LeeRoth wrote on 08/08/09 at 01:13:59:
After 4..Bb4 5.Nbd2 f6 6.a3!, Palliser notes that 6..Ba5 leaves the bishop misplaced when White begins to undermine the Black queenside with a4. He does not, however, analyze 6..Ba5.
I think that Palliser is simply wrong about that. Since 7...dxc4 is strong after any retreat of White's bishop, White has to play 7.b4 immediately, when 7...fxg5 8.bxa5 Nf6 is very interesting and I believe at least adequate for Black.
I think that 5.Nbd2 is decidedly suboptimal and that 5.Nc3! should be played. Then 5...Nf6 brings up a position that has occurred in many games, much more often by Black's playing 5...Bb4 in reply to 5.Bg5 in the standard Anti-Meran move order. It's not very good for Black. Black's moves besides 5...Nf6 have produced dismal results in my database, but entirely in low-level games.
So I think that Palliser is right that 4.Bg5 Bb4+ favors White, but for the wrong reason.
I now think that Black's best answer to 4.Bg5 is 4...Qb6. There are just two high-level games in my database: Shalimov-Berezin, Kiev 2002 was a draw, while Black actually won in Eingorn-Gurevich, Graz 1995. Both continued 5.Qc2 Nd7 6.e3 Nf6 7.Nc3, a position that has occurred in 26 games in my database, with an average score of 50%.
As for 4.Qc2, which is at least a good move, the critical test appears to be 4...dxc4 5.Qxc4 and now if 5...Nf6 6.Bg5, or if 5...b5 6.Qc2 Bb7 7.e4. In the latter case I think that += is not exaggerated. In the former, 6...h6 7.Bh4 b5 8.Qc2 Bb7 looks decent for Black, but White may be able to prove some advantage.
If all this is right then 4.Qc2 may be a somewhat better move than 4.Bg5, but not necessarily one that is likely to produce much advantage. So my latest thinking is that even against 3.Nf3, the Triangle is a good way to angle for a Meran.