Tricks with Knights & Rooks

In an earlier post, I had a messy position from the Eickelman-Slade game in the club championship. Working my way through it I came to a position wherein White has to win with a Rook, g- & h-pawns versus a Knight & h-pawn. It turned out that winning the position wasn't too hard, but it put me in mind of the famous Em. Lasker-Ed. Lasker game from the New York, 1924 tournament. Looking up that position in my trusty copy of Fundamental Chess Endings, I came across the following study.

J. Moravec, La Strategie, 1913

White to move and win

This will eventually resolve itself into a R vs N endgame. If the knight is close to the king, the defending side can often draw the game, as in the Lasker-Lasker game above. However, sometimes the knight can be separated from the king, even when they're close together at the start. To see some endgame magic, check below the fold for the solution.

[Event "La Strategie"] [Site "?"] [Date "1913.??.??"] [Round "?"] [White "Moravec, J."] [Black "?"] [Result "1-0"] [SetUp "1"] [FEN "7K/6p1/8/7p/8/8/R7/6k1 w - - 0 1"] [PlyCount "37"] {Nunn Convention in effect. Unless otherwise noted, analysis & notation are from Muller & Lamprecht.} 1. Kh7 $1 h4 ({After} 1... g5 2. Kg6 g4 {White again refuses to take the pawn. This is a nice echo of the motif of avoiding the capture of a black pawn on move one!} 3. Kg5 g3 4. Kh4 g2 5. Kh3 $1 Kh1 6. Rxg2 $1 $18) 2. Kg6 $1 h3 3. Kg5 h2 4. Kg4 g5 (4... h1=Q 5. Kg3 $1 $18 {is the reason why the g-pawn had to be preserved. If White had instead played 1 Kxg7?, then Black would have 5 ... Qh8! -+.}) 5. Kg3 $1 h1=N+ 6. Kf3 $1 g4+ 7. Kxg4 $1 $18 {Here is one example to demostrate what happens if the knight can be separated from its king. Surprisingly, White wins even with Black to move because the knight is on a bad circuit.} Nf2+ 8. Kf3 Nd3 9. Ra4 Nc5 (9... Kh2 10. Rh4+ Kg1 11. Rd4 Nc5 (11... Nf2 12. Rd5 Nh3 (12... Kf1 13. Rd2 $18) 13. Ra5 Kh2 14. Rh5 $18) 12. Rd5 Ne6 13. Kg3 Kf1 14. Rf5+ Kg1 15. Re5 $18) 10. Ra1+ Kh2 11. Rd1 Ne6 12. Rd2+ Kg1 (12... Kh3 {is met by} 13. Rd6 $1 Ng5+ 14. Kf4 $1 Nf7 15. Rd7 {, winning the knight.}) 13. Kg3 Kf1 14. Rd5 {Taking away five of the knight's possible eight squares. It can't come back to the king.} Nc7 15. Re5 Na8 ({The movement of the knight from h1 to the opposite corner a8 is a nice feature of the Moravec study.} 15... Na6 {doesn't save the knight either:} 16. Kf3 $1 Kg1 {[If the knight moves to the b-file, 17 Rb5 wins immediately. If the knight moves to the c-file, 17 R(x)c5 would also win immediately. - TD]} 17. Rg5+ Kh2 18. Rg2+ Kh3 19. Rg6 $18) 16. Kf3 Kg1 17. Rg5+ Kh2 18. Rg2+ Kh3 ({ [ Or} 18... Kh1 19. Kf2 Nb6 20. Rg6 $18 {- TD]}) 19. Rg8 1-0

I found this study particularly striking, and I hope everyone else does, too.