One thing that surprised me has been seeing the affects of either alcohol or caffeine on the type and quality of the sleep I get.

Even if the absolute numbers aren't 100% accurate the watch definitely spots when I've had even 1 beer during the evening.

I also find heart rate variability interesting. I can't put much spin on the absolute numbers but after either heavy exercise or if I've been unwell I can see the variability rate really drop.

A lot of value is in the long term trends. One bad night doesn't mean much but if our sleep quality is trending down over weeks, it's a sign you should change something.

What makes one measurement "real" and the other one "not real", if you're mainly using it as a personal metric?

The reason that the traditional definition of resting heart rate exists is people didn't have 24/7 heart rate monitors, and doctors had to measure what they are able to measure. And they still can't measure it well, because patients often have white coat syndrome and there's not enough time during an appointment for people to relax fully.

The Apple Watch, which uses the traditional definition, has to use algorithms to guess which of its measurements counts as "resting" or not, which adds complexity. In contrast, lowest HR during sleep is a more reliable and consistent measurement.