I think Nikos was referring to the PANFR line iirc. I think Black is equal in places in a few places, but play is rich - far richer (personal opinion of course) here than in the endgame variations after Qf3, etc. The downside is that they're a bit sharper and more dangerous - just depends on your disposition and inclination to study such variations.

my guess is 5...Bb6 6.Nc3 d6 7.Qe2 Nge7 8.Be3 O-O 9.O-O-O f5
(although I now see that 10.f4 fxe4 11.g3 may cause some problems)

another option is 5...Bb6 6.Nc3 d6 7.Qe2 Be6 8.Be3 Nf6 9.O-O-O Qe7 10.f3 (10.g3 O-O) O-O-O

Perhaps, but no one knows that yet. 

Maybe my taste is clouding my judgment, I really can't say. All I see here is an ending I'm unlikely to win against someone of equal strength! Again, my issue is not that this variation is not good for White or anything, just that I'm not in love with clarifying the position so early and reaching this ending regularly. There's just less play than I'd like when compared to 4...Bc5 5.Nb3!?.

After a quick think, Stockfish likes 19.Rg1 a5, presumably stopping b4 and looking to offload that a-pawn at some point.

Let's start with 9.Nd2 O-O-O. In my main ending, do you agree this is the critical line? What are you trying on move 17? There's at least 17.Be2, 17.Bd3, or 17.Kc2. Perhaps it doesn't matter that much. The latter seems the most flexible at first glance.

I generally agree with your definition. But there's some subjectivity to what "better moves" looks like to one person or another, and how difficult it is to play the position for either color depends on the person who's got to play it. Using a definition that requires some knowledge of "chess truth" to discern which side needs to play the better moves, etc, is weird. One might call a drawn, pawn up rook ending slightly better for White, but it doesn't matter if the defender playing Black knows how to draw it with literally no effort, i.e. he's studied and mastered this specific ending. That defender would annotate it a clean "=", and now since we have a convention that uses evaluations that aren't actually results, these two annotators have a useless and trivial disagreement! 

Anyway, I agree with Nikos, as there's really no better system in place for evaluating chess positions. It's an interesting topic to wane philosophically on over a few beers some time though. 



Let's not sidetrack what could be a nice discussion on the Mieses.