2015 Hugo Awards: The Kingmaker Scenario and Margin of Victory
As I finish up my investigation and predictions of the 2015 Best Novel Hugo Award, we need to seriously think about a possible kingmaker scenario. A term borrowed from game theory (which in term borrowed the term from royal politics), Wikipedia gives us a good working definition: “A kingmaker scenario, in a game of three or more players, is an endgame situation where a player unable to win has the capacity to determine which player among others is the winner. Said player is referred to as the kingmaker or spoiler.” I know we shouldn’t trust Wikipedia, but this is a well-established theoretical, and that’s as good an intro as any I could easily find.
In a normal Hugo year, it’s hard to be a kingmaker . While it’s easy to boost the #2 choice to the #1 place, you’d have to guess at who would be #1 and #2. It would be very hard to boost a #5 work to #1—but let’s imagine a year where you only have three viable candidates. How hard would it be to boost #3 into the #1 position? That’s exactly what we have in the Best Novel this year: two Puppy candidates, three “regular” candidates, and a tailor-made kingmaker situation.
For the Hugos, you can consider the “players” the supporters of the 5 Hugo nominees and the “endgame” as the final vote. It would appear—and I stress that we don’t know this for sure—that Katherine Addison, Cixin Liu, and Ann Leckie are the only viable players for the 2015 Hugo. Debate that if you will, but I think that the negative reaction to the Sad/Rabid Puppy campaigns and the use of “No Award” make it nearly impossible for Butcher or Anderson to win. I also think “No Award” isn’t viable in the Best Novel category, due to the strong support that Addison, Liu, and Leckie have received.
In a kingmaker scenario, the Butcher/Anderson voters would have the power to choose between Addison, Liu, and Leckie as the winner of Hugo. Even a smaller set of the Butcher/Anderson voters—like the Sad or Rabid Puppies—could operate as kingmaker. Of course, they would have to be organized, have to have sufficient numbers (see below for the estimate), and choose to execute a kingmaker move.
Interestingly, the Hugo voting rules make it incredibly easy to deploy a kingmaker effect. Since the Hugo operate via instant-runoff voting, you just vote for your favorites (even if they don’t have a chance), then vote for who you want to win (your kingmaker effect), and then you leave off anyone you want to lose. The instant-runoff vote will maximize your impact for you. In a first-past-the-post voting system, a kingmaker would have to be more organized (you have to leave off your favorite and vote only as a kingmaker), making it less likely to work.
So, what kind of numbers do we need? The math gets a little tricky, but if we look at the Hugo data from 2011-2014, we can get a rough sense in the main fiction categories. I left off 2010 because that’s the year Mieville and Bacigalupi tied, which messes up the stats a little. Still, that shows how close these votes can be. A single kingmaker voting for Mieville would have pushed the balance!
Here’s the data separating first place from second place, in the final pass of the instant run-off voting:
Table 1: Margin of Victory in the Hugo Fiction Categories, 2011-2014
Throw out Leckie’s domination last year, and you see that in most Hugo years, it would takes only a modest number of organized voters (between 100-200) to boost the #2 novel to the #1 position. As I said before, because of the way Instant Run-Off voting works, all the campaign would need to do is leave that #1 novel off the ballot.
It’s interesting how up and down these numbers are. Sometimes a kingmaker effect is easy, sometimes hard. I’m surprised that Short Story had such large margins of victory in recent years. Also pay careful attention to 2014: that’s the first year we had a big boost in Hugo voting numbers. Since we might double that this year, we could be in a situation in 2015 where margin of victory is so large that no kingmaker effect is possible. Ancillary Justice had a very unusual year last year, though, sweeping all the major awards. We don’t have that kind of dominance in 2015.
With somewhat higher block voting numbers, you could exert more control over the ballot. I took a look at the difference between 2nd and 3rd place in the next to last run-off (that’s what you’d need to boot the #2 work from reaching the final stage). If you add these numbers to the previous chart, that’s a rough estimate of what it might take to get work #3 to win the Hugo. This isn’t 100% accurate (we can’t do that math, or at least I wasn’t able to figure out a way to do that math), but it’s a good eyeball test.
Table 2: Margin of Victory of 2nd Place over 3rd Place in the Hugo Fiction Categories, 2011-2014
Again, widely variable by year and category, with some remarkably close and others far apart. Still, as a rough estimate, I think it’s revealing. So let’s say you wanted 2312 by Kim Stanley Robinson to beat Redshirts in 2013. It would have taken around (213+6), 219 votes to do so. It would have only taken 213 votes to give Lois McMaster Bujold’s Captain Vorpatril’s Alliance the win.
