I played an interesting Center Game in the Tampa Chess Club Championship this past weekend.  My opponent was Corey Acor, USCF 2274.  Although I lost, it had nothing to do with the opening, where my position was rated strongly by Fritz until a time-pressure blunder on the 23rd move.  Notes in the Move List.

Really good effort by both players.

I'm going to guess that the reason you tried 23. Bc3 was to chase off or remove Black's d4 knight, enabling you to play Be6, winning the queen? Just a bit too hopeful that, I guess.

Anyway, I ran the position after move 22 through Sigma Chess HIARCS 12.1 MP and it came up with the same recommendation as your Fritz at first--which 23. Ng5 line HIARCS rated at +1.95, a nice advantage for White. But after a few minutes, HIARCS found and preferred:

23. h6 Qxd5 24. Nc3 Qf7 25. hxg7 Kxg7 26. Kb1 Rh8 27. Ne4 Nc4 28. Ng5 Nxd2 29. Qxd2 Qf6 30. Qg2, which HIARCS rated at +3.94.

Because HIARCS scores this 23. h6 line as twice as good for White, I'm curious as to what your Fritz thought of this line? Does Fritz (which version?) see something wrong with 23. h6 that HIARCS doesn't?

Hi, nimble...

    I was using Fritz 8.  At the top level, Fritz rated 23. Ng5 over 23. h6 (at +1.76).  Instead of the HIARCS line, Fritz likes 25....b4 26. Qh2! Kxg7 (if 26...bxc3 27 Be6 cxd2+ 28 Kd1 Qxe6 [28...Nxe6 29. Qf8#] 29. Qg7+ and g8/Q)  and at this point, the evaluation shifts to +3.59 with the continuation 29...Kf8 30. Qb7 Nf7 31. Qxa8+ Qe8 32. Qxa7.

    Crazy stuff...but the whole CG is a very rich, unexplored area.

 It looks to me like a great surprise weapon. I'll be exploring it. Thanks.