I simply observed that I lose weight if I eat <3k kcal over long stretches. I think human metabolism has a very wide variance, though it's hard to find good quality studies about it.

You went from 4K to 3K in a single post. Needless to say you're not exactly reliable.

I count my calories and I eat between 3.5k and 4k. I had a few periods where I ate <3k. These are two separate things. The latter is significant because the fact that I lost weight makes me think that I need my usual intake to retain my weight.

Dude you need to be a 6"5 320lbs monster to lose weight on less than 3,000 calories a day.

https://www.calculator.net/bmr-calculator.html?ctype=standar...

Metabolism varies widely among individuals, and there is no clear understanding on the relationship between calorie intake and weight. All the calculators and estimates you find on the internet are based on surface level low-quality estimates that don't (and can't) take into account all of the variances at work. That said I am aware that I'm a bit of an extreme case (always been teased for that) but that's how random variables work, sometimes there are data points outside of the bulk.

The 2000 kcal reference (which I assume you're referring to) isn't for "an average man", but for an average person. For an average man, the reference value would be somewhere between 2.5k to 3k.