coon74's thread reminded me that I've been meaning to start a topic about this for a while, but don't want to hijack that thread. As long as I can remember, I've played the Bayonet against the KID. I have good results with it, but am interested in adding in some variety. I've noticed both 5. h3 and 6. h3 (after 5. Nf3) are gaining popularity since it's less explored. KID players, how do y'all feel about facing this? Anyone here play this as white?

Yeah, it's a threadworthy topic in its own right

I'm actually playing this as White in one corr game right now, shall post it for analysis when it finishes (you can look it up in my game list but obviously can't kibitz).

I'm examining Chessexplained's games in the line, and it seems to me that Black's attack against White's queenside castle is usually too slow, and the c3 knight and the pawns are quite sufficient for White to fend it off while throwing pawns on the kingside.

I liked Black's counterattack on the e-f-files in CE's blitz #1395, though.

CE also faced an exchange sac Re8xBe3 in one of the games when he was castled short, and he said the sac was standard (giving Black play on the dark squares); alas I don't remember the number of the blitz, would be grateful if someone listed it.

X-post for convenience:

Actually, CE has a short theoretical video on 5. h3 and 5. f4 KID as part of his former White 1. d4 repertoire (he's switched to the 1. Nf3 camp since then but surely Nf3 and d4 fans end up in almost all the same openings).

CE is actually obsessed by the Exchange Variation - 5. h3 O-O 6. Nf3 e5 7. dxe5 (the principal move is 7. d5) dxe5 8. Qxd8 Rxd8 9. Bg5 (Δ Nxe5; immediate 9. Nxe5 would be met by Nxe4) Re8 10. Nd5 - especially when he's tired and wants to 'keep it simple'

But 9... Na6 10. Nd5 Rd6 scores well for Black (9... Nbd7 10. O-O-O Rf8 is somewhat decent too); I wouldn't say it crushes, but Black seems to have few problems.

Btw, as you might have noticed, Na6 (preferably early and followed by e5) is a typical and recommended move for Black in h3 KID (on c6, the knight would block the key c-pawn).

A good news for White is that this seemingly bizarre move is unlikely to be played by an unprepared opponent, who will usually block the bishop by Nbd7.

I wonder why CE (and I guess other people too) recaptures with the c-pawn in the Benoni transposition 5. h3 O-O 6. Bg5 c5 7. d5 e6 8. Bd3 exd5, which is especially surprising given that the bishop has just been put on a square where it's going to be control only the queenside for long, and then he castles kingside. I'm aware that the e4 pawn prevents Black from playing f5, but White seems to totally give up thoughts of a kingside attack. Why is cxd5 followed by g4, Qd2 and O-O-O worse?

The game didn't wait for long to finish I got totally owned on the a1-h8 diagonal (underestimated Nb6(-a4) and Qf6). Exchanging pawns on b5, opening the a&b-files, would have sucked too, but a bit less. 21. Nf2?, cutting the queen off the king, was a decisive blunder, played on massive chess tilt (after getting owned by computers in blitz) after a sleepless night.

So you get an idea how precise the play should be from both sides in this variation

All right, I thought 17. Re1!? b4 18. Nd1 could cement the Q-side at least for a while, but alas 17... Nb6 would force the exchange on b5 anyway. What Houdini 1.5a suggests is actually 17. f5!? Nb6 18. g5 Bxf5 19. gxh6 Bxc3 20. Qxc3 Qf6 21. Qe1= - attack is the best defence, and Black's counterattack would take long to develop.

A questionable point is that I delayed developing the LSB, unwilling to lose a tempo like it often happens in QG where Black waits for the B to move before taking on c4, and also waiting until the dilemma between putting it on c2 or e2 gets a clearer resolution, but it looks like the B could be needed on d3 and eventually c2 or b3 for defensive purposes and to enable Rdg1 to protect the g-pawn.

The move with which I tried to introduce a novelty was 11. Nh2!? enabling f4. The only game I knew with that position was Laketic vs Bassis (2002) 1-0, where the engine move 11. Bd3 was played. Talk about the harm from deviating from a winning strategy.

But in that game, 11... b6 12. Nh2 Ne5 was played, forcing the B to retreat to e2, and my 11. Nh2 and f4 were designed to prevent just that; instead, the engine suggests castling short (and hence giving up thoughts of a mate attack) after 11. Bd3

Man, lots of good stuff in here. That first link you posted with all of the material is fantastic, I'm going to have to check it out. I owe you one for all of this stuff.

Is Nh2 theory? That seems slow and your position gets vey bad quickly, I would also think b3? is a mistake.

There's almost no theory in that position; in its only game Laketic-Bassis linked to above, 11. Bd3 was played, but the 12th move was Nh2 (followed by f4 soon) as a response to the slow b6. Probably the move order there was vital. I didn't like it that Laketic made too many bishop moves, and I decided to keep Black's knight away from e5.

I still think that my biggest mistake was allowing b4 without being active enough myself.

I was already on tilt (hence moving too fast because I generally feel conscience for making my opponents wait for too long) while making the 19th move; b3 seemed a solution against the nasty Na4, but the Bxg4 threat was more devastating, and I didn't get as much comp for the pawn as I hoped (i.e. b3 failed to shut down Black's queenside counterplay).

Fwiw, the computer lines are 19. h5 Kg8 20. hxg6 fxg6 21. g5 Na4 22. gxh6 Bxb2+ and 19. h5 Bxg4 20. hxg6+ fxg6 21. Nfg5+. It's very unclear, the engine keeps alternating between the two as I'm writing.

More precisely, it's spitting out a perpetual: h5 Bxg4 hxg6+ fxg6 Nfg5+ Kg8 Ne6 (wow, I didn't see that ) Qd7 Nxf8 Qa4 Rg1 [Re1 is met by Qxa2 obv] Bxd1 Nxc5 dxc5 Nxg6 Bf3 Bxc5 Qxa2 Qc2 Na4 Ne7+ Kh8 Ng6+ Kg8 Ne7+.

Edit: or a perpetual for Black h5 Bxg4 hxg6+ fxg6 Qg2 Nf6 Nfg5+ Kg8 Ne6 Bxe6 dxe6 Nxe4 Qxe4 Qf6 Rh2 Na4 Rxd6 Nxb2 Bxc5 Na4 e7 Qa1+ Qb1 Qc3+ Qc2 Qa1+ Qb1 Qc3+

All in all, it becomes dynamically equal :/