Further Analysis of Leveikina-McCartney Rook & Pawn Ending

Click on Move 36 to view critical position

Yes, Jenny started out the endgame wrong by playing 36.g4? but it's still not clear the endgame is lost after that. Jenny also made the following missteps which brought about her demise: 56 Ra5?  Better is 56.Ra6+ or 56.f3 slowing Black down. 57.Rb5?    Better is 57.Ra6+ and continuing to check along the ranks.The best Black can do is move his king toward White's rook which will then go to e3.  And, again, it's not clear Black can win that position 63 Kh3??  The final blunder/mistake which leads to easy victory for Black. Better is 63.Ra4 and if ...63.Re6 White has 64.f3! and again Black's path to victory is unclear. ...64.f5? leads to equality after trading the two sets of paws as trading rooks leads to the simple drawn K+pawn endgame with pawn.  And on ...64.Ke3 White once again checks along the ranks and the best Black can do is circle around to the kingside pawns (towards g6) but that position appears drawn after White plays fxe4.  ...63. Rd2 appears to be stronger, but again I don't see how Black wins.  64...Re2 can be answered by 65.Rb6 leading to ...65.Kxg4, 66. Kf1,Rc2 67.Rxf6 and again a draw appears likely.

NM Pat Sciacca

Sciacca may be right about 55 and 56, but his analysis is faulty on move 63. He says 63.Ra4 is a good move for White, but after what he indicates as the last ditch effort for Black, 63…Rd2, his follow up he has isn’t right. After 63.Ra4, Black plays 63…Rd2! (NOT 63…Re6?), but then his analysis is wrong for Black at move 65, where he claims to take the g-pawn, which is incorrect. After 63.Ra4 Rd2! 64.Rb4 If White plays 64...Ra6, hitting the f-pawn, then Black plays the immediate 64...Kxg4! and now:

A)     65.Rxf6 e3!! Intending 66…Rxf2 with a winning K+P vs K, Black’s King is in front of the g-pawn with the WK on f2, so Kh3 wins after the trade down.

B)      65.Kf1 (going based on his other line, but here, the Rook is on d2, NOT e2, and so 65…f5 wins with an easy 3-on-1.

After 64.Rb4 Re2 65.Ra6, Black doesn’t play 65…Kxg4, but instead, 65…e3!!.  Now both lines win for Black:

A)     66.Rxf6+ is easy for Black, 66…Kxg4 and 67…Rxf2 -+ (If it’s check, and 68.Rxf2 is forced, then 68…exf2 69.Kxf2 Kh3 is winning)

B)      66.Rb4+ Ke5 and White can continue to check if he wants, but Black will hide on g6 eventually.   After 67.Rb5+ Ke6 68.Rf5 (chasing the King is worse)  Rxf2+ 69.Rxf2 exf2 70.Kxf2 Kd5 71.Kf3 (71.Ke3 Ke5 Black has opposition) Kd4 -+

And so I do believe that after 62…Kf4, it truly is, Mission Accomplished!

     Patrick McCartney

May I respond to McCartney's analysis (correction?)? 

McCartney's proposed improvement ...65.e3 is met by 66.Rb4+, Ke5 67.Kf3 (instead of 68.Rb5+?, Kd4 and simply moving the king toward the rook, winning easily, the plan with ...68.Ke6 and hiding the king at g6 is not so effective; thus 69.Rb6+, Kf7, 70. Rb7+, Kg6 71.Re7! and after ...71.Rxf2+ 72. Kg3, Re2 73.Kf3, Re1 74.Rxe3! and the king and pawn ending is drawn) ,Rxf2+ 68.Kxe3 and once again we have a theoretically drawn position.

Sorry Patrick, but there's no win after ...62.Kf4, though you did accomplish your mission.

You really ought to check your analysis with a chess engine. We need to be accurate in our analysis of a position. Otherwise, we are miscommunicating the truth.


While the position after 62…Kf4 may be a draw (though I’m still not 100% convinced), an engine isn’t the answer either as engines are horrible at endgames.
For instance, strong engines like Rybka 4 recommended lines that lead specifically to Philidor’s draw (i.e. WKg1, WRb3, BKe4, BRa2, BPg5) and Rybka would say it was -1.6 (-+).
So while human analysis can be faulty (find me a chess book with no error cover to cover…doubt you can), chess engines aren’t the end all be all solution either.  Engines are horrible at openings and endgames.  They only really work when you are trying to find that magical move in the middle game that would have won it for you.
Patrick McCartney

Chess engines are best used to analyze positions and games and are thus very useful in any phase of the game. They can verify our analysis or show where it is faulty and would prevent us from making extreme statements about a move ("...65.e3!!" i.e., "...65.e3 is an outstanding move." ....65.e3 is merely another try which is easily dealt with by 67.Kf3).  Though their evaluations can be faulty in endgames due to its inability to see beyond a certain number of moves, one can usually get a sense when it isn't making progress, and it will change its evaluation after several moves.  I had a sense that the position after ...62.Kf4 wasn't won, or at least as I claimed, not clearly so, and Fritz 8 verified that for me even though it evaluated the position as having a winning advantage for Black.