Gilmoy: After <14..f5> Black seems to have a fine KID-like position. <15..Nf7 16..Bh6> is an intriguing idea to trade off Black's "bad" DSB. So where does he go wrong?
White doubles on c, Black storms K-side. <24.d4> shows the <strength> of the doubled pawn: they can hit a pawn chain <twice>.
Subtly, it also outraces Black's desperate <24..g4> just-in-time: now Rc3 defends f3 sideways, and White doesn't fear 25..gxf3+ 26.Rxf3. Now 26..fxg3 just favors White: 27.e6 Rxf3 28.exd7 <Black has no time to capture White's Q> Rf2+ 29.Kxg3 Rd8 30.Qb2+ <saving the Ne2> K<any> 31.Rxc7. And after that disaster, Black's Nh6 <still> has no useful move to join the attack.
The threat of e6 forces Black into bad pawn trades, ceding the long diagonal. On top of that, White plants a <rusty Ne6 in the knee>, concludes a successful minority attack, and still has the KID heavy-pin on c. That minority attack is already winning by itself, as simple liquidation is ending with Rc6 and Qb8+ (or Rc7 if Black abandons c).
<31..gxf3+> is the final throe: it only opens g for yet another inroad. Black might have wanted 31..g3 just to lure White's K forward and hope for a perpetual, but the Q-side has eaten all his tempi. <32..Nf5> exposes his "intriguing" idea as a loser -- this N was so lifeless that it could only offer to sac itself -- and even then, it's so weak that White would <rather keep his pawn>.
Black wanted Rg8, but White's double-on-c prevented that.
Black wanted Kf8, but White's inevitable <28.Ne6> prevented that.
Black had a focal point on g5, but White neutered it by simply <not> capturing there.
<19.Qc1 Kg7> was a painful concession: Black's N had <no escape>. White uses the tempo to lubricate his Q-side expansion.