I think a lot of the cloudiness is from absorbing water, if not from the atmosphere then from condensation, and like the info on Parks says, water contamination can impact how the oil behaves. Just like starting your car in the winter, driving only so far before it warms up and shutting it off is bad for your oil, you want to get to operating temperature to boil off the condensation, or it turns into a sludgy mess.
I think, but don't know, that if your oil, regardless of what it is, becomes very cloudy and it's not silt/solid particulate, heating it to 220F for a while to boil off any water contamination would go a long way towards "making it right" if not "like new."
But the other side of that is that if you're not noticing deteriorating performance, does it matter? If you're relying on initial test values for your heat treating processes, ie, once upon a time with a new bucket of oil and fresh thermocouple etc, you heat treated 1500f and quenched in your oil and tempered at 400 and got 60RC, but since then you just heat-quench-file check-temper and do not RC test, are you sure you'll catch deteriorating performance 1 lost RC point here or there?
So you can "play it safe" and just change your oil out to the manufacturer's recommendations, which are likely conservative. Or you can RC test periodically to verify your original process is still meeting expectations. But I don't think it's smart to both keep going with old oil and without some kind of process verification like RC testing, as variable as the file check is.
*not directed at anyone in particular, just my thoughts on the entire subject regardless of oil type/age. Processes that were once shown to be in control left to deteriorate without further verification will fall out of control, if not from a single major variable change, then from the stacking of multiple minor variable changes.