Computer Analysis Mistake. Problem 45855

[tama]

The main line in this problem goes Bf3,Bf5,Bd5+,Kh8,Qh4,Bg4, and the computer analysis says that Qxg4 is the best move but really Rf5 is mate in 4!!

This problem is also very short because during solving the computer would go after Bf3,Kh8 right away giving up the Bishop and preventing mate.

I think this is a new problem generated by the last update because it has 3 attempts.

I felt maybe toga isn't picking up on some tactics in the variations seen by Gold members, because i have seen weird things like this before were the main variation showed to gold members ends in one side having a mate in 1 or one side completely hangs a piece! lol

[drahacikfm]

There was a discussion about the moves in the Gold member variations not being good near the end of the variations here, just in the last couple days:

http://chesstempo.com/chess-forum/site_feedback/truncating_the_displayed_variations_for_gold_memberships-t453.0.html

The example you mentioned is slightly different.  The mate in your mate-in-4 happens exactly one move after the depth that Toga searched (depth 13).  After White plays 1.Bf3, then Toga sees that mate, and changes it's choice for Black to Kh8 so that it doesn't get mated.  It's just related to the search depth.  If Toga thought one move more in the starting position, it would have seen the mate and the variation displayed would have been Kh8 instead of Bf5.

Many times the problem doesn't play all the moves that you see in the variation shown for move 1.  That's because it's looking deeper as the problem progresses move by move.