large-stack ITM all-in decisions

[joshzero]

It's down to 3 in your SnG and you are the chip leader.
You face an all-in bet from a small stack (the other player
has folded). Let's say that his stack is small enough
relative to the blinds that the all-in bet, or raise,
does not necessarily indicate a strong hand.

You are faced with pot odd of X:1, and your hand rates
at Y:1 against a random hand where X>Y, but not by much.

My question is, at what point will you not play this hand
because although the tournament chip EV is positive, the
overall real money EV is zero or negative?

Has anybody done a study mining actual PP data, plotting tournament chips at different points in the tourney vs. real money equity based on the actual tournament finishes? I think this might be potentially very useful data, as I imagine the curve becomes very flat when the tournament chips approach 1/3-1/2 the chips in play, and should greatly affect one's strategy down to 4-5 handed. I have some feeling for this when I play, but would like to see it quantified for, say $30+3 PP tourneys -- need lots of
data though. Did anyone ever plot this?

josh