The standard deviation of a HU SnG is just SQRT(wW^2 + lL^2) where w is chance of winning, W is amount you win, l is chance of losing and L is amount you lose. If you're a 50/50 player, for example, the SD of a 10.50 HU SnG is ~$10.01.

For a ten-person SnG, your SD is SQRT((w1)(W1)^2 + (w2)(W2)^ + (w3)(W3)^2 + lL^2), where w1 is your chance of coming in first and W1 is what you earn for first. If you're an average player in the $11s, your SD for one trial is $16.76.

This effect actually increases as a player becomes better (which is why you need less than 5/8 the BR to play HU SnGs). I've really gotta get back to work, but you can play around with these numbers and see what I mean.