Garech: <Mturaga>
Really good effort and analysis, this is great for the team - thank you.
I have a few suggestions:
In your first line, after
<(A) 16…Qe7 17. Qb3 bb7 18. e4 g5 19. e5 (here engine room 2 (ER2) variation says 19. Re1 which I shall analyze next) 19…Bg7>
black must give serious consideration to playing f6 at some point, with counterplay, instead of sitting passively.
In the diagram you posted after
<White is a pawn plus and may look forward to the end game with optimism. So 16…Qe7 need not be feared in this variation.>
Black has a lot of positional trumps as compensation for the pawn, I.e. complete control of the a file, bishop pair on a fairly open board and potential to play the pawn sac Ne4!? leading to very sharp play. In the hands of a GM the black pieces could be very dangerous in this position.
In your next variation, with:
<Now coming to the following variation suggested by ER2: 19. Re1 (A2) 16…Qe7 17. Qb3 bb7 18. e4 g5 19. Re1 (suggested in Engine Room 2) Bg7 (here ER2 suggests Rb8 but Fritz10 does not even consider it) 20. e5 c5>
black is allowed to play c5 - this must be stopped! Also, again here black has f6!? with counterplay. IMO black is plaiyng too passively and planlessly in these variations. My engine is tied up so I cannot examine the bishop sac, but I'm sure black has strong ideas to play in this position.
Personally, I doubt we will see 16...Bb7 (your next variation) - maybe tonight we will find out. In this line you posted
<16. b4 Bb7 17.c5 a5 18.a3 axb4>
18...axb4 doesn't achieve much. It's better to double rooks on the a file (or at least play Ra8) before making this capture. This encourages white to play cxb6 instead, which is weakening.
Also in that variation, <24. Rxa3 Rd8> makes little sense if white is allowed to play e5. It would be better for black to play dxe4 and then Rd8, but this is not good either. There are more useful posts for the rook than d8 (unless black is trying to provoke e5 - but I don't see why, because the main idea of provoking this is to play f6 - which makes less sense if the rook has moved from f8).
Your next line:
<26.Nb3 Rb8 27.cxb6 Rxb6 28.Nc5 Rb8 29.h3 Bc4 30.Qc3 Be2 31.Ra1 Bc4 32.Kh2 Bb5 33.Ra2 Nxc5 34.dxc5 Bc4 35.Ra3 Bd8 36.Qe1 Bg5 37.Nd4 h5>
Features some strange play. ...Rb8 27.cxb6 Rxb6 simply would not happen. cxb6 is a terrible move and black would answer Nxb6 in order to a) not give white initiative (with Nc5) and b) exploit the c4 square.
The remaining play again seems planless. Why isn't black targeting white's weak b4 pawn - backward on an open file (another argument against cxb6 - probably the main argument, in fact).
Also, instead of "pushing the kingside pawns" to gain counterplay, black should be challenging white's a file control (which white shouldn't have anyway!) and gaining heavy piece activity on the queenside. For example in the diagram beneath your comment:
<Phase 3 completed.
White’s knight is superior to black’s bishop and White’s attack on the king side is under way. Black’s pieces are tied down to the back rank and are passive>
black should play Ra8 instead of Bg7. Like I said I can't check this with an engine (or even a board, at the moment) but as long as this manoeuvre is safe, given the kingside situation, black should defintely be doing this - or at least *something* active.
Which engine are you using? Perhaps you have the settings on "defensive" or something?
In your final two diagrams, of course white is winning - but there is no way black would let this happen, especially when black is GMP with a team of analysts and engines.
Sorry - I don't mean to shoot down your hard work, but I have to be blunt - black has much better play than what you posted.
-Garech
Kibitzing
A majority vote for 1-0 means we resign. 
, maybe 
] 
