The game started off as a French Defence with an interesting Sicilian style 2...c5 move by Black, which I hadn't seen before. I decided to grab some space with 3.d5 and made a slight blunder by responding to 3...Nf6 with 4.c4. I thought this would solidify my pawn centre, but missed the obvious 4...Nxe4.
Fortunately, Black also missed the free pawn (or perhaps chose to ignore it) and followed up with 4...b6. This was followed by a quiet period where I played some natural looking development moves while Black chose to advance Queenside pawns and prepared to castle. The resulting position after White's 9th move was
My plan here is to mount an offensive on the Kingside with my lead in development before Black can make any progress on the Queenside. Black decides to play Nh5 in order to trade off the dark squared Bishops. In response I strike with g4, both attacking the now undefended Knight and starting a pawn storm against the Black King. After trading Bishops, the Knight hops back to f6.
Now I decide to play e5?! Perhaps a dubious move in hindsight with computer analysis. The idea was to open the b2-h7 diagonal for my Bishop and provoke the Knight to vacate the f6 square, leaving Black vulnerable to a future Bxh7+. I was also concerned Black might play e5 first, locking in the centre as well as my light squared Bishop. All things considered, after 12.e5, the position is still fairly equal. Black has to be careful not to go wrong in the defense, and White has a clear plan going forward.
After 12.e5 dxe5 I play another dubious move (according to the computer), 13.h4. During the game I thought this move felt consistent with the overall plan of throwing pieces towards the Kingside. As the h and g pawns move forward, my Rooks become stronger and take aim at the Black king. Although the computer gives Black a slight edge here, it certainly isn't clear how to convert that edge into a material advantage. In response, Black plays h6 in an attempt to shoo away the annoying g5 Knight.
Now things start to get interesting! I realised that the threat of taking the g5 Knight is actually non-existent - after recapturing with the h-pawn I will have succeeded in opening the h-file for my Rook, and suddenly Black's King feels extremely exposed. This is of course made worse for Black given his dreadful development. Since I do not need to retreat the Knight, I chose to castle instead. This completes my development - the d-Rook is now perfectly placed to contest the d-file should it open up, and can also swiftly join the action on the g-file if necessary. Black finally decides to grab the g pawn now that his h pawn will not fall to my light squared Bishop, however he has chosen to open a line of attack against his own King. I responded with the natural Rdg1, but can you find the much stronger move?
Unfortunately I missed the above tactic, but the Rook move still gives Black a lot to think about. Black responded with f5, to block the light squared Bishop (of course the Knight on g5 is still of limits because of the pressure on Black's King). Perhaps better for Black was exd5! cracking open the centre, but more importantly bringing his Bishop into the game. I chase the g4 Knight away with the quiet f3. Then I take the pawn on e6 with my Knight to claw back some material. He takes back with Bishop, then I make a blunder - taking on h6 with the Queen.
I thought that it was a free pawn while threatening mate, and that Black has to spend a move defending the mate, that then allows me to recapture the Bishop (an intermezzo tactic). However, I failed to notice that after the recapture of the Bishop, my own d3 Bishop is no longer protected by the Queen, and Black can swoop in for the capture (see alternative line in diagram above)! Fortunately, Black missed the defence of the mate with the Rook, instead playing Qe7, allowing me to win back the Bishop.
At this point the material is equal but White has the overwhelming positional advantage. There is too much pressure on Black's King, two Black pieces are out of the game entirely (having not moved all game!), and all of White's pieces are on good squares. The game ends a few moves later.