1. principles
  2. Hugo Voting

Hugo Voting

By drorzel on July 2, 2009.

Just a reminder, if you're someone who's eligible to vote for this year's Hugo Awards, the deadline to do so is tomorrow. Of course, you probably already know that-- they sent out reminder emails last night. They want me to vote so badly, in fact, that I got four reminder emails last night, two with my own member number and voting PIN, and two with somebody else's...

I sent my vote in this morning. Once again, this was a year in which there was a huge gap between the category winners and the next-best nominees. It was awkwardly large, in fact-- not quite big enough to put "No Award" second, but big enough that I wanted some way to indicate that. I wonder how they would handle a ballot with no second-place vote, that went directly from "1" to "3"?

For those who care, my votes are below the fold:

Best Novel:

  1. Anathem
  2. The Graveyard Book
  3. Little Brother
  4. Zoe's Tale

This was probably the category with the biggest gap. Anathem is head and shoulders above all the rest, which all have significant flaws. The ordering of the next three almost doesn't matter-- they're good books, but not even close to Anathem.

I haven't read the fifth nominee, because the whole concept of Charlie Stross writing a late-Heinlein tribute novel makes my skin crawl. I can't bring myself to read it, but not having read it, I can't put it below "No Award."

Best Novella

  1. "Truth" by Robert Reed
  2. "The Tear" by Ian McDonald
  3. "The Erdmann Nexus" by Nancy Kress
  4. "The Political Prisoner" by Charles Coleman Finlay

Another category that isn't even close. "Truth" is better than the next three put together. The McDonald gets the edge over the Kress because Kress says silly things about quantum, and the Finlay is a lovingly detailed story about how much it sucks to be in a concentration camp, so it goes right to the bottom.

"True Names" by Doctorow and Rosenbaum got the Eight Deadly Words halfway through, but it wasn't awful enough to go below "No Award." I just didn't care about what was going on.

Best Novelette

  1. "The Ray-Gun: A Love Story" by James Alan Gardner
  2. "Pride and Prometheus" by John Kessel
  3. "The Gambler" by Paolo Bacigalupi
  4. "Shoggoths in Bloom
  5. No Award
  6. "Alastair Baffle's Emporium of Wonders" by Mike Resnick

Another category with a clear winner and a bunch of "Enh." "The Ray-Gun" is a Kelly Link-ish story where the fantastic element is just an excuse to tell a sweet story about two ordinary people. After that, there's a huge step down to the next three, which are, respectively, second-rate Neil Gaiman, an unsubtle lecture about how we're all Bad People, and a Lovecraft homage that goes nowhere.

If there were a way to put "No Award" on the ballot four or five times before the Resnick story, I would. God, that story (or, more precisely, its presence on the ballot) pisses me off.

Best Short Story

  1. "26 Monkeys, Also the Abyss" by Kij Johnson
  2. "Exhalation" by Ted Chiang
  3. "Evil Robot Monkey" by Mary Robinette Kowall
  4. "From Babel's Fall'n Glory We Fled" by Michael Swanwick
  5. No Award
  6. "Article of Faith" by Mike Resnick

Finally, a difficult decision, at least for the first two spots. Kij Johnson gets the edge in the end because the Chiang got a little too obvious at the end, but they're both very good stories.

After those two... "Evil Robot Monkey" comes third because Swanwick has some sort of artistic objection to endings, or something. Resnick, of course, comes 428th, after a whole bunch of "No Award"s.

I have no opinion on most of the other categories, due to not having enough time to read/ watch/ give a damn about the nominees. I did happen to see 60% of the field for Best Dramatic Presentation (Long Form), so I threw in a partial vote of 1) The Dark Knight, 2) Iron Man, 3) Hellboy II. I didn't get to the other two, so I have no opinion on them.

The other category for which I cast a complete ballot was the John W. Campbell Not-a-Hugo Award for Best New Writer:

  1. Felix Gilman
  2. David Anthony Durham
  3. Gord Sellar
  4. Aliette de Bodard
  5. No Award
  6. Tony Pi

Here, we see the dangers of putting together the Hugo Voter Packet-- as I don't read much short fiction, I wouldn't ordinarily have any idea who Tony Pi is. Thanks to the packet, though, I was able to read his story "The Stone Cipher," which was sufficiently awful to put him below No Award. We're talking Resnick-level bad, here.

