problem 9974 and new problem set

[drahacikfm]

I'm curious if your new generator will delete problem 9974.  1.h4! is the key move, and not so hard to find.  I found it in about 10 seconds.  Then Black played 1...Qf8, completely hanging his queen (I guess to stop some quicker mate).  After 2.Qxf8, White has a queen, bishop, and pawns against Black's lone pawns.  The problem should stop here, of course.  But no!  Now White has to find a mate in 6 more moves.  Not the mate in 7 more moves which I found, and got the problem wrong.

This problem has a very high standard rating, not because 1.h4 is hard to find, but because it's
almost impossible (and useless for practical chess) for a human to find the mate in 6 more moves from the completely winning position where he has a queen and a bishop against nothing.

But if the only first move which mates is 1.h4, will this problem be deleted?  There are mates-in-9 instead of the mate-in-8 given, but these are not with different first moves.  They are with different 3rd or 4th moves.  Will the generator take these cases into account?

For this problem specifically, the solution should be manually truncated to give the win after 2.Qxf8.

[richard]

Hi Drahacik,

Yes, this problem was removed, I didn't look at the analysis but it should have failed at least 2 tests, the longer mates and the winning position after the queen take.  To answer your question about the 1.h4 being the only move that leads to mate and whether the generator will still see the longer mates issue starting at later moves. Yes, the generator has to look for ambiguity at every position, so it repeats the tests at every move in the sequence, so for example a branch at move 3 where it was mate in 8, but an alternative mate in 9 was seen then that problem would be rejected.

Regards,
Richard.