7 steps to heaven. 7 testpositions of variable difficulty!

[Mark]

I stopped the engine after completion of 37 plies with still 3. Rxb2 as best, although the engine thought very long about 3. e7

[FEN "2b5/8/4P1p1/3p1pP1/p2k4/Bpp3K1/4R3/8 b - -"]

1... b2 2. Bxb2 cxb2 *

37/88 605:32 -1.41 3.Rxb2 Bxe6 4.Kf4 Kc5 5.Rc2+ Kd6
6.Rb2 Bd7 7.Rb8 d4 8.Rg8 Be6 9.Rxg6 a3 (30.086.449.037) 828

best move: Re2xb2 time: 620:35.094 min n/s: 828.093 nodes: 30.086.449.037

I've been running wac230 in IDeA for a long time now. I recently started re-analyzing it in IDeA with Houdini 3.

[d]2b5/8/4P1p1/3p1pP1/p2k4/6K1/1p2R3/8 w - -

After thousands of positions, IDeA has e7 as -0.33, so it does indeed seem to hold the draw. (Rxb2 is at -2.13)

[Mark]

...b2!! wins. So WAC 230 is correct after all.
That leads to one of the most interesting endgames i've seen. I hope i was not wrong on my analysis since it was made before 4 years with much weaker engines than today.
1...b2!! wins.
[d]2b5/8/4P1p1/3p1pP1/p2k4/Bpp3K1/4R3/8 b - - 0 1

Wrong!
After 1...b2 2.Bxb2 cxb2 3.e7! saves the draw
(but not 3.Rxb2?)

Are you sure that e7 is drawing?

With my Stockfish "Chessbase" Hybrid clone, just joking about that, it is admittedly a bit selective but it has almost standard Stockfish 2.3.2 eval I can go to depth 36 with it on an old Athlon in under 4 hours with a small hashtable:


[d]2b5/8/4P1p1/3p1pP1/p2k4/6K1/1p2R3/8 w - -

Engine: Stockfish 2.3.1 Hybrid, Chessbase Open Source Release Candidate 145 Athlon 2009 MHz, only 1 thread, 128 MB
by Tord Romstad, Marco Costalba and Joona Kiiski


28/63 5:20 -0.96 3.Rxb2 Bxe6 4.Kf4 Kc5 5.Ke5 Bd7
6.Rc2+ Kb4 7.Rb2+ Ka5 8.Ra2 Bc6
9.Ra1 d4 10.Kxd4 f4 11.Ra2 f3 12.Ra1 Bb5
13.Ke3 Be2 14.Kf2 Kb4 15.Rb1+ Kc5
16.Rc1+ (249.532.936) 778

29/69 8:44 -1.45 3.Rxb2 Bxe6 4.Kf4 Kc5 5.Ke5 Bd7
6.Rc2+ Kb4 7.Rb2+ Ka5 8.Ra2 Bc6
9.Ra1 f4 10.Kxf4 Bb5 11.Ke5 Bc4
12.Kf6 Bd3 13.Ra2 Kb4 14.Rb2+ Kc3
15.Rb8 a3 16.Ra8 (410.166.465) 781

30/69 16:25 -1.45 3.Rxb2 Bxe6 4.Kf4 Kc5 5.Ke5 Bd7
6.Rc2+ Kb4 7.Rb2+ Ka5 8.Ra2 Bc6
9.Ra1 f4 10.Kxf4 Bb5 11.Ke5 Bc4
12.Kf6 Bd3 13.Ra2 Kb4 14.Rb2+ Kc3
15.Rb8 a3 16.Ra8 (769.795.733) 781

31/69 25:19 -1.41 3.Rxb2 Bxe6 4.Kf4 Kc5 5.Ke5 Bd7
6.Rc2+ Kb4 7.Rb2+ Ka5 8.Ra2 Bc6
9.Ra1 f4 10.Kxf4 Bb5 11.Ke5 Bc4
12.Kf6 Bd3 13.Rd1 Be4 14.Rc1 a3
15.Rc8 Kb4 16.Rb8+ (1.196.675.880) 787

32/78 35:03 -1.69 3.Rxb2 Bxe6 4.Kf4 Kc5 5.Rc2+ Kd6
6.Rb2 Bd7 7.Rb8 d4 8.Rg8 Be6 9.Rxg6 a3
10.Rh6 a2 11.Rh1 Bd5 12.Ra1 Bb3
13.g6 Ke6 14.g7 Kf6 15.Rg1 Bd5
16.Ra1 (1.671.561.611) 794

