Disagreeing with What Is Not Said

Continuing with the game I last looked at in The 'Fischer Attack', a few moves later the players reached the position shown in the diagram. Black played 15...gxf6, and Fischer remarked,

The best chance is 15...bxc3! 16.Ne4 Qb4 17.Qg4 Bxf6 18.Nxf6+ Kh8 19.Qh4 h6 20.Ng4, threatening 21.Nxh6 with a strong attack.

Commenting on 15...gxf6, Kasparov noted,

The first, but already decisive mistake -- to lose in a sharp Sicilian, in contrast to the 'Spanish torture', Black does not have to make many mistakes. However, Fischer does not attach any mark to this move, regarding Black's position as already bad.

In Kasparov's view, this means that there is a distinct difference between '15...gxf6 (best is 15...bxc3!)' and '15...gxf6? (best is 15...bxc3!)'. The first indicates that Black was already lost, but had a variation to push off the loss; the second indicates that Black was not already lost. It's not an obvious distinction to make based on a single missing '?'.

1959 Candidates Tournament (round 10)
Benko, Pal

Fischer, Robert
(After 15.Bg5-f6(xN))
[FEN "r1b2rk1/p3bppp/4pB2/n1q1p3/1p3P2/1BN3N1/PPP3PP/R2Q1R1K b - - 0 15"]

After making this remark, Kasparov repeated Fischer's variation from 15...bxc3! through 'threatening 21.Nxh6 with a strong attack', and added

Huebner categorically disagrees with such an evaluation of the position after 20...cxb2. Here are the variations with a few refinements...

I won't give Huebner's variations. It's sufficient to say they show that Black had good play after 15...bxc3!, and it's curious that Fischer had nothing to say about 20...cxb2. On the other hand, can we 'disagree' with what is not said?