That’s a good idea . So you’re telling me that I can trust the on device numbers enough?

I've never been quite clear that the numbers are completely trustworthy, but at least the counting is more solid than the restrictions and the passwords.

One thing I've noticed on my own device, that, counter-intuitively, screentime also seems to include time when the screen is off, e.g. playing music through Safari with the screen turned off.

As an example of the fragility of the screentime passwords, the password on a child's device can often be removed by simply turning off the screentime password - it will not always prompt for the current password. Until recently, I actually thought that it was intentional that a password was not required to disable the screentime password on a child device.

I recently had a crazy experience where many restrictions and a screentime password seemed to suddenly get enabled on a child's device. We had just turned on screentime, but without restrictions or a password. It seems as though a bug caused the restrictions to revert to those used a long time ago during an exam period, with very strict time restrictions. It was almost unusable, and no one knew the screentime password. The crazy part is that I tried dozens of passwords - many on the first day, and then one a day for weeks. I was certainly the only one touching the device from that point. One day, when I tried to turn off the screentime password, it just turned off. Unlike every other day for the past few weeks, it did not prompt for a password at all.

Thanks this mirrors my experience. I’ll try your trust and verify approach then.

I wouldn't trust them. I have seen screen time over 24H at least twice.