No.21 in Nunn's Endgame challenge is busted

[Alan Tomalty]

I busted No.21 in Nunn's Endgame Challenge book (of course with the aid of Houdini Pro 2). In that book Nunn gives the best 250 studies of all time in his opinion. After this he will have to add another one. It opens up the question of how many others are busted since Nunn was using Deep Fritz and the year of publication of Nunn's book was 2002. Computers have come a long way in 10 years.
[Event "Revista de Sah 1965"]
[Site "?"]
[Date "2012.11.23"]
[Round "?"]
[White "Nunn Endgame Challenge "]
[Black "No. 21 by P.Joitsa"]
[Result "0-1"]
[Annotator "Tomalty,Alan"]
[SetUp "1"]
[FEN "5k2/8/8/8/7P/8/4Np2/1K3n2 w - - 0 0"]
[PlyCount "9"]

{Diagram [#]} 1. Ka2 Kf7 2. Kb2 Kf6 3. Ka3 Ne3 (3... Ke5 {This move also busts
the study} 4. h5 Ne3 5. Ng3 Nf5 {See Diagram 2}) 4. Ng3 {This study is busted.
Nunn in his book Endgame Challenge gives only 4....Kg6 or 4......Nf5 which
both draw for white} Ke5 $3 {Diagram [#] Now on any white knight move the
knight is either lost or the black pawn queens} 5. Kb3 $4 {mate in 11 for black
} (5. Kb2 $4 {Mate in 10 for black}) (5. Kb4 $4 {Mate in 12 for black}) (5. Ka2
$4 {mate in 10 for black}) (5. Ka4 $4 {mate in 12 for black}) (5. h5 $4 Nf5 $19
{Diagram [#]} 6. Nf1 (6. h6 {mate in 24 for black} Nxh6) (6. Ne4 $4 Kxe4) 6...
Ke4 (6... Kd4 {also wins for black}) 7. Kb3 (7. Kb2 Kd3 $8 8. Kc1 $4 {mate in
29 for black} (8. Ka1 $4 {mate in 25 for black} Ke2) (8. Nh2 Ke2 9. Kc3 {
Diagram [#]})) (7. Kb4 Kd3 8. Kc5 Ke2 9. Nh2 Nh6 {Diagram [#]} 10. Kd4 Ng4 $19)
7... Kd3 $8 8. Nh2 (8. Kb4 Ke2 9. Nh2 Nh6 10. Kc5 Ng4 $19 {Diagram [#]}) 8...
Ke2 $19 {Diagram [#] after any white king move Nh6 wins for black} 9. Kc4 $4 (
9. Kc3 $4 {see diagram 4} Nh6 $19 10. Kc4 $4 {mate in 33} (10. Kd4 $4 Ng4)
10... Ng4) (9. Kb4 $4 Nh6 10. Kc5 Ng4) (9. Kc2 Nh6 {mate in 26 for black}) 9...
Ne3+ 10. Kd4 $4 (10. Kc5 $4 Ng4) 10... Ng4 11. Nxg4 $4 {mate in 21 for black})
0-1

[Michael Yip]

Re: No.21 in Nunn's Endgame challenge is busted

http://budapestchesnews.blogspot.hu/

Analysis is blogged. Full credit to Chesstalk and Al Tomalty.

[Alan Tomalty]

Re: No.21 in Nunn's Endgame challenge is busted

Thanks Michael for doing your own analysis to check this. It isn't easy to definitely categorically state the evaluation in most chess positions and even with Houdini it takes a lot of work going up and down the variation tree. However, eventually if there isn't too many pieces on the board one can get all the lines out to a mate if one doggedly traverses the variation tree. An interesting epiphany I had today was something I am sure most GMs know intuitively. Noone has stated this in print. The critical positions (many times) in a chess endgame are the ones that have many transpositions leading back to them. These are the positions that can be forced and are the most optimal ones and are critical to understanding the piece configuration of that endgame. Sometimes there is only 1 critical position, so that makes it easier to play knowing what your 1 target is. Sometimes it is a critical position to avoid. Ex: in the Q vs 2B eg the position where the BK is on b7 and bishops on b6 and c6 or the symmetrical equivalent is to be avoided if the white queen wants to win.

[Egidijus Zeromskis]

Re: No.21 in Nunn's Endgame challenge is busted

Per blog diagram it is a six piece problem. You can check tablebase directly. No busting needed.
Later i could give a link. With a tablet it is not easy to type.

[Laurentiu Grigorescu]

Re: No.21 in Nunn's Endgame challenge is busted

For those of us who do not have Nunn's book your explanation is hard to follow. What was the study claiming? Draw? White wins? Black wins? What is your conclusion?
Michael Yip provided the useful diagram, which helps.

Thank you!

[Alan Tomalty]

Re: No.21 in Nunn's Endgame challenge is busted

All the information you request is in the FEN line. For those who cannot read FEN please click on the link in Michael Yip's reply. The study claims it is white to move and draw but black wins in all variations. As Mr. Zeromskis correctly points out the bust is confirmed by the 6 man tablebase.
http://www.k4it.de/index.php?topic=egtb&lang=en

[Egidijus Zeromskis]

Re: No.21 in Nunn's Endgame challenge is busted

It is possible to use FEN here too

[Louis Morin]

Re : Re: No.21 in Nunn's Endgame challenge is busted

Per blog diagram it is a six piece problem. You can check tablebase directly. No busting needed.
Later i could give a link. With a tablet it is not easy to type.

This is a link:

http://k4it.de/index.php?topic=egtb&lang=en

With best play, White gets mated in 33 moves.