33/78 40:06 -1.85 3.Rxb2 Bxe6 4.Kf4 Kc5 5.Rc2+ Kd6
6.Rb2 Bd7 7.Rb8 d4 8.Rg8 Be6 9.Rxg6 a3
10.Rh6 a2 11.Rh1 Bd5 12.Ra1 Bb3
13.g6 Ke6 14.g7 Kf6 15.Rg1 Bd5
16.Ra1 (1.919.255.143) 797

34/78 121:11 -1.73 3.Rxb2 Bxe6 4.Kf4 Kc5 5.Rc2+ Kd6
6.Rb2 Bd7 7.Rb8 d4 8.Rg8 Be6 9.Rxg6 a3
10.Rh6 a2 11.Rh1 Bd5 12.Ra1 Bb3
13.g6 Ke6 14.g7 Kf6 15.Rg1 Bd5
16.Ra1 (5.926.336.764) 815

35/78 190:56 -1.73 3.Rxb2 Bxe6 4.Kf4 Kc5 5.Rc2+ Kd6
6.Rb2 Bd7 7.Rb8 d4 8.Rg8 Be6 9.Rxg6 a3
10.Rh6 a2 11.Rh1 Bd5 12.Ra1 Bb3
13.g6 Ke6 14.g7 Kf6 15.Rg1 Bd5
16.Ra1 (9.533.567.332) 832

36/88 220:19 -1.45 3.Rxb2 Bxe6 4.Kf4 Kc5 5.Rc2+ Kd6
6.Rb2 Bd7 7.Rb8 d4 8.Rg8 Be6 9.Rxg6 a3
10.Rh6 a2 11.Rh1 Bd5 12.Ra1 Bb3
13.g6 Ke6 14.g7 Kf6 15.Rg1 Bd5
16.Ra1 (11.009.081.370) 832

The engine has e7 as second move always, but so far has rejected it each time. It is not totally clear this is a losing score for 3. Rxb2 though. Maybe an engine that uses endgame table bases and that can go deeper can see this better.

36/88 220:19 -1.45 3.Rxb2 Bxe6 4.Kf4 Kc5 5.Rc2+ Kd6
6.Rb2 Bd7 7.Rb8 d4 8.Rg8 Be6 9.Rxg6 a3
10.Rh6 a2 11.Rh1 Bd5 12.Ra1 Bb3
13.g6 Ke6 14.g7 Kf6 15.Rg1 Bd5
16.Ra1 (11.009.081.370) 832

After Stockfish's 7. Rb8, Kc5 is much better according to IDeA.

Kc5 is at -2.38 while d4 is at -.44.

[George Tsavdaris]

...b2!! wins. So WAC 230 is correct after all.
That leads to one of the most interesting endgames i've seen. I hope i was not wrong on my analysis since it was made before 4 years with much weaker engines than today.
1...b2!! wins.
2b5/8/4P1p1/3p1pP1/p2k4/Bpp3K1/4R3/8 b - - 0 1

Wrong!
After 1...b2 2.Bxb2 cxb2 3.e7! saves the draw
(but not 3.Rxb2?)

Are you sure that e7 is drawing?

With my Stockfish "Chessbase" Hybrid clone, just joking about that, it is admittedly a bit selective but it has almost standard Stockfish 2.3.2 eval I can go to depth 36 with it on an old Athlon in under 4 hours with a small hashtable:


[d]2b5/8/4P1p1/3p1pP1/p2k4/6K1/1p2R3/8 w - -

Engine: Stockfish 2.3.1 Hybrid, Chessbase Open Source Release Candidate 145 Athlon 2009 MHz, only 1 thread, 128 MB
by Tord Romstad, Marco Costalba and Joona Kiiski



36/88 220:19 -1.45 3.Rxb2 Bxe6 4.Kf4 Kc5 5.Rc2+ Kd6
6.Rb2 Bd7 7.Rb8 d4 8.Rg8 Be6 9.Rxg6 a3
10.Rh6 a2 11.Rh1 Bd5 12.Ra1 Bb3
13.g6 Ke6 14.g7 Kf6 15.Rg1 Bd5
16.Ra1 (11.009.081.370) 832

The engine has e7 as second move always, but so far has rejected it each time. It is not totally clear this is a losing score for 3. Rxb2 though. Maybe an engine that uses endgame table bases and that can go deeper can see this better.

Yes he is right and i have said that to the Rybka forum already(but i forgot to mention this here too): http://rybkaforum.net/cgi-bin/rybkaforu ... #pid463390

...b2 doesn't win. I had made a newer analysis about this before some years and i had found a refutation on ...b2. But(the moment i've made the topic here) i have forgotten about that. Sorry for the confusion. These are the results of being away from the computer Chess scene for so long, you forget many things.

