New test suite "Tactical Insanity"

[Dann Corbit]

If you see an am along with a bm, that means that no, really, the am move is not a solution.

[Dann Corbit]

The inspiration for this test is a mish-mash of things, but mostly these are positions where about half the engines I tried got the wrong answer after 30 seconds using 6 cores at 3.4 GHz.

Some are positions I have studied over time and found interesting.

All of them are positions which I did not have an ID for, so they are most likely not contained in other famous EPD test suites (but there are probably exceptions to that since I did not check carefully).

[Dann Corbit]

First correction (bm should be a4):
2bqkbnr/3r1ppp/p1npp3/1p6/4P3/1BN1BN2/PP2QPPP/R4RK1 w k - bm a4; id "TI.021";

Analysis:
pv = a4 b4 Nd5 Bb7 Bb6 Qa8 Nc7+ Rxc7 Bxc7 Qc8 Bb6 Nf6 Rac1 Nd7 a5 Nc5 Bd5 f6 Bxc6+ Qxc6 Nd4 Qxe4 Qxe4 Bxe4 Bxc5 dxc5 Nxe6 Kf7 Nxc5 Bxc5 Rxc5 Bd3 Rd1 Bb5 Rc7+ Kg6 g3 Re8 Rd4 Re1+ Kg2 b3 Rb4 Ra1 Rg4+ Kf5
ce = 306

[peter]

Thanks a lot, Dann!
Already the first three of them I didn't know till now.

Very first one is a good example to me, that also best moves, which are not the only ones winning the game, can be used as for testing engines as well.

Second best winning moves, with much more longer lines to win, simply aren't best.

If you start collecting test positions in a too puristic manner, you can forget about positional test suites anyhow at all. There are no single one "game changing" best moves in positional test positions, or they aren't positional.

[Dann Corbit]

Hope you enjoy it.
I expect that defects will be found (I already found one) and I hope that people have fun with it.
Some of them are kind of surprising for humans too.
Others are a lot more obvious to humans that to the machines.

[BBauer]

I tried.
Pos 56 has as bm Rc1.
IMHO Kd4, a3, Rb1 win as well.
So what? I expected something like that.
Kind regards
Bernhard

[peter]

IMHO Kd4, a3, Rb1 win as well.

You're right, Bernhard.

Rc1 is not even the shortest way to mate,

6k1/5p2/5Pp1/p2p1pP1/1n3P2/2p1K3/P7/7R w - - 0 1

Analysis by Cfish Tactical 2.0 250518 64 POPCNT NUMA:

1. +- (#26): 1.Kd4 Nxa2 2.Rh2 Nb4 3.Kxc3 Nc6 4.Rd2 Nd8 5.Kd4 Ne6+ 6.Ke5 d4 7.Ra2 Kh7 8.Rxa5 Kh8 9.Ra3 Kh7 10.Rd3 Kg8 11.Rh3 Nc5 12.Kxd4 Ne6+ 13.Ke5 Nc5 14.Rc3 Ne6 15.Rc4 Kh7 16.Kd6 Kh8 17.Ke7 Kg8 18.Ke8 Kh7 19.Kxf7 Nd8+ 20.Ke8 Ne6 21.Rc8 Ng7+ 22.Ke7 Ne6 23.f7 Nd4 24.f8Q Nc6+ 25.Kf6 Nd4 26.Qh6#

2. +- (#30): 1.Rc1 Nc6 2.Kd3 d4 3.Kc4 Kf8 4.Kc5 Nd8 5.Kxd4 Ke8 6.Kd5 Kd7 7.Rxc3 Ne6 8.Ke5 Nf8 9.Rc4 Ne6 10.a4 Nf8 11.Rc5 Ne6 12.Rxa5 Nc7 13.Ra7 Kc6 14.Rxc7+ Kxc7 15.Kd5 Kd7 16.a5 Kc7 17.a6 Kb6 18.Kd6 Kxa6 19.Ke7 Kb6 20.Kxf7 Kc5 21.Kxg6 Kc4 22.f7 Kd4 23.f8Q Kc3 24.Kxf5 Kd3 25.g6 Kc3 26.g7 Kd3 27.g8Q Kd4 28.Qb4+ Kd3 29.Qg2 Ke3 30.Qbd2#

3. +- (#35): 1.a3 Nc6 2.Kd3 d4 3.Rd1 Kf8 4.Re1 c2 5.Rc1 Ke8 6.Rxc2 Kd7 7.Rc5 Kd6 8.Kc4 Kd7 9.Rd5+ Kc8 10.Rb5 Kc7 11.Rxf5 d3 12.Kxd3 Kd6 13.Rb5 Nd8 14.Rxa5 Kc6 15.Ra7 Kc5 16.Rd7 Ne6 17.Ke4 Kc6 18.Rxf7 Kd6 19.Re7 Nc5+ 20.Kd4 Ne6+ 21.Rxe6+ Kxe6 22.Kc5 Kf5 23.f7 Kg4 24.f8Q Kg3 25.Qf6 Kf2 26.Kd4 Kg2 27.Qxg6 Kg3 28.Ke5 Kh2 29.Qd6 Kg2 30.Kf5 Kg3 31.g6 Kg2 32.g7 Kf3 33.g8Q

[Vinvin]

Cfish Tactical 2.0 250518

Where can I find this engine, please ?

[Dann Corbit]

I tried.
Pos 56 has as bm Rc1.
IMHO Kd4, a3, Rb1 win as well.
So what? I expected something like that.
Kind regards
Bernhard

Thanks for the defect report.

[peter]

Hi Vincent!

Cfish Tactical 2.0 250518

Where can I find this engine, please ?

Link to Immortal obviously wasn't allowed to be given, before I give it another try with MZ- Forum, how about that?