Something that I read on another website in summary and filtered by my limited understanding:
When a sector fails on a drive without TLER (or some other software/firmware of the same purpose), the drive doesn't know, and the RAID doesn't know. So, theoretically, you have 4 WD Green drives set up in RAID5, and one drive has a bad sector, but doesn't know it and doesn't report it. Since that file is rarely accessed, the problem remains unknown. Then, two months later, a different drive in the array dies completely. The array tries to rebuild itself with a new drive, but when it finds the bad sector in the first disk we mentioned, it views the array as lost, and therefore the data is lost.
What the person on the other website who described this scenario is doing to counteract this is running a program that reads all data on all drives in his RAID once a day, so a bad sector would be discovered within a day, so he can replace the bad drive before a second fails and the array is lost.
But, if the drives have TLER (or some other brand's version of it) like the Red drives, it knows when there is an error and can rebuild that sector by itself, right away. And correct me if I am wrong, but if a drive fails and there is a bad sector on another drive, you can at least salvage the rest of the data. So Red drives are a huge improvement compared to the ticking time bombs of a RAID of Green drives. Those who have had good security so far have had good luck (and I hope we all have good luck in this regard).