or other defence. Was I on a break point!?
33 Qxb6+ Kg7 34 Qxc5 Qf1+ 35 Kf3 Rd3+
[pos]8/5pkp/6p1/2Q5/4P3/3r1KP1/PPR2P1P/5q2 w - - 3 36[/pos]
What do YOU play?
36 Kf4
36 Kg4 was possible – but why give him two or three checks instead of none? Didn't seem logical.
This also didn’t seem the logical moment to have my think, better to reserve the time for when I can see his actual threat. Besides, at the moment I’ve got a draw with Qe5+ as f6?? gets mated: Qe7+ Kh6 Qf8+ Kg5 g4+ and mates.
36 … Qg2!
[pos]8/5pkp/6p1/2Q5/4PK2/3r2P1/PPR2PqP/8 w - - 5 37[/pos]
This threatens a quick mate with Qf3+. It also scuppers my perpetual with the queen (with the BK on h5, the black queen stops my g4+) so the first thing I did was look for another one. What about the rook? Qe5+ f6 Rc7+ Kh6 is menacing but Rxh7+ Kxh7 Qe7+ does the trick, as with the h-pawn gone Black can’t run to h5.
I can’t run with the King as f6+ mates; and Kg4 invites pawn checks. Not seeing anything obvious, I took the draw with Qe5+ etc.
Result: 1/2 - 1/2.
Did I have better? In the post-mortem we thought White had 37 Qe5+ f6 38 Qe7+ Kh6 39 Kg4! now that the BK blocks that dangerous h-pawn check. True Black now has a perpetual with Qf3+ Kh3 Qh5+ and Qf3+ etc, but the question arose whether Black had a clear win or with the tournament situation perhaps inviting risks (only two results short of the required miracle), the game might happily continue along the precipice. Black has a decent winning/losing try with the other pawn check 39 … f5+! 40 Kf4 (evidently best…) Qf3+ Ke6 Qxe4+ Kf6 Qd4+
[pos]8/4Q2p/5Kpk/5p2/3q4/3r2P1/PPR2P1P/8 w - - 2 43[/pos]
43 Kf7 (Keep running! Qe5 loses for White)
43 … Qd5+ (Now Qg7+ loses for Black as after Ke8 Qg8+ the queen interposes with check)
44 Kf6 (not Qe6? the quiet move Qd8 wins).
And now it’s Black who has to repeat moves!
The most interesting line is that ‘illogical’ Kg4 I declined earlier, back at move 36:
[pos]8/5pkp/6p1/2Q5/4P1K1/3r2P1/PPR2P1P/5q2 b - - 4 36[/pos]
36 Kg4 is however logical to Houdini, which after 3 minutes evaluates as pretty much a White win (+0.90) despite the checks. For example 36 .. h5+ (which also stops my perpetual check) 37 Kh4 the most obvious move is Qd1 but h3 stops the immediate threats and White is 3 pawns up. We got that far in the PM.
It seems I missed a good opportunity to beat an IM. But as usual in this ending, ‘seems’ isn’t enough – the players and kibitzers were wrong. The computer is wrong! There is no win! Though, I need the computer to prove it…running it over longer than the time it would have in a game.
After 36 Kg4 Black plays Qd1+ 37 Kh4 h6:
[pos]8/5pk1/6pp/2Q5/4P2K/3r2P1/PPR2P1P/3q4 w - - 0 38[/pos]
Despite the near-decisive negative computer evaluation, Black is able to plug away and the advantage dissipates as White keeps shedding pawns to stop mate threats. In one line B gets his pawn to h3, in another he gets the e4 pawn then W has to do Qf4 and B trades queens, doubling the f-pawns then with the R on the seventh, takes on b2 - with a probably drawn ending!
I’ll just give two sample computer lines arising from this line (with genuine apologies to the Barefoot Bronstein in Edinburgh):
1) 38 f4 Kh7 39. Kh3 h5 40. Qf2 Qg4+ 41. Kg2 h4 42. Qe2 h3+ 43. Kf1 Qd7 and White has to play 44 f5 Kg7 45 Rc1 Rd2 46 Qc4 Rxh2
[pos]8/3q1pk1/6p1/5P2/2Q1P3/6Pp/PP5r/2R2K2 w - - 0 47[/pos]
2) 38 Rc1 Qe2 39 Kh3 Qxe4 40 Qc4 Qf3 41 a4 Kh7 42 Re1 Rd2 43Qf4 Qxf4 44 gxf4 Rxb2
[pos]8/5p1k/6pp/8/P4P2/7K/1r3P1P/4R3 w - - 0 45[/pos]
This heavy piece endgame would have been fun to play in time trouble, especially if the championship had happened to depend on the outcome! So here is my suggestion for championship tie breaks - all the players involved play preset, computer-generated ‘precipice’ positions, with five minutes on the clock...