As the dust settles from this year's Hugos, I thought it worth revisiting my two posts from earlier this year assessing how the nominees had been rated in different ways online.
The LibraryThing/Goodreads statistics proved a good guide this time round, with Redshirts, the Best Novel winner, pretty far ahead of the field on both sites. It has to be said that LibraryThing/Goodreads prognostication is not always so successful. LibraryThing called 2312 for the Nebula this year, and Goodreads was not far off; but the process failed completely for the BSFA Award and the Clarke Award - both winners, Jack Glass and Dark Eden respectively, were fourth on the Goodreads ranking and fifth on LibraryThing. Basically this is a good way of identifying books that have built up a wide audience, but won't take you much further.
My survey of blog posts got two winners in the fiction categories right and failed to spot the other two. The overwhelming consensus from bloggers for "Mono no Aware" for Best Short Story, and the strong consensus for The Emperor's Soul for Best Novella, were reflected in the voters' choices. But only one blogger of my original survey went for "The Girl-Thing Who Went Out For Sushi" for Best Novella, and none at all for Redshirts, though each of them drew support from two (different) commenters to my post.
This is not unusual. In my 2011 survey, the blogging consensus converged correctly on "The Lady Who Plucked Red Flowers beneath the Queen’s Window" for Best Novella. But the strike rate was unimpressive in other categories, with majority support for ultimately unsuccessful nominees in the other short fiction shortlists, and not a single blogger in my survey voting for the eventual Best Novel winner, Blackout/All Clear. Basically, blog surveys are a pretty blunt tool, covering only the articulate voter who posts in forums which I can see. If there is a strong consensus around a particular nominee, it is often right. If there is no strong consensus, it is certain that most people are wrong. (Using "right" and "wrong" as shorthand for "correct [bzw. incorrect] reflections of the outcome of the actual vote" rather than any judgement of individual choices here.)
Despite the demonstrably limited value of these surveys, I expect I shall continue doing them; it is interesting to identify front-runners, especially when it turns out that they do not win.