This is a messy, MESSY puzzle, and
it is ENRAGEING to include it in
any puzzle book! (giving alternative
move to Garcia's 24 Bxa6?)

1 Na5!

'1...Qa7 loses to 2 Rc7.'

No further comment here. Apparently
it is obvious that after 2...Re8,
that 3 Nc6? wins. Yet Black has
3...bxc6! 4 Rxa7+ Kxa7 with
sufficient material for the Q, indeed
one can go so far as to state that
BLACK is better here, with a strong
compact pawn mass, two potentially
very powerful bishops (notice White's
3(!) weak isolated pawns!) and moves
like ...c5 and ...f5 upcoming.

So it's 3 Qa4 then, right? (after
2...Re8) Doesn't that easily win a
piece? Yet Black has 3...Qb6!, when
after 4 Rxb7 Bxb7 5 Qxe8+ Ka7
6 Qxe7 Qxa5 7 Qxf7 Qe5 White also
doesn't have enough to crow about.

3 Qc4--? - trapping a bishop?
No, 3...Qb8! defends that. (Not
3...Kb8? 4 Nc6+! now wins.)

Nunn again:

2 Nc6 (the most convincing, altho
2 Be2 is also quite good)'

No comment from Nunn on Black then
playing 2...bxa5!
White still wins after 3 Qxa5 Qa7
4 Rc7 or, what is much more difficult
to see, 3...Qf2 4 Rc7! (Not 4 Bxa6?
Qa7) Qxe2 (4...Kb8 5 Rxe7 or 5 Rd3
is effective) 5 Qb6! Qxd1+ 6 Ka2
wins.
But was any of this obvious?
Was 5 Qb6! - saccing another rook
with check- obvious? Remember that
Nunn have no further commentary at
all after 2 Be2 !
I at first thought that 2...bxa5 had
been missed by Nunn. Actually, maybe
it was. But White wins anyway.
Remember that this is supposed to be
a puzzle. It is instead an actual
game, and a messy one at that.

Finally, after 1...b6 2 Nc6 Bb7
3 Be2 Qf2, Nunn gives 4 Bxa6, even
giving it an exclam, tho the move
does nothing more than desperado the
bishop after 4...Bxa6 5 Nxe7.
Of course, much better is simply
4 Nxe7 Re8 (4...Qxe2 5 Rxd6) 5 Rxd6!
(was this missed?) Qxe2 6 Qxb6,
winning, or 5...Rxe7 6 Rd8+ Ka7
7 Qxe7 Qxe2 8 Rcc8 b5 9 Qc5 mate, OR
5...a5! 6 Qb5 Rxe7 7 Rd8+ Ka7 8 Rcc8!
with mate or decisive material gain.