After having demonstrated fairly convincingly that 10...Qd7 in the main-line of the Norwegian variation (after 1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Ba4 b5 5.Bb3 Na5 6.0-0 d6 7.d4 Nxb3 8.axb3 f6 9.Nc3 Bb7 10.Nh4) is at best difficult for Black, Greet in his "Play the Ruy Lopez" moves on to Black's other major option:
This may be Black's most natural way of facing the Qh5+ threat. However, I have always thought it inferior to 10...Qd7 as I quite early was taught that the black queen needed access to f7, and this knight move obviously interferes her direct route.
Greet recommends this move, which probably is stronger than 11.f4 b4 12.Nd5 Nxd5 13.exd5 Bxd5 when Black should be able to defend, e.g. 14.Qe2 Qd7 (14...Qe7?!) 15.fxe5 dxe5 16.dxe5 0-0-0 (16...Qb5!? 17.Qf2 fxe5 18.Nf5 g6 19.Ne3 Be7 may be best) 17.Qxa6+ Bb7 18.Qc4 fxe5 19.Be3 and White had some advantage in Vihinen-O.Moen, Gausdal 1994.
11...fxe5?! 12.f4 gives White a dangerous attack.
The somewhat less natural 12...Qc8!? may transpose and in some lines even improve: 13.Be3 (13.Rd1 h5 14.h3 Qe6 simply transposes to the 12...Qd7 line with 14.Be3) 13...h5 14.h3 Qe6
and now a split:
a) 15.Rfd1 g5 16.Nd5 0-0-0 17.Nxe7+ Bxe7 18.Qf5 Qxf5 19.Nxf5 Rxd1+ 20.Rxd1 Bf8 21.f3 h4 was better for Black in Nyysti-L.Johannessen, Gothenburg 2003.
b) After 15.Bc5!? Kf7 (Heinrich-H.Martin corr 1971) Greet gives 16.Rfd1 when Black still is far from equality.
13.Rd1 Qe6 14.Nd5
14.Be3 h5 (14...g5 15.Nxb5 axb5 16.Qh5+ Qf7 17.Rxa8+ Bxa8 18.Rd8+ Kxd8 19.Qxf7 gave White a clear advantage in Spassky-Taimanov, URS-ch Moscow 1955) 15.Bc5 Qg4 16.Qxg4 hxg4 17.Bxe7 Bxe7 18.Nf5 g6 19.Ne3 Bd6 20.Nxg4+= Langeweg-S.Johannessen, Beverwijk 1965.
14...Nxd5 15.exd5 Qf7
After 15...Qd7 16.Nf5, Greet concludes that White is slightly better. This seems to be supported by the game 16...a5 17.c4 a4 18.bxa4 bxa4 19.b3 c6 20.Rxa4 Rxa4 21.bxa4 cxd5 22.cxd5 Qxa4 23.Bb2 += Chambers-Draba, corr 2002.
So far Greet's coverage has been quite convincing, and now he just manages to get his main analytical point right as he informs the reader that Anand recommends 16.Nf5!, when e.g., 16...g6 17.Nh6 Bxh6?! 18.Bxh6 offers White the better chances.
However, for those wishing to see the conclusion of Anand-Agdestein Baguio City 1987, it now gets quite confusing as Greet claims that the game continued 16.Be3?! Be7?! 17.Nf5! Rd8 and now the impossible 18.Be3. If Greet instead had given 18.c4 g6, it would at least have been a quite reasonable transposition back to the actual game, allowing the reader to make sense of the rest of the score. However, according to BigBase2007 the game actually continued 16.c4?! Be7?! (16...g6!) 17.Nf5! Rd8 18.Be3 g6 (18...0-0 19.Bh6!) 19.Nh6 Qg7 20.Qg3 with a clear advantage to White.
So - how bad is this for the book? Well, in my opinion it mainly raises some questions about Everyman's proof-reading routines. But it isn't really an analytical problem, as it occurs after the analysis has been concluded. And a reader who tries to play through the game and gets confused should be able to find the correct gamescore in a database without too much effort. So in itself this must be classified as an ugly but still quite minor blemish in such a monumental work. The real question is how many problems of this kind will turn up after a more detailed scrutiny of the book. My preliminary guess is that there will not be too many as it appears to be quite well researched.