[Stumpy]

I'll start with the 99+,AQ+ range for CO.

MP2 calling is always a bad thing for you.
I'd say he'll call with QQ+, AKs. You're 25% against that range.

When he calls, the side pot is 39,614 each:
39,614*(0.25) - 39,614*(0.75) = -19.6k

The main pot is: 71,380. Against QQ+,AKs and 99+,AQs+,AQo+ you're 20% to win.
71,380*(0.20) - 24,490*(0.80) = -5.3k

So when he calls, you're losing 25k on average.

If his initial call was pretty wide, like top 15% of hands, then he called initially with about 200 hands.
QQ+,AKs = 22 hands, or about 10% of that 15%.

So 10% of the time you lose 25k, or -2.5k on average.

You need to make that up the other 90% of the time. That means you need to win 25k*(0.1) / (0.9) = 2.8k on average when heads up.
To win 2.8k, you need: 2.8k = 46,890*wp - 24,490*(1-wp). (46,890+24,490)*wp = 24,490 + 2,800 = 38.2% to win the heads up pots.
You're 35.6% against 99+,AQ+. You'd be 39.5% if his range was 77+,AQs+,AQo+.

I would say that's reasonable. Basically it means he just has to be a little light once in a while.
If MP2 is much tighter, it gets ugly, because he's calling a bigger fraction of the time, and that -25k is a bitch to offset.

You can assume MP2 would not always flat call QQ+,AKs, so that would discount the chance he has those hands above. That would obviously help you. Overall, the numbers show it is very close, and your reads on how tight these two players are would be pretty important. Without that, I'd rather err towards aggression. Having a huge stack heading into the final table about 35% of the time is not bad at all. And when you lose you'll have 40k left at 2k blinds most of the time, which isn't horrible either.

The math to show shoving is better than calling is a bit trickier, and I'm not sure how to do it and have it actually mean anything. Reshoving looks really strong though, so MP2 will fold almost everything. Anything he'd be borderline on calling is really bad for you, like 99-JJ, AKo, so shoving seems better.

-Stumpy