In reply to Offwidth:
> All are well known scientific reasons why such data sets can be problematic.
I agree completely with your last paragraph. My argument is not that the databases are inherently trustworthy and free from bias. My point is that bias exists everywhere. Humans are more likely to believe somebody that they like, rather than somebody they don't like, even if the unpopular person is objectively more correct. It's why we have 'alternative facts' and 'fake news' and all that. Thus, guidebook writers, in deciding whose opinions to trust, will tend to side with people they like the most, and not necessarily with those whose opinions are more valid.
BTW, and with the greatest of respect, the example you chose wasn't very convincing. Firstly the 'average view of most experienced climbers' is a very difficult statement to justify (particularly re. my point about bias above), despite your experience both as climber and guidebook writer. Secondly, I really couldn't tell you difference between low-VS and mid-VS. I'm not sure I ever could, even when VS was my maximum grade.
I'll throw a counter example to you:
The Strait Gate (E2 5b) at Mother Scary's is given E1 in Rockfax guides. I climbed it years ago and thought it was the living end for E1. I found out afterwards that CC guides have it as E2 (at least, early ones did, I haven't checked the newer ones). So, two different sets of 'experts' have differing opinions. Who to believe?
Well, as it turns out, of the 117 votes it has on the UKC logbooks, it seems that most people think it is low-end E2, as opposed to high-end E1. So I gave myself an E2 tick, and felt a bit better!