Here is the best try for black:
1...b2 2.Bxb2 cxb2 3.e7 b1Q 4.e8Q Qg1+ 5.Kf3 Qf1+ 6.Kg3 f4+ 7.Kh2 Qh3+ 8.Kg1 Qg3+ 9.Kh1 Qf3+ 10.Kg1 Bd7 11.Qe5+ Kd3 12.Rf2
[d]8/3b4/6p1/3pQ1P1/p4p2/3k1q2/5R2/6K1 b - - 0 12

And now black has 2 main options to choose:
•12...Qg4+ 13.Kh2 d4 14.Rd2+! giving the rook! since after 14...Kxd2
[d]8/3b4/6p1/4Q1P1/p2p1pq1/8/3k3K/8 w - - 0 15

White has a "crazy queen" draw. E.g: 15.Qxd4 Ke1 16.Qc3+ Kf2 17.Qc5+ Kf3 18.Qc3+ Ke4 19.Qc4+ Ke5 20.Qc5+ Ke6 21.Qc6+ 1/2-1/2 as black can't capture the queen!
[d]8/3b4/2Q1k1p1/6P1/p4pq1/8/7K/8 b - - 0 21

•12...Qf4 13.Qb2 draw

►►►►►►►
About the line Stockfish analyses, it's a win for black, i.e 3.Rxb2 loses since:
After 1...b2 2.Bxb2 cxb2 3.Rxb2? instead of the drawing one 3.e7, black wins by 3...Bxe6.
[d]8/8/4b1p1/3p1pP1/p2k4/6K1/1R6/8 w - - 0 4

With many many complicated variations. For example:
4.Kf4 Kc5! 5.Rc2+ Kd6 6.Rb2 Bd7! 7.Rb8 d4 8.Rg8(8.Rd8 Kc6 9.Rg8 Kb5 10.Rxg6 a3 11.Rh6 a2 0-1) Be6 9.Rxg6(9.Ra8 Bb3 10.Ra6+ Kc5 11.Rxg6 a3 12.Ra6 a2 13.g6 Kb5 14.Ra3 Kb4 0-1 #40 for black ) a3 10.Rh6 a2 11.Rh1 Bb3 12.g6 Kc5 0-1

[George Tsavdaris]

So to make up for the wrong position(the one with the ...b2 wins , that doesn't win) i give here a difficult position, a highly anti-computer one.

White to play and draw!!
[d]1k6/3p4/1B6/4Pp1p/1p5R/1p4p1/pP3n2/K6n w - - 0 1

With the removal of the wrong WAC.230 ...b2 position and the addition of this, as i understand computers are only be capable of solving the first 2(the easy ones) of the 7. That's unfortunate and i hope to see someday a program to solve 2-3 of these 5 tough ones....

[George Tsavdaris]

Avoid 1.Bxc7? that loses as also all other moves that lose too and play 1.c6! that draws.
[d]8/p1pp3p/8/2P5/5B2/p2K4/7P/1k6 w - - 0 1

1.Bxh7! wins! Classic Bxh7 sacrifice, but here the win is more difficult than the classic ones.
[d]r1bq1rk1/1p1nbppp/p3p3/2ppP3/5B1P/2PBP3/PP1N1PP1/R2QK2R w KQ - 0 1

1.Bf8+! wins!! Easy one for computers i guess.
[d]6R1/8/2pB3k/2P4p/5p1q/5P2/4P1K1/8 w - - 0 1

Avoid 1.Kh7? that draws and play 1.Kh8! that wins! Easy too.
[d]6K1/6P1/3P4/7k/8/P7/1R2P3/3r2r1 w - - 0 1


For humans only! Don't use engines, it's pointless as it's too easy for them.
How you would proceed? White to play and win! Only one move wins.
[d]8/7K/P2P4/8/8/4k3/7B/r7 w - - 0 1


Continuation from before:
White wins with 1.Bg1+
And after 1...Rxg1 white has only one move that wins. Which one is it?
[d]8/7K/P2P4/8/8/4k3/8/6r1 w - - 0 2


For humans only again!
How you would proceed? White to play and win!
[d]6rk/1p3p1p/2p1p3/3NP2R/2PR1P2/2P3Pb/1bKBQ3/q6r w - - 0 1


Continuation from before:
White wins with 1.Rxh7+!!
And after 1...Kxh7 2.Qh5+ Kg7 how you would continue the attack?
[d]6r1/1p3pk1/2p1p3/3NP2Q/2PR1P2/2P3Pb/1bKB4/q6r w - - 0 3

[peter]

Sorry, George, but I don't see, how White can keep this.
Of course, if Black takes the white e-pawn and the white bishop too soon, the white rook can give serial check and must not be taken because of stalemate, but how can Black be forced to do so?
Don't get it

[George Tsavdaris]

[d]1k6/3p4/1B6/4Pp1p/1p5R/1p4p1/pP3n2/K6n w - - 0 1

Sorry, George, but I don't see, how White can keep this.
Of course, if Black takes the white e-pawn and the white bishop too soon, the white rook can give serial check and must not be taken because of stalemate, but how can Black be forced to do so?
Don't get it

White can't force to do that, because black can choose to do something else. But still white can force the draw! It's rather easy for humans if you think about it.

