That having been said, the total blocks found (on a global, DGB network basis) for scrypt is clearly lower than the rest for this 10 hour period. (I don't have block explorer info going any further back than that, but I will keep adding to the spreadsheet I've started.)
I really noticed this difference when I was at the tail end of my NiceHash Scrypt order that I had running pre-block 145K. I let it run for some time longer after block 145k, but noticed that the block discovery was WAY slower than it had been pre-block 145k. So I stopped it as from an economic point of view it made more sense to spend the same money on SHA hashes from NiceHash. It's as if the difficulty retarget calculations for the other 4 algos are correct, but still not right for Scrypt?