[RedDevil]

Hi CapriRacer, thanks for that.
In the meantime I edited my post and added another measurement of the circumference, for reasons mentioned in the post, and that gave a smaller circumference and hence a much smaller difference; less than 1.7%

It is safe to assume that the tread diameter went down as the lugs wear. The reduction of the difference between rolling and actual diameter for my worn tires suggests the rolling diameter has not changed much.

But, as I wrote in my first post on this topic, the rolling distance has actually changed over time; my commute became about 0.6% shorter by the odometer, give or take 0.2% but definitely shorter even though the route has not changed one bit.

So the rolling distance did increase over time, probably due to gentle stretching and settling of the belts.

Googling and reading into rolling circumference, everyone seems to agree that it varies with load and inflation pressure. Tread wear does not get mentioned, at least not by the experts.

This puzzles me.
I can't see how load influences rolling distance.
The contact patch would always be flat unless the road surface gives way (mud, sand, snow). Even if it does deform, as long as it approximates flatness it will still be very near constant as the difference reduces quadratically.
Pressure may stretch the belts a bit, but every tire has some pressure so they are already tight.