A friend wanted to paint my coro props and just used a brush and standard acrylic paint which, no surprise, didn't stick very well and was coming off as I pushed the pixels. We then put an anti-graffiti coating on and had no more issues with it chipping etc. I suspect the anti-graffiti coating soaks through the acrylic and sticks to the coro because even though the acrylic was dry it became a little 'runny' till the anti-graffiti clear coat fully dried.
So, after a season of working with some painted coro it held up pretty well. Not very flexible and cracked where there was a lot of flexing and a little chipping while pushing pixels. Considering most of my props are sitting in rain, cold (10 - 30 F) or snow I was very happy.
Just a thought and not sure how practical it may be but what about the automotive paint used for the 'rubberized' bumpers cars have now-a-days? Guess on my part but I suspect there's an additive to make/keep the paint flexible. Regarding adhesion, I scuff the surface using a fine ScotchPad/steel wool or 400 grit sandpaper typically.
What was the drying time between the coats? The manufacturer dry/'cure' times stated between coats are I've found very dependent on temperature and humidity. Also, how thick were the coats? For a flexible surface thinner is better I think.
I've been rubbing mine with fine 2000 grip sand paper to take the shine off, then cleaned with 100% isopropanol, spray and haven't had any dramas. Sometimes it does chip away when bumped or pushing the odd pixel but it's unnoticeable from a distance. The gold paint comes out horrible though, it really shows the lines in the coro. QLD here so we have plenty of heat, but at the moment it just seems to be raining more than sunny.... Sunshine coast my left foot!
I'm about to do some more and might use a gray primer this time, see how that goes.
I've never pre-cleaned (just brushed off noticable dirt) and simply apply the Rustoleum directly without any primer layer.
No sanding in between, just 1-2 coats as required for desired look.
So the bubbling may be the temps as suggested earlier.