It seems like an impractical but useful thought experiment, which I think makes a good point about why these lists should be taken with a grain of salt. To take things further, we could improve the experiment by randomly assigning books to founders when they start! HN folks are quick to point out when many scientific results as “correlation not causation”, and this isn’t any different, in my opinion.

Yes :). I'm just saying that it could theoretically be done properly in a way that addresses GP's criticisms, but it would be impractical.