Recently I read a thread about differences between the Sicilian and the KID vis a vis White attacks on the KS. Comments were made about the attack "playing itself" with f3, Be3, Qd2, g4, h4, etc.; I was never clear, either, exactly why this worked in the Dragon but not in the KID. In any case, I've been fooling with my travel chess computer trying to work this out, and it seems like White is always slower than Black when it comes to opposite side castling and attacks on the enemy king. The following board should illustrate the point, but note that while the position which develops is correct, the initial move order may be slightly different from my game, so if it seems retarded that may be my fault:
And now things get messy on the White kingside very quickly. The rook is going to c8 and the queen to a5, among other things. White simply doesn't have time to defend properly and play a bunch of pawn moves on the kingside. Does he?
Note also that I played Bg5 instead of Be3. Was this a mistake forcing d5 rather than cxd5? It seems to me that I played Bg5 after having problems with Be3 but can't at the moment recall what these were and don't have enough online time to reconstruct this. Anyway, that ought to be sufficient for one post to chew on.
Well, just for the sake of completeness I'll add a copy of my other game here also:
Today I decided to play it using Be3 instead of Bg5, and while it certainly didn't play itself as some claim, after a bit of trial and error I had good results against the travel chess computer. The move order is verified, but the order of the first few is a little unusual because full book was turned on; however, the game quickly settles out along more regular lines:
The computer was set for an average of 1 minute per move. However, that figure is averaged over a significant number of forward moves, so in the initial stages of the game the computer often takes two or three times that. Eventually a point is reached where the average collapses and moves made under time-pressure occur: these may be relatively weak. The computer also processes a position during my thinking time on my moves, but if I move quickly and the computer is also under time pressure, that may result in an occasional weak move.
10.Nh3 I found through trial and error that this knight needs to be out of the way when the time comes for a pawn storm attack on Black's kingside. From here it can go to f2 or f4.
12...Qe8 Possibly one of the time pressure moves. In a second reconstruction the computer played ...Re8 here. The queen turned out to be misplaced and wasted time being repositioned.
15.Nf2 From here, the knight helps guard the g4 square. This turns out to be important. For example, Black soon has a queen and two knights aimed at that square, from which my queen on h6 could be attacked by a knight.
21.Nd5 turned out to be necessary to stop the action of the rook along the 5th rank, where my g pawn needed to move to drive away the knight from f6.
21...Nd7 This may have been one of the computer's quick (time pressured) moves. It certainly helped because now I didn't have to worry about the knight moving to f7 and attacking the queen once pawns were exchanged on g6.
I truncated the end since it's an obvious win for White. My experience was that with less than perfect play by Black, White may be able to get a good attack on the Black kingside, but only if details of development and defense are taken care of first. Those who attempt hasty pawnstorms against a strong Black defender may end up being driven on the defensive before they can consolidate an attack.