Test position What do you need to find Rxh2

[JAMerolle]

1k1r1b1r/p1p1q3/2p1p3/4Ppp1/4nP2/P2BQ3/1PP3PB/2KR3R b - - 0 1


I was wondering which engines and or computer set up do you need to find Rxh2. Thanks Joe

[zullil]

[d]1k1r1b1r/p1p1q3/2p1p3/4Ppp1/4nP2/P2BQ3/1PP3PB/2KR3R b - - 0 1


I was wondering which engines and or computer set up do you need to find Rxh2. Thanks Joe

[yanquis1972]

if the move is not objectively best, engines will not display it on their primary pv...i am not sure Rxh2 and the amazing combination that seems to follow is best with precise play from white. my engine analysis gives -1.50 for Qc5 and -.14 for Rxh2 after 5 minutes in multi-pv mode. the evaluation actually decreased somewhat for Rxh2 over time.

[Mike S.]

[d]1k1r1b1r/p1p1q3/2p1p3/4Ppp1/4nP2/P2BQ3/1PP3PB/2KR3R b - - 0 1

I found this in my database, but it may be an improper test position.. Despite the specatcular variation, Rybka 232a thinks that 1...Qc5, simply threatening a fork with Nf2, is stronger:

Analysis by Rybka 2.3.2a 32-bit (d19 after some backwards analysis):

1. -/+ (-1.40): 1...Qc5 2.Qxc5 Bxc5 3.Bxe4 Be3+ 4.Kb1 Rxd1+ 5.Rxd1 fxe4 6.Bg1 Bxg1 7.Rxg1 gxf4 8.Kc1
2. = (-0.22): 1...Rxh2 2.Rxh2 Qxa3 3.bxa3 Bxa3+ 4.Kb1 Nc3+ 5.Ka1 Bb2+ 6.Kxb2 Nxd1+ 7.Kb3 Nxe3 8.fxg5

Probably Rybka and I are missing something?

The position after 7...Nxe3 in the 2nd variation:

105: Andrejew - Doluchanow, 3/93PC-63 MR-Test2 1935
[d]1k1r4/p1p5/2p1p3/4Ppp1/5P2/3Bn3/2P3PR/2K5 w - - 0 1

Analysis by Rybka 2.3.2a 32-bit :

8.fxg5 Kb7 9.Rh6 Re8 10.Be2 Nxg2 11.Bh5 Rg8 12.Rxe6 Nf4 13.Rh6 Rxg5 14.Be8 c5
=/+ (-0.60) Depth: 5 00:00:00
=/+ (-0.56) Depth: 22 00:02:27 9543kN

Analysis by Komodo32 1.0 JA:

8.fxg5 Kb7 9.Rh6 Re8 10.Be2 Nxg2 11.Bh5 Re7 12.Rf6 Rg7 13.Rxe6 Nf4 14.Rh6 Rxg5 15.Be8 c5 16.Kb2 Kc8 17.Rh8 Kd8 18.Bb5+ Ke7 19.Bc4 Ng6 20.Rh7+ Kd8 21.e6 Ne7 22.Be2 c6 23.Rf7 Rg2
-/+ (-1.17) Depth: 1 00:00:00
=/+ (-0.62) Depth: 22 00:01:17 45523kN

Not very dramatical.

(After 8.fxg5, 8...Rg8? would be met by 9.Ba6, and Black's rook has to guard the 8th rank.)

[zullil]

1k1r1b1r/p1p1q3/2p1p3/4Ppp1/4nP2/P2BQ3/1PP3PB/2KR3R b - - 0 1


I was wondering which engines and or computer set up do you need to find Rxh2. Thanks Joe

Here's what Stockfish-1.6.2 (8 threads) says at depth 28:

Code: Select all

info depth 28 score cp 266 time 901674 nodes 7519633306 nps 8339636 pv e7c5 e3c5 f8c5 d3e4 c5e3 c1b1 d8d1 h1d1 f5e4 h2g1 e3g1 d1g1 g5f4 b1c1 h8g8 g1f1 g8g4 c1d1 b8b7 d1e1 b7b6 e1f2 b6c5 f1d1 g4g5 b2b4 c5c4 d1d7 g5g3 d7c7 g3a3 c7c6 c4d4 c6e6 a3c3 e6e8 c3c2 f2f1 d4d5 e5e6 e4e3 e6e7 d5d6 b4b5 c2f2 f1e1 f2g2 e8f8 d6e7 f8f4 e7d6 f4f7 

I'll let it go some more, just for fun.

[JAMerolle]

the position is from the marathon test suit in F12 position 10 which claims the best move is Rook takes bishop. I was wondering if any of the I7s find this.

My dual process does not find it with any of the engines

After studying the position it looks like that bishop is very threatening which makes me believe it has to go.

[yanquis1972]

it's not really a question of processing power i don't think. analyze both Qc5 and RxB lines with your computer forwards and backwards, in multi-pv if you wish. it's certainly possible the engines are just wrong, but iwas getting growing returns for Qc5 and diminishing for Rxh2; i don't think Rxh2 is bad but i am not convinced it is best. it is an amazing combination however.

the only thing i can guess is that there is a refutation to Qc5 that the computers are somehow missing, which would make RxB best.

edit -- the moves in both variations all seem to be more or less forced, to me. i think this is a position where computers should be top notch & are most probably correct.

