|Sep-11-04|| ||mjk: 66.♔d6 and 68.♔e8 were not best defense, but should still lose. White strays from the fastest win as 69...♘e5 mates in 30. Finally 76...♘d5? 77.b4! lets White escape.
<-- Nalimov Tablebase Server>|
Maybe Lilienthal studied after this game, but he played out a draw in Smyslov vs Lilienthal, 1940
|Jul-30-07|| ||YouRang: K+2N vs. K+P is a notoriously difficult game to win. I suspect Lilienthal knew he wasn't going to win that way, so he hoped to set up a blunder trap (diagram after 81...Kd7 - white to move):
click for larger view
Black is hoping for 82. b7??? Ne6#, but white didn't fall for it.
As <mjk> pointed out, black had a win. In fact, at move 75 he had mate-in-28 with 75...Kf6, but he played 75. Nf3? (mate-in-40), then gave up the win completely with 76...Nd5.
The winning technique in such a game is:
(1) Block the pawn with one knight.
(2) Use the other knight and king to force the opposing king into a corner. (The further advanced the pawn, the harder it is to find a corner and the right tempo that mates.) This is very difficult! The king and knight must do a very long and strange dance to work this out.
(3) At the right time (when the opposing king is sufficiently cornered), the knignt that was blocking the pawn approaches the cornered king with mating attack. This isn't quite as difficult as step 2, but it's riskier! If you time it wrong, the pawn may promote -- and then you lose!
|Feb-08-09|| ||WhiteRook48: 82. Kg7!! is one of the greatest drawing moves|
|Feb-07-12|| ||FSR: <YouRang: ... If you time it wrong, the pawn may promote -- and then you lose!>|
Well, maybe. King and two well-placed knights can hold the draw against a queen. Weirdly, there seem to be no examples of that ending in the database, although there are nine endings with queen versus bishop and knight (the worst set of two minor pieces with which to oppose the queen). Endgame Explorer: Q vs BN
|Feb-04-13|| ||YouRang: <Well, maybe. King and two well-placed knights can hold the draw against a queen. Weirdly, there seem to be no examples of that ending in the database>|
That is rather weird that there's no games ending with Q vs 2N.
And yes, reviewing the tablebase shows that there are lots of positions where the 2N can hold the draw. However, regardless of the forced outcome, it's not an easy task for either side. For example, the position below:
click for larger view
As white, you can force mate in 43 moves -- but only if you find Qe2!
|Feb-15-14|| ||FSR: <YouRang> Queen against two knights is apparently quite rare. CG.com's Endgame Explorer specifies that the pieces must remain on the board for three moves (six ply) for the endgame to be counted. This avoids scooping up, for example, (1) the conclusion of NN v. P endings where the pawn queens and then the other side immediately mates, or (2) the pawn queens and a knight captures it, or (3) more complicated endings where the queen sacrifices itself to produce a drawn NN ending (for example, I saw a game with Q v. RNN where the side with the queen sacrificed the queen with QxR+ to avoid the risk of loss). I found only five games in Mega Database 2013 with Q v. NN that meet the 3 moves/6 ply criterion. As I recall, there are about 5.5 million games in that database, so it's literally a one-in-a-million type phenomenon.|