Start Your Engines

[Spliffjiffer]

4...be5 leads to a mate in 26 (at least) finally, i can confirm this !

[Ovyron]

Which would mean, provided M26 is correct, nobody had it right

Which engines even find Rc8+ on their own, i.e., without launching a browser and Googling the FEN?

Houdini 6 with Contempt=2 doesn't even consider anything else.

Sounds like a bug.

Okay, sorry, yeah, I meant to say "it doesn't output anything else"

[zullil]

Which would mean, provided M26 is correct, nobody had it right

Which engines even find Rc8+ on their own, i.e., without launching a browser and Googling the FEN?

Houdini 6 with Contempt=2 doesn't even consider anything else.

Sounds like a bug.

Okay, sorry, yeah, I meant to say "it doesn't output anything else"

What evaluation does the engine show??

[peter]

Hi!

After 1...Kb7 and some backtracking, I get this output:

[d]r1Rb4/pk1p2p1/1P6/5P2/8/2K5/8/5B2 w - - 0 2

Analysis by Houdini 6.03 Pro x64-popc:

2.Lg2+ Kxb6 3.Txa8 Lf6+ 4.Kc4 Le5 5.Kd5 Lf6 6.Tf8 Lc3 7.Tf7 Kc7 8.Lf3 Lb2 9.Ld1 Lf6 10.La4 Kb6 11.Txf6+ gxf6 12.Lxd7 Kc7 13.Ke6 a6 14.La4 a5 15.Ke7 Kb6 16.Kxf6 Kc5 17.Ke5 Kb4 18.f6 Kxa4 19.f7 Kb5 20.f8D Kc4 21.Da3 a4 22.Dxa4+ Kc5 23.Ke6 Kb6 24.Kd6 Kb7 25.Db4+ Ka7 26.Kc6 Ka6 27.Da3#
+- (#26) Tiefe: 66/71 00:13:20 40896MN

And after one more ply backward this one in 2MV mode:

[d]rkRb4/p2p2p1/1P6/5P2/8/2K5/8/5B2 b - - 0 1

Analysis by Houdini 6.03 Pro x64-popc:

1. +- (#26): 1...Kb7 2.Lg2+ Kxb6 3.Txa8 Lf6+ 4.Kc4 La1 5.Tf8 Le5 6.Kd5 Lc3 7.Le4 Kc7 8.Tf7 a6 9.Lc2 Lf6 10.Kc5 Lb2 11.La4 La3+ 12.Kd5 Lb2 13.Txd7+ Kb6 14.Td6+ Kb7 15.Ke6 Kc7 16.Tc6+ Kb7 17.Kd7 Ld4 18.Lc2 Lb6 19.Le4 Ka7 20.Te6 a5 21.Kc8 a4 22.Ld3 a3 23.Te7+ Lc7 24.Kxc7 a2 25.Te1 g5 26.Ta1 Ka8 27.Txa2#

2. +- (#22): 1...Kxc8 2.La6+ Kb8 3.b7 La5+ 4.Kd4 Lb6+ 5.Kd5 Lc7 6.f6 gxf6 7.Ke4 Le5 8.Kf5 Ld4 9.Kg6 f5 10.Kf7 Lb6 11.Ke7 Kc7 12.bxa8D Lc5+ 13.Kf6 Ld4+ 14.Kxf5 Kd6 15.Db8+ Kc6 16.Dc8+ Kd6 17.Lb5 a5 18.Lxd7 Ke7 19.Dc6 Kf7 20.Le6+ Kg7 21.Dc7+ Kh8 22.Df7 a4 23.Dg8#

[Rebel]

IMO it's (also) a M27, if no objections I will update the ranking accordingly.

[zullil]

IMO it's (also) a M27, if no objections I will update the ranking accordingly.

I don't deserve to get a point for submitting an incorrect answer. It's M26, not M27. (Probably ) So I should finish the competition with a perfect 0.

[Spliffjiffer]

actually i am able to replicate Louis analysis exept for 3 lines still:
4..Kc7
4...Bg5
4..Bh4
im still trying to figure this out...every other lines after 4. Bf6+ lead to a mate in 26 or less in the end...so far im on Louis side and believe him and his analysis...but these 3 variations driving me crazy ...let me some time Louis, i will ask u for help if i do not manage it for my own...impressive analysis by you !

[peter]

Hi Louis!

It's M26, not M27. (Probably )

For me it's #27 as in my H6-output- line some postings above,

http://www.talkchess.com/forum/viewtopi ... 06&t=66359

, but it's a matter of keeping the whole way backward to the starting position with correct mate- score in hash and output, which is really hard to be realized, at least with just 32G hash to be used, in this way H6 to me seemed to be best so far as for this position.

And I don't really want to compete here as for proving this single one move more or less to be true or not, especially seeing Kubbel's fine study clearly not as a mate in -x -problem at all.

Yet, if you'd like your main line to get backtracking compared with my still stored 32G H6- hash by me, please show it here again, so I don't have to look for it somewhere else.

Of course I have a rather big .pgn of many lines to mate stored already, I don't give it here at once neither, I'd have to check all of them once again to be sure not to have lost or gained the one or the other move anywhere, and so would have anybody else wanting to judge on their own too.

As for Ed's fine Start Your Engines- Site, I'd rather not register there neither, I probably would spend much to much time with it then right after starting.

Kubbel's study I had in my own database since quite a while, and I guess I so have most of the better known studies, which are hard for engines also still already.

Yet thanks to all writing about the site here too, bringing especially interesting positions along too now and then again.

[zullil]

Hi Louis!

It's M26, not M27. (Probably )

For me it's #27 as in my H6-output- line some postings above,

Hi Peter,

I misunderstood your previous post. I saw M26 without noting that was not from the root position.

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

Might look one more time when I can. Don't have access to that machine now, but the search is still running.

Louis

[Spliffjiffer]

4...bh4 leads also to mate in 26 in the end...bg5 and kc7 remaining...