Sometimes Even An Engine Is No Help

     The position below is a position from the game Shirov vs. Piket, Amsterdam 1995, that was first published in New in Chess and later in his book, Fire On Board. Shirov wrote that sometimes analyzing a game gives more pleasure than than the game itself. That's what he felt uncovering the mysteries of 23...Qc5 24.Bg5! Since Shirov mentioned that he was using Fritz in his analysis, when I first read the book and came to this game, I analyzed it with the Xtreme Chess engine and it was unable to find Shirov's suggested 24.Bg5. 
     When I came across the game again the other day I was curious to see if his statement that no engine would find 24.Bg5 is still true, so checked it out with Komodo 8 and Stockfish 6. Neither engine found it! ...at least not after 15-20 minutes or so. In fact, Komodo 8 dropped its evaluation from slightly over a Pawn to a little more than half a Pawn when I actually played the move.
     After spending a couple of hours looking at this position with Stockfish and Komodo I was unable to come to a clear conclusion about Shirov's 24.Bg5 even using engines that are much strong than whatever version of Fritz he was using back in 1997.