1.h3 Da2 2.h4 Da1 3.h5 Da2 4.h6 Da1 5.h7 Da2 6.h8=C Da1 7.Cf7 Da2 8.Cd8 Da1 9.Cb7 Da2 10.Cxc5 Da1 11.Ce4 Da2 12.Cd6 Da1 13.Cxc4 Da2 14.Ca5 Da1 15.Cxb3# 

I thought engines are not allowed....

Found mate in 11. You must just push the pawn 1 square at a time. thus the queen will be on a1 and you can mate him. I didnt use a engine

ChessMaster solved it in less than 5 seconds!

only movable peice in black is queen and traped so white will promot to KNIGHT and smother mate black.

To the reader posting the mate in 5 moves...

Remember that is it NOT black that is to mate white here. (and btw your sequence suggests less than optimal moves from white here).

It is white that with optimal moves can mate black in 15 moves and no move sequence shorter than 15 moves can lead to white mating black (again assuming optimal play from Black here as well).

cyrano found the solution (mate in 15) in 39 seconds.

An engine I am developing which has thus far defeated Deep Rybka 4 and Houdini 1.5 solved the problem in less than 10 seconds

BTW there is a mate in 5 as well

Kf2 e1=Q+

Kxe1 Rxh2

Kf1 Nd2+

Kg1 Nf3+

Kf1 Be2#

My new chess engine, Deep Saros 2.3-c2, found the solution in small time:

M15 1.h3 Da2 2.h4 Da1 3.h5 Da2 4.h6 Da1 5.h7 Da2 6.h8=C Da1 7.Cf7 Da2 8.Cd8 Da1 9.Cb7 Da2 10.Cxc5 Da1 11.Ce4 Da2 12.Cd6 Da1 13.Cxc4 Da2 14.Ca5 Da1 15.Cxb3#  (46.91)

CPU: AMD Turion X2 - 64bit

Download Deep Saros for test: http://www.mediafire.com/?cq512le4hf4hgs0

Carotino

I also tried this problem on my own engine  - MEngine. As expected it does not find the correct solution, but only manages a draw.

billwall, good work. I have tried with Shredder 12 - it also found the mate in 15. Fritz 11 did not and rybka 3 did not find it either.

The mate in 14 line was wrong. It is mate in 15 as first stated. Remember - you should assume best play from black and white as always when solving mate problems. Yes, there is a mate in 14 if black does not play optimal.

Mate in 14?

Could you send me that line so I can veryify that claim - all engines that have solved it says mate in 15 so I am bit sceptic.

Its Mate in 14.Togall found it.

Hint.(Queen must not be in a2 when crusial check arrives..).

rich - that is actually an interesting observation. We should try and add that aspect to our engines: the ability to simply disbelief a given position due to lack of realism. Should save well on the amount of necessary calculations. 

56. Nh8, [game aborted due to unrealism]

billwall, thx. I have looked at other positions, but this seems to be the best I have found so far. There is a clear difference between engines on this one.

Good test.  I tested it against over 50 UCI engines and 5 engines found the mate in 15 in less than 2 minutes.  The rest found the easy draw after queening.    Any other positions to test engines?  This was the first one that demonstrated a clear advantage with a mate in a short amount of time.  I tested the position against abrok, alichess, anmon, arason, ares, aristarch, baron, biglion, bright, booot, comet, crafty, cyrano, daydreamer, deep frenzee, delphil, ECE, firebird, fritz 10 and 12, fruit, gibbon, glass, glaurung, hamsters, Hiarcs 9, list, little thought, movei, N2, naum, patzer, pawny, pharaon, philou, pro deo, rubka 2.2 and 3.3, Shredder 7, simplex, slowchess, spark, spike, stockfish, sungorus, tao, toga, twisted logic, ufim, umko, wildcat, and zappa.  Anmon, Glauring, Hamsters, Movei and Pro Deo found the solution.   There are over 100 chess engines out there.  Anyone else tested their engines on this problem?

This wouldn't materialise in a real game.