Play count: tracks with crazy high numbers

[Sharky-PI]

Through years of mostly unsuccessfully migrating across music players, my play counts have generally been reset each time, sadly.

Happily, strawberry saves these to the ID3s, which is great going forward.

However I've noticed that some of my files (about 1990 of 43878 tracks) have play counts increasing from the twenties, to hundreds, thousands, and lots of tens of thousands, 5 digits, specifically with blocks at
10496
16896 (loads of these)
25856
Max = 65335

Then I have two, one with 6384640 & one with 6881400.

Whole albums typically have the same playcount.

I'm not blaming strawberry for this, just wondering if there's any underlying structure/logic which could be extracted.
Ideally some simple mathematical conversion to play count. But I'm not holding my breath!

Any ideas, genius, wisdom, etc, appreciated.
Cheers!

[Sharky-PI]

More sleuthing has uncovered something. Puddletag shows various files with Playcounts with 2 values e.g. 0\0, 0\258, up to 0\65335.

I tried to delete those screwy counts in puddletag by overwriting them with 0 but it doesn't seem to have taken. Or they've regenerated.

Checking in kid3 it turns out that one of those files, with 0\258 has
FMPS_PLAYCOUNT=0
FMPS_Playcount=258.

If I listen to it once in strawberry:
FMPS_PLAYCOUNT=0,
FMPS_Playcount=259.
So strawberry increments FMPS_Playcount but not FMPS_PLAYCOUNT.

I wonder what the official/correct tag name is?
https://www.exiftool.org/TagNames/ID3.html this says it's "PCNT" for id3 v2.4
According to https://github.com/strawberrymusicplayer/strawberry/issues/1074 , FMPS_Playcount (Strawberry spelling) is CORRECT, so logically something else has been using FMPS_PLAYCOUNT.

But why do they (second entry, FMPS_Playcount) go up to 65335?
And does overwriting them with 0 in puddletag not work because it's displaying 2 variants of the same tag?
In puddletag I can't create separate columns for the 2 spellings because it automatically uppercases them but displays both.