Play appears to be very hard to develop on cross lock based systems. I submit for example my M16. My M16's bar has worn all the way and now impacts the opposite side of the handle. It, however, has absolutely no play in any direction, and can easily handle even hard spine whacks (many, many times).
A properly designed crosslock should compensate for the length lost during wear (going more diagonal, farther away from its liner of origin) by the shape of the wedge at the tang that impacts it. That's my theory anyway, and it seems to hold true, at least in regards to my M16.
I've used my little 2 inch urban shark liner lock for several years now, and abused it, and it has had 0 liner wear at all, that I can notice anyway. Still no play.
My para has been flicked very hard thousands of times by now, but it had slight vertical play new. I don't believe there has been any wear in the liner, but it's more difficult to say since you can't actually see the lockup, just estimate it. The vertical play is VERY small (though still bothersome in a 100+ dollar knife). I've been working on a theory to explain why it's so common place in compression locks. But I've disproven all my ideas. It could just be simple poor geometry, but I'm inclined to believe Spyderco wouldn't make that a consistant trait in their entire class of knives that use these locks (all of which are reasonably high end).
The liners are quite substantial and bigger than you might think. They're also embedded with two big bars running through it. I've used my para a lot and I've got no fears with it.
The compression lock is less friendly than the axis lock, and yes, it does take a little (not much) more skill to axis flick. The axis lock is still number 1, even among the axis clones, in user friendliness and action.