Rob,
I've only seen one statistic from AS: AS successfully places 99.9% of snipes using a 5 second lead time (1 out of 1,000 isn't placed). AS' considers a snipe to be successful if it gets placed (barring bid increments). The actual "winning" of a snipe, is dependant upon how much the sniper is willing to spend, which is outside of AS' control.
Would a "win/loss percentage correlated with snipe times" be useful? How can a 5 second lead time without a sufficient bid amount be compared to a 15 second lead time with a sufficient bid amount, or the reverse?
I might suggest another comparison - how much does sniping tend to reduce the closing price of an auction. Difficult to impossible to quantify. I think that most of us "feel" that we are spending less on a auction by sniping. If we didn't, I doubt we would be sniping and I doubt we would be paying AS a fee to snipe for us.