Start Your Engines

[peter]

Cfish keeps insisting it's M26. That's with 32 GB hash and days of back and forth analysis, reaching depth of 111.

So Cfish might be right probably, in output of lines to mate modern engines seldomly give too short lines, too long ones are more common.

Maybe Franz Huber with Chest and or Olaf Jenkner with Gustav could have a look at it too.

[peter]

Cfish keeps insisting it's M26. That's with 32 GB hash and days of back and forth analysis, reaching depth of 111.

So Cfish might be right probably

Found your output- line we are talking about probably and started asmFish for backtracking, which still is somewhat faster than Houdini (but yet maybe not quite as thorough?).

I can take #26- score back to 6th move only in your line, from there onwards 6...Bc3 takes one move longer than ...Ba1 in asmFish's output till now.

Will start H6 with the stored 32G hash again as soon as I find the time to.

[peter]

I can take #26- score back to 6th move only in your line, from there onwards 6...Bc3 takes one move longer than ...Ba1 in asmFish's output till now.

Will start H6 with the stored 32G hash again as soon as I find the time to.

So here is the output with the reloaded hash after your output- line

rk1b4/p2p2p1/1P6/2R2P2/8/2K5/8/5B2 w - - 0 1

1. Rc8+ Kb7 2. Bg2+ Kxb6 3. Rxa8 Bf6+ 4. Kc4 Be5 5. Kd5 Bf6 6. Rf8 Ba1 7. Rf7
a5 8. Rxd7 a4 9. Bf1 a3 10. Rd6+ Kc7 11. Ra6 Bb2 12. Bc4 a2 13. Rxa2 Bf6 14.
Ra3 Bb2 15. Rb3 Bc1 16. Ba6 Kd8 17. Rb7 Bh6 18. Ra7 Bf4 19. Rxg7 Ke8 20. Ke6
Kd8 21. f6 Be3 22. Rd7+ Ke8 23. Rc7 Bb6 24. Rc8+ Bd8 25. Bb5+ Kf8 26. Rxd8#

at 6...Ba1:

[d]5R2/p2p2p1/1k6/3K1P2/8/8/6B1/b7 w - - 0 1

Analysis by Houdini 6.03 Pro x64-popc:

7.Tf7 a5 8.Txd7 a4 9.Lf1 a3 10.Td6+ Kc7 11.Lc4 Lb2 12.Ta6 Kb7 13.Kd6 Lc1 14.Tc6 Le3 15.Kd7 Lb6 16.Te6 a2 17.Lxa2 Ka6 18.Kc6 Ka5 19.Te4 Ld8 20.Lb3 Ka6 21.Ta4+ La5 22.Lc4+ Ka7 23.Txa5+ Kb8 24.Ta3 g5 25.fxg6 Kc8 26.Ta8#
+- (#20) Tiefe: 68/58 00:00:44 2245MN

and here after one ply more bacckward:

5R2/p2p2p1/1k3b2/3K1P2/8/8/6B1/8 b - - 0 1

Analysis by Houdini 6.03 Pro x64-popc:

6...Kc7 7.Rxf6 gxf6 8.Bf1 Kb6 9.Kd6 a5 10.Bc4 Kb7 11.Kxd7 Kb6 12.Ke6 Kc5 13.Ba2 Kc6 14.Kxf6 Kd6 15.Kg7 Ke5 16.f6 Kd4 17.f7 Ke3 18.f8Q Kd3 19.Qa3+ Kd4 20.Qxa5 Kd3 21.Qa3+ Kd2 22.Bc4 Kc2 23.Qb4 Kc1 24.Qc3+ Kd1 25.Qb2 Ke1 26.Qe2#
+- (#20) Depth: 56/48 00:00:13 814MN

Tried the 6...Bc3 I had here sooner too, got #26 after some forward backward there too now.

5R2/p2p2p1/1k6/3K1P2/8/2b5/6B1/8 w - - 0 1

Analysis by Houdini 6.03 Pro x64-popc:

7.Rf7 Kc7 8.Bf1 Ba1 9.Bb5 Kb6 10.Bxd7 Bc3 11.Rf8 Ka5 12.Kc4 Bf6 13.Rc8 Kb6 14.Rc6+ Kb7 15.Kd5 a5 16.Bc8+ Kb8 17.Ke6 Ba1 18.Kd7 Bd4 19.Ra6 Ba7 20.Rxa5 Bd4 21.Rb5+ Ka7 22.Kc6 Bf2 23.Bb7 Kb8 24.Rd5 Ka7 25.Ra5+ Kb8 26.Ra8#
+- (#20) Depth: 62/47 00:00:06 323MN


So I'd give in to #26 as for the more probable number of moves to mate, even if I have corrected my lines only from 6th forwards, at 4th I'd have to check at least 5 more alternative ones, but that is too much work for too little interest to me.

One more time: that's not a mate in x- problem.

[Spliffjiffer]

hi Louis :--)
im still trying to figure out how white proceeds after 4...Bg5 (and 4...kc7)...ill try to figure it out within the next hours...if im not successful might u be able to help me at this point (giving a line) ?

[Spliffjiffer]

4...bg5 leads also to a mate in 26 in the end...only 4...kc7 remaining...

[Spliffjiffer]

4...Kc7 is also leading to a mate in 26 in the end (at least)...
so far i can confirm Louis's analysis and can tell that 3...Bf6+ leads to a mate in 26 in the end !!...ill try to solve that back later to the root position...the probability that original position is (at least) a mate in 26 moves is extremely high from my point of view !!?...ill report later about my backsolving issue

[Spliffjiffer]

hi Louis
i was able to do the job myself fortunately ....i will probably be able to confirm your analysis pretty soon !?

[zullil]

hi Louis
i was able to do the job myself fortunately ....i will probably be able to confirm your analysis pretty soon !?

That's good, because my computer has moved on to other things.

[Spliffjiffer]

im now able to confirm Louis's statement that the original position is NOT a mate in 27 but a mate in 26 moves or less !!...thank you Louis for making such an incredible deep analysis !..in my eyes youre the only one who deserves points (if any, probably not ) for the task that was made...u inspired me to dig deeper and u were simply right..full credits to your amazing analysis !

regards, Peer

[zullil]

im now able to confirm Louis's statement that the original position is NOT a mate in 27 but a mate in 26 moves or less !!...thank you Louis for making such an incredible deep analysis !..in my eyes youre the only one who deserves points (if any, probably not ) for the task that was made...u inspired me to dig deeper and u were simply right..full credits to your amazing analysis !

regards, Peer

Thanks, but I really did very little. Credit should mostly go to Cfish and Intel.