![\"qb1\"](\"https:\/\/danamackenzie.com\/blog\/wp-content\/uploads\/2014\/09\/qb1.jpg\")
<\/a>FEN: Q7\/P6p\/2K3p1\/8\/8\/3kb3\/5p2\/8 b – – 0 88<\/p>\nPosition after 88. Kc6. Black to move.<\/em><\/p>\n Here I was expecting Black to play 88. … f1Q 89. Qd8+ Kc2 90. a1Q Qc4+, which I thought was likely to lead to a draw. But I guess the computer saw some way for me to escape the checks (maybe 91. Kd6, planning to run to the kingside?). Instead it played 88. … Bxa7!?!?<\/strong><\/p>\n When I saw this move, I thought, “Oh damn, the computer hoodwinked me again.” Of course I saw that 89. Qxa7 f1Q?? 90. Qa6+ would lose for Black. But the real question is what happens after 89. Qxa7 Ke2<\/strong> (diagram 2).<\/p>\n Position after 89. … Ke2 (analysis). White to move.<\/em><\/p>\n Here, unfortunately, my analysis was totally superficial. I thought, “Queen against an f-pawn on the seventh rank. Draw.”<\/p>\n Which it would be, if Black didn’t have the g- and h-pawns. But he does! And that changes everything. His standard drawing resource in such endgames is to move the king to h1 (if necessary) and set up a stalemate after Qxf2. But because he has other pawns on the board, it won’t be stalemate here!<\/p>\n This is why it’s so important in chess not to just memorize formulas like “queen against an f-pawn on the seventh rank is a draw.” You’ve got to think about why<\/em> it’s a draw.<\/p>\n In fact, it’s pretty easy to work out that White wins in diagram 2 after 90. Qa2+ Kf3<\/strong> (I’ll let you work out 90. … Ke1, but in my opinion Black’s best defense is to head for g2) 91. Qc4!<\/strong> (Shredder’s nice move. If Black lets the queen get to f1, White is definitely winning.) Kg2 92. Qg4+<\/strong>. And here it is. If Black steps in front of the pawn with … Kf1, White can start moving his king closer. If Black goes to the h-file with 92. … Kh2 93. Qf3 Kg1 94. Qg3+<\/strong>, we get the classic position where 94. … Kh1 95. Qxf2<\/strong> would be stalemate. Except here it’s not, and White wins.<\/p>\n This morning I started wondering how far advanced Black’s pawns would have to be in order to make the game a draw. That’s the sort of question the Nalimov tablebase (which gives perfect play for both sides in positions with 6 or fewer pieces) is meant for.<\/p>\n I was quite surprised at the answer. Black’s pawns don’t have to be far advanced at all. In fact, if Black’s pawn were at h5 instead of h7, it would be a draw. (Maybe that’s not such a big surprise, because at h5 the pawn takes away the key move 92. Qg4+ in the above line.) Or if Black’s pawn were at g5 instead of g6, it would be a draw. Or — get this — if the g-pawn were off the board<\/em> instead of at g6, it would be a draw!<\/p>\n I know you won’t believe this, so here’s a sample line to give you the idea.\u00a0 (Obviously I can’t show you all the variations.) Take the g-pawn off the board in diagram 2. Then after 1. Qa2+ Kf3 2. Qc4 Kg2! 3. Qg4+ Kf1! 4. Kd5 h5!! Black draws. If White takes the pawn it’s the regular book draw. If 5. Qf3 h4! 6. Ke4 Kg1 7. Qg4+ Kh2 White is reluctantly forced to take the h-pawn: 8. Qxh4+ Kg1, and we’re back in the book draw. Notice that if Black still had a pawn at g6, it would not be a draw! That’s why removing Black’s g-pawn made such a difference.<\/p>\n Pretty amazing stuff. But that’s not the end of this fascinating endgame!<\/p>\n Because I was determined not to get hoodwinked, I declined the bishop on a7, which meant that I did<\/em> get hoodwinked. The game went on with some mistakes for both sides until we got to this position.<\/p>\n Position after 108. … Bd6. White to move.<\/em><\/p>\n You will look in vain for an endgame like this in Reuben Fine’s Basic Chess Endings<\/em>. At this point I thought it was a dead draw, because I couldn’t imagine how White could make progress. How can I ever chase the bishop away from the defense of the g3 pawn? How can I possibly use both my king and my queen for a mating attack when one or the other has to constantly watch out for the advance of the g- or f-pawns?<\/p>\n I was surprised to see, though, that the computer rated the position at 2.2 pawns in my favor — a winning advantage. What the heck, I thought. Let’s forget about playing the game, and let’s see why the computer thinks that White can win. The next eight moves were basically Shredder versus Shredder.<\/p>\n<\/a>FEN: 8\/Q6p\/2K3p1\/8\/8\/8\/4kp2\/8 w – – 0 90<\/p>\n<\/a>FEN: 8\/8\/3b4\/1k6\/8\/6pQ\/5pK1\/8 w – – 0 109<\/p>\n