MNb wrote on 04/17/07 at 02:46:14:
See the link I gave above. Markovich thinks White's game easier to play after 8.Ne4 etcetera, though the resulting position might be equal.
Short and Beljavsky only play the Traxler at "fun" occasions like the Spartakiad - when Karpov choose the Italian instead of his normal Ruy Lopez.
Yeah, as I recall, the critical Fritz line (with 8...Ne6! -- Berliner's attack looks very doubtful to me) leads to an ending where White has two pawn islands and Black has three, including isolated a- and c-pawns. To defend such an ending is not really why I play 1...e5, but perhaps Black is ultimately equal. I wish a stronger player would give his opionion on this. Personally as White, I would play on and on, since there appears to be scant risk of White's losing.
But this is why I think that 5...Na5 is the most correct. However I think that 5...b5 is a great move for young and improving players to play, and I have my students playing it, premised mostly on their opponents' never knowing White's optimal course of action. It's nice that MNb's 8...Ne6 analysis is there in case they do, since Black develops a beautiful game against White's merely intuitive replies. There is also the speculative 6. Bf1 h6!? 7. Nxf7! Kxf7 8. dxc6 Bc5 9. Be2! h5!? or 9...Ne4!? though objectively it seems that Black doesn't have enough in either case.
I consider the Traxler just plain bad after 5. Bxf7+ and then 6. Bb3 or 6. Bd5. We've been over this, of course. I would never play 5. Nxf7, particularly since I doubt that it wins.