Nope, sounds like you didn't look at the references I linked at all. If you want to follow something right, try reading the references, or at least casually skimming them. https://coinshares.com/insights/beware-of-lazy-research-bitc... is one of the sources for the percentages cited in the Harvard Business Review article explaining their methodology (in 02018). It's the polar opposite of what you're saying.
It's also true that there are fundamental economic reasons that Bitcoin mining is mostly done with renewable energy; it's not just a weird coincidence. But it's not just a theoretical prediction.
> retail prices are only vaguely correlated with production prices. Many places subsidize the crap out of energy prices
Yeah, that does happen. Also, outright theft happens; the cheapest energy is energy that you steal. But it turns out that unprofitable subsidies and stolen energy aren't very scalable, because places where the government pays for most of the energy ① tend to have strict oversight of industrial-scale operations, ② don't generate much electricity in the first place, and ③ tend to consider entrepreneurs criminals by default because they increase wealth inequality, so they shut them down when they get found out. When Bitcoin mining (or, say, Litecoin mining) starts to scale up in such places, as it did here in Argentina some years back, the energy subsidies have a way of evaporating.
> I'll also note that there was a story a few weeks back about a coal powered plant being recommissioned solely for mining bitcoin
No, that was gas; my comments on it are https://news.ycombinator.com/item?id=27450889.
You're right that there are a variety of unusual circumstances that can make fossil-fueled Bitcoin mining profitable at small scales in one or another place, so it's easy to find anecdotes of it. But it's not the majority of the network.
If I'm following this right, the only evidence for Bitcoin being mostly based on renewable energy is because people believe it's uneconomical to not be renewable, and not based on any actual attempt to figure out what energy types are actually being used by large mining groups?
Because, especially when you compare across international countries, retail prices are only vaguely correlated with production prices. Many places subsidize the crap out of energy prices, and electricity in a subsidized place could well be much cheaper than in a non-subsidized place, even if the first plant is burning coal and other plant is using wind.
For some specific data points, I believe a large clutch of miners operated out of Inner Mongolia, whose electricity capacity increases in recent years has been dominated by coal-fired power plants (because it's a coal-rich region that doesn't give two whits about environmental effects). I'll also note that there was a story a few weeks back about a coal powered plant being recommissioned solely for mining bitcoin.