View Full Version : Spectacular endgame puzzle

C.Antolis

04-07-2009, 12:11 AM

i just wanted to share a spectacular puzzle i came across (many of u will probably have seen it before, but will be nice for people who havent):

It is white to move and win...

2b4k/8/5Pr1/5N2/8/8/8/K1B5 w - - 0 35

for those who want to see the solution (in white):

1. f7 Ra6+(forced)

2. Ba3!(decoying rook) Rxa3+

3. Kb2 Ra2+!

4. Kc1(Kc3 is a draw- nonstop checks coming, K cant go on d-file or up to the 7th rank) Ra1+

5. Kd2 Ra2+

6. Ke3 Ra3+

7. Kf4 Ra4+

8. Kg5 Rg4+

9. Kh6 Rg8

10. Ne7 Be6

11. fxg8=Q+ Bxg8

12. Ng6#!

Duff McKagan

07-07-2009, 12:39 AM

Nice puzzle Ceddy... good to see you practising your chess also!

black

07-07-2009, 12:53 AM

Is this from Secrets of Spectacular Chess? Does look familiar.

C.Antolis

07-07-2009, 07:19 AM

Is this from Secrets of Spectacular Chess? Does look familiar.

ummm... i dont know if it is in that book or not, but i read this off 'endgame magic'

michael.mcguirk

05-11-2009, 11:33 PM

Yeah it is in that book, one of the more awesome studies I admit :)

V. Korolkov

1st Prize, Lelo 1951

Here's the alternate lines for anyone interested in the reasoning.

If 1...Rf6? 2.Bb2 wins.

If 2.Kb2? Rf6

If 3...Rb3+? 4.Ka2

If 4.Kxa2 Be6+ =

If 4.Kc3? Rc2+ 5.Kb4 (5.Kd4 Rd2+ followed by 6...Rd8) 5...Rb2+ 6.Kc5 Rc2+ 7.Kb6 Rb2+ and White can't make progress as 8.Kc7 allows 8...Rb7+ and 9...Rxf7.

If 4... Rc2+ 5.Kd1 and Black's out of options.

If 9.Kxg4? Bxf5+ 10.Kxf5 Kg7 11.Ke6 Kf8 12.Kf6 =

If 9.Kf6 or 9.Kh5, both allow 9...Rg8 =

If 9...Rg6+ 10.Kxf6 Bxf5+ 11.Kf6 wins.

If 10...Rf8 11.Nf6 mate.

If 10...Rg1 11.f8=Q+

If 11. Ng6+ Rxg6 12.Kxg6 Bxf7+

Desmond

06-11-2009, 08:57 AM

Just putting the mainline into the viewer...

35. f7 Ra6+{forced} 36. Ba3! {decoying rook} Rxa3+ 37. Kb2 Ra2+! 38. Kc1 {38.Kc3 is a draw- nonstop checks coming, K cant go on d-file or up to the 7th rank} Ra1+ 39. Kd2 Ra2+ 40. Ke3 Ra3+ 41. Kf4 Ra4+ 42. Kg5 Rg4+ 43. Kh6 Rg8 44. Ne7 Be6 45. fxg8=Q+ Bxg8 46. Ng6#!

Czentovic

15-01-2011, 05:08 PM

i just wanted to share a spectacular puzzle i came across (many of u will probably have seen it before, but will be nice for people who havent):

It is white to move and win...

2b4k/8/5Pr1/5N2/8/8/8/K1B5 w - - 0 35

for those who want to see the solution (in white):

1. f7 Ra6+(forced)

2. Ba3!(decoying rook) Rxa3+

3. Kb2 Ra2+!

4. Kc1(Kc3 is a draw- nonstop checks coming, K cant go on d-file or up to the 7th rank) Ra1+

5. Kd2 Ra2+

6. Ke3 Ra3+

7. Kf4 Ra4+

8. Kg5 Rg4+

9. Kh6 Rg8

10. Ne7 Be6

11. fxg8=Q+ Bxg8

12. Ng6#!

1...Ra6 is not forced, and is actually incorrect! As you attempt to prove, Ra6+ leads to a quick checkmate by White (allegedly in 12 moves), so Black will choose to play 1. Rg8 instead, avoiding a short mate, and forcing White to play 2. fxg8Q+ Kxg8 and now White has a forced mate in 33 starting 3. Ne7+, for a total of 37 moves, which gives Black vastly better chances of getting a draw or winning on time than your 12-move mate above. Actually, that forced mate-in-12 is also wrong. If we start 1. f7 Ra6+ 2. Ba3 Rxa3+ we can query the online Nalimov database at http://www.k4it.de/index.php?topic=egtb&lang=en to find that White has a forced win in an extra 18 moves, making a total move sequence of 20 moves, not 12. The optimal move sequence starts 3. Kb2 Ra2+ 4. Kc1 Rc2 (not Ra2?) 5. Kd1 Rc7 6. f8Q+ Kh7 etc. and White eventually mates on move 20.

Black's best chance is to hope that White has forgotten how to play the Bishop-and-Knight mate correctly.

Capablanca-Fan

18-01-2011, 04:14 AM

But this is not a fair comment. A study is supposed to illustrate a clever idea. It's an accepted rule that any variation leading to a book win can be ended right there, regardless of how many moves it takes.

Tony Dowden

19-01-2011, 07:02 AM

But this is not a fair comment. A study is supposed to illustrate a clever idea. It's an accepted rule that any variation leading to a book win can be ended right there, regardless of how many moves it takes.

Agreed. It is easy to see that 1...Rg8 loses trivially to an eventual K+B+N checkmate whereas 1...Ra6+ is the critical line because it is the only sensible way to attempt to avoid losing immediately.

Rincewind

19-01-2011, 09:23 AM

I agree with Jono and Tony. However it is clear that Czentovic is thinking of practical considerations of winning on time or via 50-move draw against an opposition over the board. However, while both rook moves lose even from practical considerations going into book loss is not good except against weak opposition. A strong opponent can bang out a BN endgame pretty quickly because it is something they have practiced. The line in the study certainly requires more thought and is not likely to be something they have seen before, certainly prior to the composition of this study. So even considering practicalities you can't just count the move to mate, you need to consider how many problems you can give your opponent that they have to solve over the board and how easily they may go wrong.

Igor_Goldenberg

19-01-2011, 03:18 PM

I would definitely try 1...Ra6, more practical chances.

Once I deliberately walked in a forced checkmate in two (which wasn't easy to spot, though) as alternative was a clear loss (rook down followed by massive liquidation into exchange).

paulo101977

23-06-2020, 01:16 AM

i just wanted to share a spectacular puzzle i came across (many of u will probably have seen it before, but will be nice for people who havent):

It is white to move and win...

2b4k/8/5Pr1/5N2/8/8/8/K1B5 w - - 0 35

for those who want to see the solution (in white):

1. f7 Ra6+(forced)

2. Ba3!(decoying rook) Rxa3+

3. Kb2 Ra2+!

4. Kc1(Kc3 is a draw- nonstop checks coming, K cant go on d-file or up to the 7th rank) Ra1+

5. Kd2 Ra2+

6. Ke3 Ra3+

7. Kf4 Ra4+

8. Kg5 Rg4+

9. Kh6 Rg8

10. Ne7 Be6

11. fxg8=Q+ Bxg8

12. Ng6#!

I believe that Kne7, because it prevents the capture of the pawn (the rook would be embedded in the same line as the bishop) and supports its advance.

Powered by vBulletin® Version 4.2.5 Copyright © 2021 vBulletin Solutions Inc. All rights reserved.