Would an organized Hugo campaign want to do something like that? I’ll let you decide.
So, what does that mean for this year? Keep in mind that we’re due for perhaps a doubling of the Hugo vote, which could double margin of victory, making a kingmaker effect harder to pull off. Second, the vote between Liu, Addison, and Leckie is a 3-way, not a 5-way battle, and that may further concentrate votes. But . . . I have a feeling we’re in for a close contest, which might offset the increased number of voters.
I’ve previously estimated the Sad/Rabid Puppy campaigns at around 300-400. However, that was only for the nominating stage; we don’t know what that number will be in the final vote. Doubled? 1.5 times higher? What’s the ratio of Sad/Rabid Puppies? Will those groups chose to pull a kingmaker? The Sad Puppies, to the best of my knowledge, have made no move to suggest how their group should vote. The Rabid Puppies, on the other hand, seem to have made just such a move, given Vox Day’s Best Novel post. Note that this is an exact kingmaker play: Leckie is left off the ballot completely. The result will be to push the award to either Liu or Addison if the Rabid Puppy vote is larger than Leckie’s (possible) margin of victory. If Addison were to be in a close race with Liu, this would also give Liu the win (if the Rabid Puppy block size is larger than Addison’s margin of victory). So, that leaves two unknowns: how big is the Rabid Puppy block, and will they stick together? I could make guesses, but I try to avoid sheer guesswork on Chaos Horizon.
The Hugo is remarkably amenable to kingmaker scenarios. In fact, you could argue kingmaker scenarios are more favorable to block voting groups than outright sweeps, as it allows them to exert more control of the final result. In a sweep, No Award is an option; in a kingmaker scenario, it really isn’t. This is an issue the any proposed changes to the Hugo votes should keep in mind; you could implement a change that makes kingmaker scenarios even easier to set up. The best way to defuse kingmaker scenarios is to have more viable players in the game, but, even then, it’s easy to make sure one player doesn’t win if you control a block vote of any size.
So, in conclusion, in close years and in close categories, a kingmaker effect is remarkably easy to deploy. It is even easier to deploy if a campaign places 1 or 2 of their works on the ballot. In fact, if someone was trying to dominate the Hugos, that’s the situation you would probably want: 2 of your picks on the ballot, 3 other picks, and then you could chose between the three other works to see who would win. You get to make your point with your picks and then shape the outcome of the award. There’s also relatively little risk in deploying it: you don’t hurt your own choices’ chances. On the other hand, we don’t know if the increased number of Hugo voters will make margins of victory so large that kingmaker effect doesn’t come in to play. Stay tuned to Chaos Horizon; I’ll run the numbers when they’re published after the Hugos, and see how things worked out.
SHEER GUESSWORK TIME
I’m guessing the first thing to be eliminated will be The Dark Between the Stars – I haven’t see anyone rank it first and it seems far behind Skin Game.
Then No Award – there’s a small group that’s going to no award all the categories with any puppy nominees, and I’m betting it’s bigger than DBS fans.
Then Skin Game, having received a few of the DBS votes. I’m guessing this is going to be a decent block. You’ll get most of the puppies and some general Fantasy fans. I’m guessing this splits into roughly thirds – one going to Goblin Emperor, one going to Three Body, and one dropping out (puppy-only voters). Probably much fewer going to Goblin Emperor than the other two.
Then it all comes down to the standings of the other three nominees. I’m guessing the original preferences will be roughly equal, with Ancillary Sword in the lead by only a little bit, then Three Body Problem, and then Goblin.
Both Goblin and Three Body seem to have more overlap with Ancillary Sword than with each other, so if either is in third place after Skin Game’s elimination Ancillary Sword wins.
If Ancillary Sword’s in third place, it’s basically up in the air, probably with a slight advantage to Goblin.
Numbers from a hat time (for winning chances):
Ancillary Sword 60%
Goblin Emperor 20%
Three Body Problem 15%
Skin Game 4%
No Award 1%
The Dark Between the Stars 0%
If the SP/RPs did half the scheming their critics ascribe to them, they wouldn’t have time to put any of those schemes into effect.
Every Sad Puppy supporter I’ve talked to, myself included, plans to vote Three-Body Problem #1 (it’s simply better than the SP nominees, IMHO). Vox Day did the same on his vote, which means that if the Rabid Puppies follow his lead and the Sad Puppies meetup I attended was a good barometer of how the group will vote, then both factions will put their weight behind it. What that means depends entirely on the size of those groups compared to the electorate as a whole, of course. AFAICT nobody knows for sure the meaning of the huge swelling of supporting memberships. Should be an exciting election!