As for the others, Thunderer (included in the packet) was a better novel than Acacia (which I read a year or so ago). I enjoyed Gord Sellar's story about jazz musicians playing for aliens, and he gets bonus points for doing a Many-Worlds thing without including a tedious and incorrect lecture about how it works. Aliette de Bodard's Aztec-themed stories were unobjectionable-- not great, but not shriekingly awful, either.

And that's how I voted, and why.

Tags
Books
Pop Culture
SF

More like this

Skipping straight from 1 to 3 would be handled exactly the same as going from 1 to 2: if your first choice isn't the winner, than your next-highest choice gets your vote.

(Waring: voting system geekery; check the blog leastevil.blogspot.com)

This method is called instant runoff voting (IRV) and, despite claims to the contrary on the Hugo Awards page, it does not guarantee that the winner is preferred by at least 50% of the voters.

If there is as large a gap as you say between the top-place contender and the rest of the field, than everything should work out fine; similarly, IRV can handle two strong contenders. However, it is known to have problems when there are more than two strong contenders.

The nominations process is pretty smart though; it's similar to a method called approval voting. The one difference is that, in proper approval voting, you aren't limited to how many candidates you can approve, unlike Hugo's maximum-of-five. Removing that restriction might create some paperwork headaches, but could slightly improve the field of nominees.

However, I would advocate replacing the final IRV round with score voting. It works like this: you give each candidate a score in some range (such as -3 to 3, or 0 to 99). Highest average score wins. This allows you to express your degree of preference, to easily express equality of preference, to fairly express "no opinion" without unfairly negatively (or positively) effecting a nominee, and can handle of field of three strong contenders.

For instance, you could (assuming a 0 to 9 rating scale) give Anathema a 9, the next three a 6, and "no opinion" on the fifth; that more-accurately matches your honest opinion than the current rank-order based method.

(Clever folks may have noticed that approval voting is just score voting with a range of 0 to 1.)

Also: Anathema was the only sci-fi I read this year, but I'm confident in saying it deserves to win every award it gets nominated for. It's one of the best books I've ever read.

I guess I should look over Anathem again. After Cryptomonicon, I got tired of Stephenson's quirks.

I subscribed to Asimov's for a year as an experiment. Didn't renew... but the 28 Monkeys story was one of the few things in that year that made me smile.

Was the Swanwick related to his _Dragons of Babel_? Because at least two chunks of that had been published earlier as short stories. (If you haven't read it, do. You'll like it.)

Doug M.

Fascinating analysis.

Instant runoff voting has worked well for Hugo Awards and I wouldn't change it. It sure is nice to be able to rank all your choices -- it's history in lots of other big elections is helpful to see too.

Oh, how I envy you. I HAVE read the Stross book. Yuck.

Dale is correct. Instant Runoff Voting is quite poor at measuring voter preference. Score voting is a vastly superior method.

Here are Bayesian regret figures that show that:
http://ScoreVoting.net/UniqBest.html

It is kind of strange that a group of presumably scientifically adept people would choose a poor and relatively complex voting method over a simple method that is essentially the best one known.

JB,

Well, there are lots of voting methods that use ranked ballots -- IRV happens to be one of the worst. Various Condorcet methods and even less known methods like Bucklin are all far superior to IRV.

But it turns out that all ranked voting methods are subject to problems as proved by e.g. the Gibbard-Satterthwaite theorem and Arrow's theorem. They tend to be particularly problematic with regard to tactical voting.

Score voting allows you to express not only order of preference but intensity. Plus, that information is less tainted by voter dishonesty (tactical behavior) and tabulation flaws (e.g. IRV ignores asymptotically 100% of ballot data).

When you say that IRV has "worked well", I can only cite Bayesian regret figures, which show exactly how well IRV performs. If you call a social utility efficiency of around 50% "good", then you should really like score voting, when it produces social utility efficiencies more in the neighborhood of 90%.

Doug -

Almost diametrically opposed in some of the short fiction categories, but we were pretty close on the novels and, yes, agree about Resnick. And we had the same top two in short story, albeit in different order.

I'm still stunned that Anathem did not win. The Gaiman novel was fine and I quite enjoyed reading it, but it was utter fluff by comparison.

Susan
(arriving belatedly in the comments after following back a hit on my short fiction post from here)