Wednesday, August 30, 2006

Public Choice At Work: The Hugos

by Tom Bozzo

Via Gary Farber, here's the voting data for the 2006 Hugo award for best novel, deservedly won by Robert Charles Wilson's Spin — which even made it through my reading queue before last weekend's announcement of the Hugo winners.

Hugo voting uses what my former prof Dennis Mueller calls a "simple alternative" to majority rule, a form of instant runoff voting a/k/a the Hare system — though it's evidently not so simple that this web page doesn't try to dispel a view that the Hugo voting procedure is too complicated.

The World Science Fiction Society's openness is refreshing, especially compared to the seeming a mystery (at least to a cursory web search) of whether, say, better nominees split a quality-film vote to allow "Crash" to take home this past year's Academy Award for Best Picture. In fact, there's research showing that plurality rule (let alone, ahem) breaks down quickly as the choice set epands, even for the typical five or so award nominees.

The Hugo electorate put Spin solidly in first place, with Charlie Stross's Accelerando (an Ad-Hoc Book Club selection from earlier in the year) solidly in second — perhaps to Stross's surprise. I think the voters got that right. Stross's self-criticism is basically right in implying that Spin is the more cohesive novel.

Wilson leaves the technology behind Spin's central mystery, involving some alien engineering that threatens to advance the end of the world within the Social Security trustees' non-infinite analytical horizon (*), as magic; his accomplishment is in constructing a plausible story of the human reaction to the catastrophe along with a cracking scientific detective story that otherwise rigorously plays within the story's rules. (Little of what's good about that can be discussed without major spoilers; if you're interested in such things, just go ahead and read it.)

Stross, in contrast, starts off extrapolating from a very recognizable near future as the world's aggregate computing capabilities increase exponentially (which in its own way leads to the end of the world as we know it), and up to a point plays within something resembling the limitations of known science. It could be argued that the future shock interferes with the development of characters from the first part of the book (**). This can cut a couple ways, as Ken MacLeod's fourth-place Learning the World, set some ten thousand years in the future with nothing remotely resembling faster-than-light space travel (***), didn't seem to me to have enough future shock to show for the elapsed time.

Last, the Hugos demonstrate that fancy voting systems can have somewhat strange results. While first and second place were solid, the nominee that finished third in the first-place tournament ended up fifth when the dust settled; the fourth and fifth place finishers from the first-place round each moved up a notch. Perhaps more curious are the 17 cranky voters (3% of the electorate) whose first choice was "no award" given the more than respectable field. Go figure.


(*) The eponymous Spin is a temporal anomaly that dilates time as experienced on Earth such that the death of the Sun billions of years hence would occur in a few decades per terrestrial clocks. Don't even think of asking how such a thing might be done.

(**) That, IIRC, is what lost fellow Clubber Brayden King around the halfway point.

(***) But, in the end, an interesting take on the 'anthropic principle.'
Yeah, I've enjoyed playing with some of the figures for years, teasing out which blocks of voters who like X clearly next most like Y, in various categories; it's always been helpful in understanding taste in the field in all the various areas.

"Perhaps more curious are the 17 cranky voters (3% of the electorate) whose first choice was 'no award' given the more than respectable field. Go figure."

This is often a puzzler. Notice that the Best Editor category also started with 15 No Award votes. Does this mean that 15 think all the named editors suck, and that 17 people think all the novels suck?

Possibly, but my suspicion is that a high percentage of those votes in many cases (not always; there are other situations at other times; there are definitely times when a protest vote is clear) are simply voters who think it means "I can't decide" or "I don't know enough," and that they have to vote in each category. (This has often most clearly been the case in the Fanzine category, since a far smaller percentage of the voters are familiar with fanzines than the fiction, for instance.)

The funny thing is that since I'd already read the list of winners before I pulled that specific iteration, I actually overlooked that the break-downs were included right off; for almost all of the past 35 years or so (sheesh!) I've been looking at the results in real time, the final break-downs weren't made public for a couple of months, since the committee was always too burned-out to get around to it until then. The idea that they'd be efficient enough to release them with the results is so alien to me that it never even occurred to me that it would happen.

But, what the hell, according to all reports, many other Worldcon traditions, such as not having a competent pocket program design, were still maintained.

Not to mention that people are still arguing furiously about Harlan Ellison.
Gary, thanks for the comment.

I certainly can see the "no award" option arising in other categories as you suggest. It's also easy to imagine it arising as a cutoff between one or two top choices and "no preference" or "no information" choices. It's still a litle interesting, though, for the headline fiction category, given the self-selecting electorate.

While it would clarify things to allow voters to register flat preferences over some subset of choices, that creates a variety of technical challenges (admittedly still on the remote side) in reducing the probability of ties and other 'indecisive' voting system outcomes.

