I do have some familiarity with how such old systems would misbehave with corrupt data, such as from bad data cables. I didn't see any of those kinds of symptoms and so do not think there was any significant storage decay from being on the shelf at least 5x longer than the 5 year interval mentioned up thread. I tend to think that such a failing drive must have had some other kind of damage and not just some kind of spontaneous demagnetization
Refreshing every five years for magnetic storage is probably overly cautious to me from my own personal experiences. But FWIW if it's data I really care about it doesn't spend years sitting around on just magnetic media. For my own personal risk tolerance putting important data on say an external hard drive and sticking it on a shelf isn't a great way to archive data. I might do it with data I'm OK with eventually maybe losing, but if it's something I know I'll want in 20+ years it gets a few copies with one being on some highly reliable optical media like M-DISC.
Your hard drive might go a long, long time without losing a bit. And maybe if you do lose some data, it's a couple of typos in some help file. Maybe it's some rare code path in an executable that almost never gets called. Maybe it's some b-frame in a video that you'd never even notice. Maybe it's some important bits of the last photo you had of a loved one. Maybe it's some important field a database. Maybe it was empty space. Who knows.
Most of the bits will probably be fine. Some of them probably won't be. The rate depends on a myriad of factors. Even within a brand different models will be better or worse, and it's not exactly a metric you'll find listed. And then yes the way they're stored is also a factor and even the environment when the data was written.
I've had drives from the 80s still work and still pretty much have all the data. I've had files get corrupted on modern drives with no other explanation than bitrot. Even just cosmic rays can eventually flip some bits.