Alas! It's Not Up To Me: OSL Dating South African Caves

It seems almost foolish to say such things. But, if I ran the world (even if only the archaeological world), there'd be some calling up on the carpet during my reign.
     For example, I'd make sure that people claiming great antiquity for modern human behaviour in southern Africa were called out for using optically stimulated luminescence to date cave sediments. After all. The assumption of OSL is that the mineral grains subjected to stimulation were, at some point, exposed to the direct sunlight for long enough to allow all of the trapped electrons in their crystal lattices to have escaped, and the 'clock' effectively set to zero. Otherwise the OSL age determination is going to be older than the reality by some indeterminate amount. How, I ask, is it possible to know that grains found in a cave where, presumably, direct sunlight is at a premium could ever be presumed to have been exposed long enough? Complex 'correction' algorithms notwithstanding, how do they know? If I were the monarch of all things archaeological they wouldn't be able to report age determinations without serious caveats accompanying them. And, I ask, what's the point if you have to say, 'but they might be wrong?'
     Furthermore, given that, regardless of location, daylight gives way to night at least half the time, how do we ever know that mineral grains were exposed to daylight long enough to 'zero the clock'? Really. Half the lifetime of any given mineral grain has been spent in darkness. What are the odds that any given (especially) sand grain was exposed, eroded, deposited and buried in a single night? Happens all the time in the desert. No? Well, then it can happen to a cave or a rockshelter. And in the case of sandstone rockshelters, how do you ever decide what's a sand grain from somewhere else, or a sand grain that was exfoliated from the bedrock? Pray, tell me. 
     So, if I multiply an error-prone estimate by an error-prone estimate, what do I get? It very well could be an accurate age determination. But who's going to know?
     Help me on this one, folks! There are claims for modern human behaviour beginning upwards of 75 to 100 kya in South African caves and rockshelters. What happened between then and about 45 kya, when Europe and indeed the entirety of Asia including Australasia were sprouting human populations almost simultaneously, where no humans trod before? Tell me. Someone. Please. Why a 30 to 65 kyr lag? Hmmmm?
     I guess you'd have to call than an open question. Well, you might. But I think there's nothing open about it. I think the dates are suspect. Pure and simple. Suspect and then some.