What should my move 12 be?

"Everything's been OK for white so far...???" Why did you back off of the Knight with 6 Bc4? That drops a tempo for nothing that should have been used to develop something: maybe d3, d4 or Re1--almost anything but Bc4.

Okay, but is the position itself after black's move 11 bad for white? Or is it (as it seems to me) a fairly even position?

And what should my move 12 have been?

And how can I help myself to see possibilities like that knight sacrifice? What's the pattern I should have noticed that would have alerted me to that black move?

I would give white an edge in this position because of the bishop pair and the number of strong pawn breaks available.  The pin on the knight is annoying, sure; your first priority should be to prevent ...nd4 ideas, and c3 does that while preparing the d4 break.  Following that, your bishop will be happy on g5 and both a4 and d4 are good plans (assuming black castles kingside).

In general, some factors to help you see when the sacrifice is strong:  black's knight can come in through d4, white does not have a second knight that can easily go to d2, white is castled kingside, black is not castled kingside.

Shoop, I believe that was extremely helpful.

It appears to me one fundamental problem was simply that I had no awareness whatsoever of the relevance of his distant knight on the situation around my king. Question: Do good players know things like the following off the top of their heads, or would it not really occur to them as relevant to good play:

"The square three diagonal spaces away from a knight can be attacked by that knight in two moves."

Is that an example of a simple kind of "pattern" that good players see fairly instinctively? Or is that not what is meant when people say the good players see "patterns" moreso than reading out board positions several moves ahead?

Also, can you say more about why the castling positions were relevant to seeing the strength of the sacrifice?

I am guessing your thinking, well the bishop is worth 3 and 1/2 points and the knight is 3. Yes and no.  I think one problem is trading off your knight for a bad bishop which you did at move 11. in a closed position, A bad bishop means that it cannot find active play because it is blocked by a center pawn. Look at blacks pawn structure and you will see at this moment the bishop is capable of nothing, so your knight is a superior minor peice. You traded a very active knight for a bishop that is absolutely useless in this position, if anything trading that bishop off is doing black a favor more than hurting him. The bishop is trapped, it would be good to let it live so black has more problems to deal with since that bishop is only in the way of your oppenent.

In open positions or positions where there few pawns in play if your going to trade off in that position you would want to trade off your knight for bishops if possible because the bishops have long range capabilites, in a closed position where many pawns are in play bishops have a hard time finding space and knights are superior so you would want to trade your bishop for a knight. When you trade keep this in mind. If you are forced to trade a bishop you might want to try to keep pawns on the board, when you have bishops trade down pawns to get an open positon again. I hope this helps. Taking the knight would have made more sense, especially taking the knight with a bishop.

Also white has a temporarily weak d3 square due to the pin on the knight by the bishop on h5 and . In this position I would try to break the pin by Qe2 then Qe3, and before that I would pin the knight with whites black squared bishop because it helps develope another peice and it is inferior to the knight. Ideally you would want to trade off the white squared bishop since it is the bad one, but the other bishop would be my second choice for a trade off either way in this position. and exchange it when they try to break it since as I said before knights trump bishops in this position.

Well, now you've learned the problem with h3/g4 to drive off a Bishop. I bet you won't do that again.

On a more practical level, keeping the Knight off d4 with c3 is probably best. Then Qe2-e3 breaks the pin.

Sadly, I don't think I have learned that problem. The bishop pinning the knight is something I simply don't know how to deal with. As far as I can tell, it's [i]guaranteed[/i] to mess up the pawn structure around my king--either by my using those pawns to drive it off, or else by trading for the knight forcing me to open my king's file and double my pawns.

This game hasn't taught me any different, though perhaps it should have. It appears to me my choices were either drive the bishop off with the pawns, or else accept an open king file and doubled pawns on f. My allowing the c knight to get involved was an additional problem here, but right now I'm not seeing anything in this game to learn particularly about how to deal with the bishop in front of my castled king. (Not saying I shouldn't see it--just that I don't see it.)

Well then, how about just adopting a general rule that you shouldn't mess up the pawns in front of your king? Do just about anything else.

 Sure--I already do have that general rule. But my problem is, I don't see how to avoid messing up the pawns when it comes to that bishop attacking that knight. This, it seems to me, gives my opponent total discretion over whether my pawns will be messed up. If I don't drive him off, then he has, at his sole option, the power to screw up my pawn structure by taking the knight.

How do people wiggle out of this?

In quite a few cases, it's ok to allow bg4 pins - either your king isn't castled kingside and you can chase the bishop with h3/g4 without endangering your king, or you have a knight on b1 that can come to d2, defend your knight on f3, and let your queen move to release the pin without messing up your pawn structure.  In this case, by move 12 the pin is removable (c3 followed by Qe2-e3 at some point, mixed with attempts to play in other areas of the board via a4, d4, bg5, etc).  However, it's often bad to allow these pins if neither a queen maneuver, a knight on d2, or a h3-g4 push is feasible.  That's why in most e4-e5 lines, the knight doesn't go to c3 so fast - it's nice to reserve the option of c3 and nd2.

(To answer your earlier question, it's easy to see that pushing the h- and g- pawns is more dangerous when they're in front of your king, but it's also more dangerous if your opponent's king is elsewhere on the board.  If his king is in the center or on the queenside, he can push his kingside pawns in response, and open up the area in front of your king even more without endangering his king.)

There was no real threat to that Knight that c3 wouldn't have prevented. It was protected by the Queen, so you could have just recaptured with the piece. Then the "unpin" manouver of Q-e2-e3 would have freed the piece from further menace and off you go to the middlegame. That's the wiggle.

I would probably have played Bg5 to remove the Knight guarding d5 and then move my light squared bishop there, winning a Knight assuming Black captured your other Bishop with his Queen. 

(Edit for typo, because Bb4 is illegal. Stupid keyboard. Also d5, not e5. FML)

This is what I think of it. g4 is a bit weakening but with accurate play I think it's playable. Especially if some pieces (the queens) come off the board.

Oh and on move 14.Kg2? 14. c3 prevents Nd4 and then the position may be defendable.

Move 9. Is one of the least pointless moves you made. It forces the black bishop to show what he is going to do. If he goes back to the queenside, it shows that he is not going to keep the pin, but if he keeps the pin, then it shows that he doesnt plan  helping on the queenside. note that in the 4 knights defense, you typically should play Bxc6 which results in doubled pawns for black. That is the main line. You didn't play the mainline which can result in problems. You should also note that the sacrifice that was played was unsound because after the knight takes on g4, then the pawn takes and the bishop takes, c3 would have prevented the knight from getting into the action. Kg2 was completely pointless because it didn't really defend because the Knight+Bishop=6 points while your defenders, the queen and the king=9+infinity so there attackers are worth less and you shouldnt trade. On move 12. I suggest that you play: Bg5. This develops a piece and pins the knight. your kingside is safe and there are no back rank threats because you played h3. You would have a solid position. It is usually a bad idea to be moving pawns in front of your king because the whole point of castling is to make defense and by doing that, you are getting rid of your defense, so castling would be pointless. that is a general rule, and of course if you have a good reason to break it, then you should. I hope I helped

12. c3 and white looks better.

14. c3 and white still looks better. 14. Kg2 was a blunder.