Is it possible to give Rebel a position to analyse?

I am just asking because with Rebel 16.1 this is not going well in the Shredder interface (which for me is still the standard in UCI, like it or not )

I give Rebel one of Vincent's Hardchess positions:

4k1br/1K1p1n1r/2p2pN1/P2p1N2/2P3pP/5B2/P2P4/8 w - - bm Kc8; id "Hard-Talkchess-2020.074" ;

I press the "Analyze" button:


4k1br/1K1p1n1r/2p2pN1/P2p1N2/2P3pP/5B2/P2P4/8 w - -

Engine: Rebel-16.1 (64 MB)
gemaakt door Ed Schroder and Chris Whittington (C) 2

19/40 0:01 -0.90++ 1.Lxg4 d4 2.a6 (1.696.723) 999

19/33 0:01 -0.66++ 1.Lxg4 (1.760.828) 1001

19/34 0:01 -0.99 1.Lxg4 Pe5 2.Pd6+ Kd8 3.a6 Pxg6 4.a7 Le6
5.a8D+ Ke7 6.Pf5+ Lxf5 7.Da3+ Ke6
8.Lxf5+ Kxf5 9.Dd3+ Ke6 10.Dxg6 Txh4
11.a4 Txc4 12.Dg7 Th2 13.Dg8+ Kf5
14.a5 (1.764.291) 1001

20/36 0:01 -0.83++ 1.Lxg4 d6 (1.833.355) 996

20/35 0:01 -0.67++ 1.Lxg4 (1.849.666) 997

20/38 0:01 -0.43++ 1.Lxg4 (1.909.661) 1000

20/38 0:01 -0.07++ 1.Lxg4 (1.967.918) 1002

20/42 0:02 0.00 1.Lxg4 Pe5 2.Pd6+ Kd8 3.a6 Pxg6 4.a7 Le6
5.a8D+ Ke7 6.Pf5+ Lxf5 7.Da3+ Ke6
8.Lxf5+ Kxf5 9.Dd3+ Ke6 10.Dxg6 Txh4
11.a4 Txc4 12.Dg7 Te8 13.a5 Te7
14.Dg8+ (2.098.670) 1002

21/44 0:02 0.00 1.Lxg4 Pe5 2.Pd6+ Kd8 3.a6 Pxg6 4.a7 Le6
5.a8D+ Ke7 6.Pf5+ Lxf5 7.Da3+ Ke6
8.Lxf5+ Kxf5 9.Dd3+ Ke6 10.Dxg6 Txh4
11.a4 Txc4 12.Dg7 Te8 13.a5 d6
14.Dh6 (2.359.938) 999

22/42 0:02 0.00 1.Lxg4 Pe5 2.Pd6+ Kd8 3.a6 Pxg6 4.a7 Le6
5.a8D+ Ke7 6.Pf5+ Lxf5 7.Da3+ Ke6
8.Lxf5+ Kxf5 9.Dd3+ Ke6 10.Dxg6 Txh4
11.a4 Txc4 12.Dg7 Te8 13.a5 d6
14.Dh6 (2.816.923) 986

23/48 0:03 0.00 1.Lxg4 Pe5 2.Pd6+ Kd8 3.a6 Pxg6 4.a7 Le6
5.a8D+ Ke7 6.Pf5+ Lxf5 7.Da3+ Ke6
8.Lxf5+ Kxf5 9.Dd3+ Ke6 10.Dxg6 Txh4
11.a4 Txc4 12.Dg7 Te8 13.a5 d6
14.Dh6 (3.325.880) 983

24/41 0:03 0.00 1.Lxg4 Pe5 2.Pd6+ Kd8 3.a6 Pxg6 4.a7 Le6
5.a8D+ Ke7 6.Pf5+ Lxf5 7.Da3+ Ke6
8.Lxf5+ Kxf5 9.Dd3+ Ke6 10.Dxg6 Txh4
11.a4 Txc4 12.Dg7 Te8 13.a5 d6
14.Dh6 (3.851.817) 977

24/41 0:05 +0.16++ 1.a6 (5.397.609) 959

24/34 0:05 +0.32++ 1.a6 (5.629.709) 959

24/38 0:06 +0.56++ 1.a6 (5.859.637) 960

24/42 0:06 +0.92++ 1.a6 (6.119.026) 959

24/44 0:11 +1.46++ 1.a6 (10.342.919) 934

The solution is for many engines apparently very hard, that is not the issue here. (Probably Stockfish also can't find it, I'm trying Eduard's engines for instance, no luck with them after three hours.) But if I now (after a few minutes in this case) give in the first move of the solution, Kc8, then Rebel 16.1 would still respond with this:


2K1k1br/3p1n1r/2p2pN1/P2p1N2/2P3pP/5B2/P2P4/8 b - -

Engine: Rebel-16.1 (64 MB)
gemaakt door Ed Schroder and Chris Whittington (C) 2023

24/43 0:02 -2.70++ 1...a6 2.gxf3 (2.173.703) 1056

24/38 0:02 -2.81 1...a6 2.Pe5 (2.205.127) 1056

25/40 0:02 -2.81++ 1...a6 2.gxf3 (2.881.777) 1059

25/43 0:02 -2.75 1...a6 2.Pe5 (2.931.891) 1058

26/38 0:03 -2.75++ 1...a6 2.gxf3 (3.761.700) 1060

26/38 0:03 -2.75 1...a6 2.Pe5 (3.805.009) 1060

27/39 0:04 -2.75++ 1...a6 2.gxf3 (4.916.504) 1046

27/37 0:04 -2.75 1...a6 2.Pe5 (4.990.492) 1045

28/45 0:05 -2.75++ 1...a6 2.gxf3 (6.020.892) 1041

28/50 0:07 -2.98 1...a6 2.Pe5 (7.175.360) 1014

29/51 0:12 -2.59 1...a6 2.Pe5 (12.052.684) 977

30/54 0:14 -2.75++ 1...a6 2.gxf3 a7 (14.072.528) 960

30/49 0:17 -2.70 1...a6 2.Pe5 (16.408.254) 955

which clearly is illegal. I know there were some bugfixes but I just wanted to check if there is now a legal response?

This position is known, it is also part of the ERET test.

I have looked at Rebel's main variations and can not see that there is an illegal move here.

However, I may be mistaken.

Thanks for checking Dio!