I downloaded the spec sheet for one of Seagate's 1TB-platter drives. It said it had an areal density of 625Gbit per square inch. Assuming that bits are roughly square/circular in size, then we're looking at a linear density of around 790Kbit/in. A 3.5" drive has a circumference of around 11" or so near the outer edge, so we're looking at something like 8-9Mbit per track.
The problem with SMR is that when you write track x, you destroy track x+1. And when you go and rewrite track x+1, you destroy track x+2, and so on. Eventually, we reach a track, x+n, that isn't shingled (otherwise we must rewrite the entire disk). The question is, how many tracks are in a single "band"? If there are too many tracks in a band, then there's going to be a lot of re-writing that needs to happen. If there are too few tracks in a band, then you won't get much in the way of density benefits from shingling.
The illustrative graphics that Seagate provided all showed 7 tracks in a band. No idea if this is real or if it's just something that the illustrator came up with. But with 7 tracks per band, and, say, 9Mbit per track, that means that, in the worst case (if writing to the start of a band), we'd need to rewrite something around 63Mbit of data, or around 8MB. That's a lot, seeing as how the NAND erase blocks on a SSD are only about 2MB in size. And of course, HDDs are going to be much slower at this to begin with...