How much lower is the Std Dev of AIEV than bottom line earnings?

NOTE TO MODS: Sorry for the very general question. I did my best to search the forum overall, this sub specifically, and the "stat checkup" containment threads, and couldn't find much discussion on this. Feel free to move, merge, or link me to other threads/subs as appropriate.

In extrapolating results, I never know how to account for AIEV. Since AIEV removes one (relatively small) dimension of variance, it stands to reason the standard deviation would be lower when using that number instead of your actual winrate. So for example, when using a poker variance calculator, instead of plugging in 100 as your standard deviation, perhaps you could plug in 90, which would allow your results to converge on your true winrate over a (slightly) smaller sample.

I understand intuitively that it is far less than half of poker variance but is it 75% as variable? 90%? 95%?

I presume stack depth has a large effect here. At 10bbs deep with tons of flips, I imagine AIEV could actually make up fully half of your variance, whereas at 1,000bbs deep I imagine there are far too few AIPF and AI OTF scenarios to make a meaningful difference. Let's assume standard cash game stacks (100-200bbs deep).