I thought I would share my ICM analysis on this situation. ICM is one of the advanced tournament concepts that I think I understand in theory but haven’t dug into in practice to the point where I feel fluent. Let me know if you spot any errors.

I used the ICM calculator at icmpoker.com to look at the three possible outcomes when the big blind villain shoves. a) hero folds. b) hero calls and wins and c) hero calls and loses.

If hero folds, the stacks and $EV will be respectively
BB – 77K / $1963
SB – 131.5K / $2361
CU – 89.5K / $2068
BU – 342K / $3178

If hero calls and wins, the stacks and $EV will be
BB – 77K / $2343
SB – 0 / $1210
CU – 89.5K / $2430
BU – 473.5K / $3582

If hero calls and loses, the results are
BB – 77K / $1920
SB – 233.5K / $2802
CU – 89.5K / $2022
BU – 240K / $2826

If hero calls and wins, he gains $404 in $EV ($3582-$3178). But if he loses, the hit to $EV is only $352 ($3178-$2826). Since the reward is greater than the risk, hero needs less than 50% equity to call. The the break even point is actually 46.6% (352/(352+404))

I was initially surprised by this result. One central implication of ICM theory is that when two big stacks collide, the players who gain the most are the ones not in the hand. I expected ICM to conclude that hero should be much more conservative in calling. In ICM analysis the difference in equity between $EV and cEV calls can be dramatic.

What I failed to consider is the great benefit to the chip leader of busting the second biggest stack.
Knocking out any player guarantees that hero at least 3rd place. Winning this pot gives hero a huge advantage at the top prize with 3/4 of the chips in a 3-handed game.

Since overbets like this are rare, it’s hard to estimate villain’s range. But with a wet board, I’m assuming top pair is definitely good enough to call.

I called villian had K2 for pure air. He didn't improve and I had 75% of the chips with 3 players left.