Brian has a provisional rating after four games, and I seriously doubt that his actual strength is reflected by his rating.
I generally have bad luck against this type of opening schema online because at faster speeds it is like a psychological opening/ploy where you don’t want to trade your Black bishops for White’s knights.
3…c6 was unlike me to play, but I did it because he hadn’t committed his Ng1 yet, and this Qd1-h5 diagonal was still open for his king, so I was playing the opening shrewdly.
5…Ne7 continuing with the plan of a possible future …f5 push.
8…0-0. I didn’t even want to concern myself with possible Nxe4 tactics, although they lose, because I don’t want to spend too much time in the opening even at this G/90, Inc 30 time-control.
9….b5 Played to keep him honest, to prevent him from castling queenside, but he chose to anyway.
10….Nd7 I passed the strategic “psychological dilemma” test by not freaking out after giving up the minor-exchange (bishop for knight). I looked at 10…Bg4, 11.h3 Bf3, 12.BxB exB, 13.e4?, not realizing 13…b4 would give Black an advantage. But if I should learned anything from Paul A. by now it’s that it’s all about simple chess – 13.Qd2-d1 followed by 14.Qd1xf3 picks up the pawn.
13…Bb4 My blitz-move was going to be 13…b4, but I didn’t see the follow-up after 14.Nd2? because I figured …c5 would then open up the position for most of his pieces, missing 14…Rxf2. This type of move, wow, I really shouldn’t be missing it in my calculations.
14…Nb6 During the game, I did see 14…BxNc3, 15.BxBc3 Nb6, 16.Kd2 Nxa4, 17.Ra1 (thinking he’d get his pawn back, but 17…Na4xc3 takes care of that). Actually, I thought this line was most prudent, and that the computer would confirm this, but at the end you are left with a more strategic position than a mating attack, and I didn’t see the follow up after this of …e5, and if d4xe5, then …Ne7-g6xe5. Often, OTB, it’s like you can choose to evaluate a position well, or you can find the lines correctly, but it seems difficult to stay objective enough, and have clock-time to do both.
18…Ba3?? I could blame that I was playing on four hours sleep, but I drank a Dr. Pepper during the game and felt great by this point. My “blitz move”, from the previous move, was to follow-up with 18…bxc2, but I didn’t notice that after this move that the Rc1 is attacked. Possibly, some mental fatigue caused me to economize my thinking process, since I was looking at 18…Rxf2, and must have noticed for a moment that 18…bxc2, 19.Rdf1 would defend the f2 square. Naturally, I only played this mental trick on myself to examine the variation 18…Rxf2, followed by 19…Rfc2+, but I should have “erased the mental chalkboard” better after writing off 18…Rxf2. If 18…bxc2, 19.Rdf1, then this was a free-move for Black, and there is no need to save the possibility of …Rxf2 as a resource for Black, as Black has enough pieces on the queenside to attack the king with, and should be trying to keep the kingside bottled up to prevent counterplay. After 19.Kxc2 comes …Rxf2+, and after 19.Rd2, …Nc4 will hit that piece with tempo.
The more remarkable part, after 18…bxc2, 19.Rdf1 is where Black continues the attack by retreating the bishop to either a5 or d6. GM Daniel King said he likes how GM Michael Adams makes defensive moves during his attack, and this is a great example of one. Another such defensive move could be …Rf6-h6 at some point, which would allow Black to continue defending the e6 square from the bishop, prevent a rook trade on f6, and be played with tempo as it attacks the Ba3.
LinuxGuy_on_FICS 
His actual rating is probably around 1200.
White’s queenside castle was a mistake of course.
I like 18… Rxf2, if he takes on b4 then Rxc2+ and knight goes to c4.
Too tired now to see all the consequences.
I just looked at it again and couldn’t see why that wasn’t winning either, although that’s a harsh assessment, I saw it more as unclear, but that Black should be winning somehow (Black is very winning according to Houdini, in your line). However, after 18…Rxf2, 19.cxb3, you don’t have to analyze any further in the line you gave, since it won’t appear unless White messes up.
Incidentally, this position is difficult to blindfold since at the moment it has an irrational pawn structure. 19.cxb3 is easy for the human mind to miss, we think that pawns are stationary targets, but they aren’t always. The pawn structure is the most important thing, yet hardest thing to focus on. Also, there is probably an implicit bias in stronger players to look at piece-moves first. We examine winning piece move continuations, but ignore that the lowly sluggard pawn will be the one, defensively, to have the say in this position as to the outcome of the game, yay it is even more imporant, this pawn’s move, than the captures on the board, because it changes the entire of structure of the king’s defense and the attacks against it. It seems weird to the human mind that this would trump material considerations, and it even trumps mating attacks because it calls off many mating attacks.
I figured, OTB, that he would probably go wrong in complications, such as in playing Qb5?, which I correctly figured was a blunder. From looking at a lot of GM games that were decisive, I have this feeling that people often lose it, defensively, in these sort of situations, and don’t find the best moves. It’s easier to find best attacking moves than defensive moves, if the position is crazy enough.
I posted my last game.