Esp. with online voting, I'd have assumed that the tabulation was at least semi-automated, which naturally would facilitate distribution of the detailed results.
I didn't vote this year, continuing a twelve-year trend that started accidentally with Glasgow (1995), where it took me less than four minutes in the SFWA suite afterward to find five people who would have voted for Chris Priest's The Last Deadloss Visions over I. Asimov (not necessarily first, but that doesn't matter in an Australian ballot), but had not received their ballots in time.

For all my joy at David winning Best Editor this year, I might have been inclined to vote "No Award"--he's had better years, and the other contenders were also Usual Suspects who did nothing extraordinary this year as opposed to last. (Sheila probably deserves the most credit; no one I've heard is speaking of a "precipitous decline" in Asimov's

But I'll stop neeping now, and go back to pretending I don't follow such things.
Just found out about the "argument." Who the f*ck would argue in favor of him?

Details here, with appropriate comparison.
"For all my joy at David winning Best Editor this year, I might have been inclined to vote "No Award"--he's had better years, and the other contenders were also Usual Suspects who did nothing extraordinary this year as opposed to last."

As innumerable of us have pointed out for years, what needs to be done is to, if not split the "Best Editor" category into those for "Periodical" and "Books" (and I'm always reluctant to add more categories), to at least do what would be necessary in that case, anyway, which is publish and distribute to all voters and eligible nominators a list of what books each book editor was responsible for in the given year. Those who pay attention or work in the field may know, but most voters do not.

Meanwhile book editors are totally discriminated against, and judged only when their name is on an anthology, which is ludicrous.

And I have to scratch my head at the above comment, though: the award is for "Best Editor" of the year; it's not for "Most Extraordinary Work By An Editor," or such. While it's certainly true that one could easily argue that there's no One Best -- the fact is that that's usually a winning argument in most all of the categories; the times it isn't are exceptions, not the rule. But it doesn't take being extraordinary to win the award; just the most popular that year.

On the other hand, everyone gets to vote by their own criteria, of course, and I've certainly heard endless sillier ones than voting only for extraordinary work. (In particular the dread "because I know X," since it's not supposed to be an award for Most Popular Person, though they frequently all inevitably are to some degree; can't be rooted out, after all.)

"Esp. with online voting, I'd have assumed that the tabulation was at least semi-automated, which naturally would facilitate distribution of the detailed results."

I believe so. I used to be on the smofs list, and whatnot, back before I went pretty much gafia circa 2000 or a bit earlier, and I seem to recall that them thar newfangled "com-poo-ute-ars" had been used in recent years.
I know all five of the nominees this year, having worked with two (David and Gordon at NYRSF), cat-sat for Ellen, and lived down the block from Shiela. (I've met Stan Schmidt at SFWA functions and David's house over the years, but wouldn't claim familiarity.)

The award is called Best Editor. It's not "best" editing to just keep a magazine publishing each year. (If it were, people should vote Gordon every year, he being the only O&O on the list.) Can we look at 2005 and say, "Dangitall, Analog really went to the next level in retrotexts" or "Ellen's anthologies this year included all the best short work published in the field" or "You know, since Sheila took over, I'm enjoying reading Asimov's again"?

The award has been a popularity contest for years—which is one reason Gardner kept winning—but if one were taking it seriously as Best Editor, there would have to be something that differentiated it from (Same Old)^2.

One of the aspects of Australian ballot (as Don Keller wisecracked when I ran the Crawford Award that way) is that it effectively eliminates taste. But the starting point of the Preferential Ballot is still having a preference. Since the award is for the Editor, not the magazine, my preference should be influenced not by "I read F&SF when I can and Asimov's when I must" but rather, "I usually read F&SF before IAM and then Analog. This year I (a) am more certain than ever that I'm right, (b) will switch the order, (c) would rather read Ellen's anthologies, or (d) want to concentrate on the novels David edited.

If it's none of those, then we're rewarding magazines, not editors.
I don't think we're disagreeing about anything, Ken; if we are, it's obscure to me.

"The award has been a popularity contest for years...."

Well, all the Hugos have never been anything other than a popularity contest, have they? It is what it is.

"The award is called Best Editor. It's not 'best' editing to just keep a magazine publishing each year."

Yeah, that's part of why we changed it from "Best Magazine," after all.

I otherwise agree with everything you've said.

Don and I go back to when he was co-editing Phantasmicom with Jeff Smith (whom I was pleasantly surprised to hear from by e-mail last month, as I posted, after I posted about the Tiptree bio, by the way) and living in Philly, let alone later when we were all in Seattle together, incidentally.
It turns out, by the way that the split of the "Best Editor" category into Long and Short Form did pass the Business Meeting this year, although it will still have to be confirmed next year.
Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?