KEG: A nice escape by Maroczy from what appeared to be a crushing attack by Pillsbury.
Pillsbury's 12...Bg4 was indeed the opening of Pillsbury's attack as keypusher has noted, but the prosaic 12...exf4 may have been the theoretically better move. Had he played thus, however, Pillsbury's attack would probably never have developed.
The two big questions in analyzing and enjoying this game are: (1) how did Maroczy get into such trouble; and (2) how did he escape from Pillsbury's clutches.
The answer to the first question begins with Maroczy's surprisingly poor 18. Nh2. After Pillbury's 15...Be6 and the exchange of White-squared Bishops, Maroczy had if anything the better game. But instead of trying to build on his advantage with 18. Rfe1 or 18. d4, Maroczy went passive with 18. Nh2. It did not take Pillsbury long to capitalize on this, especially after Maroczy's doubtful 19. c4 (19. Qe2 was best) and his very poor 20. f3 (20. Qb4 was much better).
Maroczy got into real trouble after 21. c3 (counter-attack by 21. Rab1 was indicated) and his unfortunate 22. Rfb1 (22. Rab1 would still have left him a defensible position).
With 20...h5, Pillsbury's King-side steamroller was off and running, and after Maroczy's very poor 23. Kf1 (his only chances lay in 23. Qe1 or 23. Qc2) Pillsbury played 23...Rg8 and his attack seemed unstoppable (23...g4 immediately also looks crushing).
After Maroczy's 26. Rh1, the game seemed over.
But from this point on, Maroczy's defense was superb, and Pillsbury's one major slip cost him an otherwise well-earned victory.
Pillsbury should probably have played 26...Rh7 (instead of his actual gxf3+), but he still had the game well in hand.
In what appeared to be a lost position, Maroczy tried 30. Qd1. This may not have been theoretically best (Fritz plays 30. Kd1), but it proved to be his salvation, since it allowed him to begin a counter-attack with his Queen.
Pillsbury's fatal slip that allowed Maroczy to draw was his 30...Ng4. The Tournament Book recommends 30...Qd7 to stop Maroczy's Qa4, and this does indeed seem to keep Pillbury's winning chances alive, but best of all would have been 30...Qg6.
After 30...Ng4, Maroczy gave Pillsbury no time to consummate his King-side attack. His 31. Qa4 turned the game around, and after Pillsbury's seemingly powerful 31...Ne3 Maroczy refused to be distracted with 32. Qxa7 (which may also have been sufficient to draw) and played 32. Qe8+.
Remarkably, Pillsbury had only one saving move in this position, 32...Rg8 (as he played). I was shocked to discover that anything else loses. What a turnaround!
Maroczy's 34. g3 was lovely (though Fritz claims a draw for White with 34. d4 as well).
The key to Maroczy's defense came with his wonderful 38. g4. While 38. gxf4 may also work (or so Fritz claims), Maroczy's 38. g4 again forced Pillsbury to find the only saving move (38...Qh3).
After these fireworks, the game ended in a spectacular draw. As keypusher has shown, even in the final position, Pillsbury had to be careful, since 41. Qe6+ Rg6 would have lost.
Despite a few mistakes, this was a fine fighting draw.