[peter]

Hi George!

I did try some variants and even if I admit that you're right about White's plan as for human thinking as told in my previous posting, I still think there could be the one or the other mistake in your winning variants.
As a matter of fact, I guess if this is a study, it could be cooked.

Would you please show your solution or tell about the origin of the position

[George Tsavdaris]

Hi George!

I did try some variants and even if I admit that you're right about White's plan as for human thinking as told in my previous posting, I still think there could be the one or the other mistake in your winning variants.
As a matter of fact, I guess if this is a study, it could be cooked.

Would you please show your solution or tell about the origin of the position

The only move that draws for white is 1.Rxh1!!
[d]1k6/3p4/1B6/4Pp1p/1p5R/1p4p1/pP3n2/K6n w - - 0 1

After 1.Rxh1 black can play:

•1...g2
[d]1k6/3p4/1B6/4Pp1p/1p6/1p6/pP3np1/K6R w - - 0 2
So then after 2.Rg1 Nd3 3.Bf2 Nxf2 4.Rxg2 it's a draw


•1...Nd3
[d]1k6/3p4/1B6/4Pp1p/1p6/1p1n2p1/pP6/K6R w - - 0 2
So then after 2.Rd1 Nxe5 3.Be3 it's a draw


•1...Nxh1 Main continuation for black.
[d]1k6/3p4/1B6/4Pp1p/1p6/1p4p1/pP6/K6n w - - 0 2

So then after 2.e6! dxe6(forced) 3.Bc7+! Kb7 4.Bxg3 it's a draw since the white bishop is immune of being captured. And white will keep his bishop moving from e1-h4 diagonal while black can't do nothing about it.
Final position after 4.Bxg3. Draw!
[d]8/1k6/4p3/5p1p/1p6/1p4B1/pP6/K6n b - - 0 4

--------------------
Note that 1.Bc7+? doesn't draw also, since black has 1...Kb7! that wins and not 1...Kxc7? that draws.
[d]1k6/2Bp4/8/4Pp1p/1p5R/1p4p1/pP3n2/K6n b - - 0 1

After 1.Bc7+? Kb7! now white can't draw and loses since 2.Rxh1 doesn't work anymore.

[lech]

With the removal of the wrong WAC.230 ...b2 position and the addition of this, as i understand computers are only be capable of solving the first 2(the easy ones) of the 7. That's unfortunate and i hope to see someday a program to solve 2-3 of these 5 tough ones....

My testing Sting SF 3 (bred on Stockfish 2.1.1 - modified null move search) gets it in 50 minutes (2 threads 1056M hash).
[d]1b6/3kB3/8/1KpPp2p/4Pp1P/5P2/8/8 w - - 0 1 bm Bf8;

Code: Select all

info depth 55 seldepth 63 multipv 1 score cp 177 nodes 1295351015 nps 1205485 ti
me 1074547 pv e7c5 b8c7 c5b6 c7b8 b5c4 b8d6 b6f2 d6c7 c4b4 d7d6 f2c5 d6d7 c5g1 c
7d6 b4c4 d6e7 g1f2 e7d6 f2c5 d6c7 d5d6 c7a5 c5b4 a5b6 c4d5 b6d4 b4d2 d4b2 d5c5 b
2d4 c5d5
info depth 56
info currmove e7c5 currmovenumber 1
info nodes 1693686623 nps 1191569 time 1421391
info currmove b5c5 currmovenumber 2
info currmove d5d6 currmovenumber 3
info nodes 1751065480 nps 1192722 time 1468125
info currmove b5c4 currmovenumber 4
info currmove e7d8 currmovenumber 5
info currmove e7g5 currmovenumber 6
info nodes 1752196588 nps 1192578 time 1469250
info currmove e7f6 currmovenumber 7
info nodes 1764634282 nps 1191854 time 1480578
info currmove b5b6 currmovenumber 8
info currmove e7d6 currmovenumber 9
info currmove e7f8 currmovenumber 10
info depth 56 seldepth 0 multipv 1 score cp 185 lowerbound nodes 3627347184 nps
1189172 time 3050313 pv e7f8 b8c7 b5c5 c7d8 f8g7 d8h4 g7e5 h4f2 c5c4 h5h4 c4d3 f
2e3 d3e2 h4h3 e2f1 d7d8 e5d6
info nodes 3627347184 nps 1189172 time 3050313
info currmove e7f8 currmovenumber 1