Looking at the ACE2 alignment from the article, it looks like there are 55 spots where the sequences have different proteins. Of those 55 spots, 7 of them are identical between the bat and sars2. On the other hand, I count 20 matches between sars and sars2.

Using a uniform distribution over the 20 proteins, and saying the probability of two proteins matching is 1/20, then the binomial probability of getting 7 or more matches out of 55 is 0.02, whereas the probability of getting 20 or more matches is less than 0.000001.

I can see 0.02 being achieved by chance, but 0.000001 seems pretty unlikely to happen by accident. So, there is some sort of non accidental relationship between sars and sars2. Maybe 1) sars2 is descended from sars, or maybe 2) it is lab engineered.

If the bat coronavirus is likely the more proximate ancestor to sars2 than sars, then #1 seems unlikely, which makes #2 the more plausible hypothesis.

Also, to return to your argument about restricting our 'lab origin' hypothesis to known viruses, it seems that if WIV found a very effective bat coronavirus, and intend to create a bioweapon from it, this is exactly the situation when they would not share the sequence. I do not understand why you think people creating a bioweapon would want to share their materials with the world.

You may also find this other article by the same author interesting, pointing out the evidence strongly points to RaTG13 being faked. https://nerdhaspower.weebly.com/ratg13-is-fake.html

There is also this interesting tweet from Jonathan Jacobs that the sample data for RaTG13 does not match the assembled genome. https://twitter.com/bioinformer/status/1252813532850081792