cxd4 is the move. If anything can be learned from yesterday, it is that more of white's moves need to be analyzed prior to his move. We can't assume that if we analyze only what we think is the strongest move then therefore we will naturally be ready for (assumably) weaker moves.

He has a few responses:

cxd4

exd4

Nxd4

(post if you see anything else that is playable)

Luckily there has a good deal done yesterday on cxd4, and some work started on exd4. Nxd4 looks mistake-ish right now, but a deep analysis could prove otherwise.

ill grab what I can from last moves comments, and add where I can in a bit.

to get the ball rolling...

15...cxd4

16. exd4

16...Nd5 17.Qe4 g6 18.Qh4 Qf4 we're okay

16...Na5 17.Ne5  needs more analysis

16...Na5 17.Nd2 Qc6 18.f3 we're okay 

16. cxd4 

16...Nb4 17.Rdc1 Qb8 18.Bb5 

18...Rxc1+ 19.Rxc1 Rc8 20.Ne5 Rxc1 21.Bxc1 a6 22.Bc4 Qc8 23.Qb2 a5 24.Bd2 Ng4! 

"follows Oliveira-Helbich and I think we can do better" -Summerstorm

18...a6 19. Bc4 Qd6 20. Ne5 Nd7 21. Nxd7 Rxd7 22. Ba3 "white is better" -gmeillonc

16. Nxd4  a bad move in every line I looked at

16...Ne5 17.Bc2 Nc4 18.Bc1 e5  if he tries to keep the B pair it's a disaster

16...Ne5 17.Nb5 Qc6 18.f3 Nxf3!  the motif of picking up the Bd3 if the Rd1 moves to protect K-side

16...Ne5 17.Nb5 Qb8 18.Bc2 a6  and we invade with N to c4 and keep White blocked

here is what i could grab from the comments... i tried to give credit to who said it... everything is in one diagram right now, and if someone has a better PGN editor than the one on this side, it would be great to separate them into his three likely replies... (is there a good free pgn editor on the internet somewhere?)

cxd4 has the most lines, so some work would be good on exd4. Though, some more work on the main line of cxd4 (or what im calling the mainline... the one with Rdc1... would be beneficial... there is only one continuation line for it currently...
if Nxd4 can be absolutely proven to be weak as it has so far, we may just want to come up with a likely line (as drakpete has done already), a suggested next move, and focus our time elsewhere

The supposed mainline in the diagram in post #4 includes threats that end up leaving our pieces worse-placed. Eg the first move Na5 moves from a situation where both knights are normally developed to one where white's is central and ours is offside (and doesn't justify it). At the end (which we avoid), Nc6 seems obviously better than Nb3 which gets us into trouble on the next move. After cxd4 Nb4 Rc1 Qd6 Ba3 Be4 Bxe4 Nxe4 Qe1 leaves us in trouble, I think

@andrewlong I think after 16.exd4 Na5 he plays 17.Ne5 not Nd2 and then has a great centralised N that sees the c4 square (he could play c4 if he wanted) and also supports K-side attack (a more likely plan from Onischuk) pushing K-side pawns with LSB and DSB support

At this point 16...Nd5 looks safer

I agree with Vegetableman that exd4, Nd5 leaves us better off.  Seems like cxd4 is his best move after reviewing the diagram.  I'll see if there's anything I can contribute on cxd4 during lunch...

It's true Ne5 is the best move in that position, my later line then continued Bc2, with no contination from anyone. So it could have gone, 16.exd4, Ne5 17.Bc2 Nf4? 18. {Queen moves}.

The theats seem mostly gone from that move when the fork is impossible.

The intention of the line is to see what white *might* have in that line.

Please feel free to ask if you don't understand what I am talking/diagramming about, thanks for mentioning my line, I realize it is very unrealistic.