I agree that without history we want to be tighter than normal against huge 3bets, typically I see these larger sizes from very strong ranges. I like folding AQ here, and might even shove as tight as JJ+, AK.

Plugging this into ICMizer the answer depends a lot on villain ranges, but AQo is rarely doing great (all calcs made with 10% antes):

If CO opens 22.6% and calls 99+,AQs+,AK while BTN 3bets 11.3% and calls 88+,AJs+,AQ+  chipEV = -0.08bb

Tightening range for CO open to 21.2% and BTN 3bet to 8.3%  chipEV = -1.64bb

Loosening CO to 26.7% and BTN to 13.7%  chipEV = +0.75bb (33+,AT+,KQ and suited Broadways are profitable)

Keeping loosest ranges with wider call range of (CO) 88+,AQ+ and (BTN) 77+,AJs+,AQ+  chipEV= +0.41bb 

Keeping loosest ranges (but original calls) and changing BTN 3bet size to 480  chipEV= +0.18bb 

Lowering our stack size to 24bb it becomes much harder to make realistic villain ranges that have AQo losing, but we’re unlikely to have a chipEV> +0.3bb.

I could play around with this all day, but I think that gives an idea of how each type of change affects our profitability. Some conclusions:

– Even with a hand as strong as AQ, most of our EV comes from fold equity

– 11.3% BTN 3 betting vs CO is somewhat rare at low stakes, 8.3% seems much more common

– With a more standard 3bet we need BTN to be wide and overfold for AQo to break even

– The call ranges I used vs our shove are pretty wide, we probably have more fold equity than my calculations

– I need to cold 4bet AQ much less at 25-35bb, I think I do it more often than not in this exact situation.

– If we are in BB vs SB 3bet and BUT open, those ranges likely can’t call enough to make shoving AQ bad.

I think the most likely villain ranges are close to:

CO opens 25.2% and calls TT+,AQs+,AK while BTN 3bets 8.14% and calls 99+,AQ+

if 3bet is 720, AQ  chipEV=  -0.04bb 

with a 480 3bet, AQ chipEV= -0.47bb