[metax]

the position is from the marathon test suit in F12 position 10 which claims the best move is Rook takes bishop. I was wondering if any of the I7s find this.

My dual process does not find it with any of the engines

After studying the position it looks like that bishop is very threatening which makes me believe it has to go.

No. It's rather that the rook H1 is diverted from the first row so that the amazing ...Qxa3!! is possible in the next move.

[Mike S.]

Engines don't choose Rxh2 because Qc5 is the better move (see my other posting). If you switch to multi-pv where available, you will probably find Rxh2 as the second best move, quickly.

Andrejew resigned to soon. According to strong engines, the last position of the game after NxQe3 is less than +1 for Black. It's a very nice combo but actually it gains less than Qc5.

[zullil]

1k1r1b1r/p1p1q3/2p1p3/4Ppp1/4nP2/P2BQ3/1PP3PB/2KR3R b - - 0 1


I was wondering which engines and or computer set up do you need to find Rxh2. Thanks Joe

Here's what Stockfish-1.6.2 (8 threads) says at depth 28:

Code: Select all

info depth 28 score cp 266 time 901674 nodes 7519633306 nps 8339636 pv e7c5 e3c5 f8c5 d3e4 c5e3 c1b1 d8d1 h1d1 f5e4 h2g1 e3g1 d1g1 g5f4 b1c1 h8g8 g1f1 g8g4 c1d1 b8b7 d1e1 b7b6 e1f2 b6c5 f1d1 g4g5 b2b4 c5c4 d1d7 g5g3 d7c7 g3a3 c7c6 c4d4 c6e6 a3c3 e6e8 c3c2 f2f1 d4d5 e5e6 e4e3 e6e7 d5d6 b4b5 c2f2 f1e1 f2g2 e8f8 d6e7 f8f4 e7d6 f4f7 

I'll let it go some more, just for fun.

Code: Select all

28     +2.67   15:01    7519M Qc5 Qxc5 Bxc5 Bxe4 Be3+ Kb1 Rxd1+ Rxd1 fxe4 Bg1 
                              Bxg1 Rxg1 gxf4 Kc1 Rg8 Rf1 Rg4 Kd1 Kb7 Ke1 Kb6 
                              Kf2 Kc5 Rd1 Rg5 b4+ Kc4 Rd7 Rg3 Rxc7 Rxa3 Rxc6+ 
                              Kd4 Rxe6 Rc3 Re8 Rxc2+ Kf1 Kd5 e6 e3 e7 Kd6 b5 
                              Rf2+ Ke1 Rxg2 Rf8 Kxe7 Rxf4 Kd6 Rf7 
29     +2.71   29:26   14698M Qc5 Qxc5 Bxc5 Bxe4 Be3+ Kb1 Rxd1+ Rxd1 fxe4 Bg1 
                              Bxg1 Rxg1 gxf4 Kc1 Rg8 Rf1 Rg4 Kd1 Kb7 Ke1 Kb6 
                              Kf2 Kc5 Rd1 Rg5 b4+ Kc4 Rd7 Rg3 Rxc7 Rxa3 Rxc6+ 
                              Kd4 Rxe6 Rc3 Re8 Rxc2+ Kf1 Kd5 e6 e3 e7 Rf2+ Ke1 
                              Ke6 g4 Kd7 Ra8 Kxe7 Rxa7+ Kd6 Ra6+ Kd5 Rf6 Kc4 g5 
                              Kxb4 
30  >  +2.91   57:15   28586M Qc5 Qf3 Nf2 b4 g4 Ba6 Nxd1 Rxd1 Rxd1+ 
31     +2.91 2:32:56   76334M Qc5 Qxc5 Bxc5 Bxe4 Be3+ Kb1 Rxd1+ Rxd1 fxe4 Bg1 
                              Bxg1 Rxg1 gxf4 Kc1 Rg8 Kd2 Kb7 Ke2 Kb6 Kf2 Kc5 
                              Rd1 Kc4 a4 Rh8 a5 Rh2 a6 c5 Kg1 Rh5 Rf1 Rf5 g3 f3 
                              Kf2 Kd4 c3+ Kxe5 Rh1 Rf8 Rh5+ Kd6 Rh1 Kd5 Ke3 c6 
                              g4 Rg8 Rd1+ Ke5 Rd7 Rb8 
32     +2.99 5:29:39  164195M Qc5 Qxc5 Bxc5 Bxe4 Be3+ Kb1 Rxd1+ Rxd1 fxe4 Bg1 
                              Bxg1 Rxg1 gxf4 Kc1 Rg8 Kd2 Kb7 Ke2 Kb6 Kf2 Kc5 
                              Rd1 Kc4 a4 Rh8 a5 Rh2 a6 c5 Kg1 Rh5 Rf1 Rf5 g3 f3 
                              Kf2 Kd4 c3+ Kxe5 Rh1 Rf8 Rh5+ Kd6 Rh2 Kd5 Ke3 c4 
                              Rd2+ Ke5 Rh2 Rg8 Rh5+ Kd6 Kf2 
Nodes: 206645823180
Nodes/second: 8214405
Best move: Qc5
Ponder move: Qxc5