This is my third year nominating and voting for the Hugo Awards for the year’s best science fiction and fantasy, and it’s taken about that long to feel comfortable with some of the logistics. Earlier this year, I took a look at the E Pluribus Hugo nomination system, explaining what was going on under the hood in some counterintuitive cases and sharing any strategic implications. As the system was designed to minimize the impact of strategic voting, the list of implications was blessedly short.
Today, I want to dive further into the actual voting process: Instant Runoff Voting, also known as Ranked Choice Voting. Because this is used outside the Hugo Awards, it’s a little bit more familiar, and the basic concept is easy enough to explain. That said, the particulars can get a bit confusing, and I rarely see discussion of some of the strategic implications of the voting system, so I wanted to take a little bit of time and clarify as much as possible. I try to communicate everything as clearly as possible, but I’m certainly diving a bit into the weeds, so if you just want to know the strategic implications, skip straight to the last section.
Instant Runoff Voting
Instant Runoff Voting is designed to capture more information than just a voter’s top pick, thus allowing people to vote for an unpopular favorite without worrying that its quick elimination will leave them without a voice in deciding among the realistic contenders. It’s not perfect, but there’s no ranked voting system that is, and it generally accomplishes those goals.
The mechanism by which it accomplishes those goals is reasonably straightforward, with both names suggesting something of the details. Instead of making just one selection of a favorite, voters rank the candidates in order of their preference (the “Ranked Choice”), and if no option receives more than 50% of the overall vote, the last-place option is eliminated, and those who voted for the eliminated selection have their votes moved to the next choice on their ballot (the “Instant Runoff”). This process is iterated until one option receives over 50% of the vote, or until all remaining options are tied. After a winner is determined, the votes for the winner are removed and the process begins again to determine second place. This repeats to determine third, fourth, fifth, and sixth.
The Hugos do have the additional wrinkle in the form of the “No Award” option. Voters may rank up to seven options: the six finalists and No Award. “No Award” means just what it sounds like—if it wins, no award is given for that category—and actually has two chances to win. In the tallying of the votes, No Award is treated the same as any other selection, and it will win if it gains more than 50% of the vote in any stage of tallying. But even if another selection wins the instant runoff, there is a final comparison between the winning selection and No Award, considering only the relative ballot placement between the winner and No Award. If the winner is higher on more than 50% of the remaining ballots, it wins. If No Award is higher on more than 50% of the ballots (note: I don’t believe this has ever happened), no award is given.
Practically, this voting method punishes the divisive options–the ones that most voters place either at the top or near the bottom of their ballot–and it’s fairly common to see a selection receive the most first-place votes and fail to win the award. One of the clearest examples came in 2021, when “Helicopter Story” received the most first-place votes in the Best Novelette category and yet finished in fifth place. Though it began in the lead, it gained significantly fewer additional votes when lower-ranked stories were eliminated compared to eventual winner “Two Truths and a Lie,” seeing it drop out of first place well before reaching the 50% threshold. In the same year, Seanan McGuire’s Come Tumbling Down suffered a similar fate, picking up the most first-place votes but relatively few high rankings among voters who did not place it first, leading to a third-place finish overall.
While all of the ballot details aren’t public knowledge, the Hugo administrators do release fairly detailed information about each stage of the voting, and if you’re the right kind of nerd, you can peruse to get a feel for how the voting works and find interesting foibles (for instance, in the 2021 Astounding Award voting, Jenn Lyons’ elimination sent a whopping 49% of her votes to Emily Tesh, with more voters leaving the rest of their ballot blank (24%) than voting for Micaiah Johnson (15%) or Lindsay Ellis (12%). This easily made up Tesh’s deficit to Johnson, who had started with the most first-place votes and had increased her lead upon Simon Jimenez’s elimination. Tesh would take home the award by a nine-vote margin).
Practically, if the first-place votes are widely split, a candidate who picks up most of the second and third-place votes is likely to win. There are extreme edge cases where this does not apply—for instance, if a candidate picked up every single second-place vote but was the only one without a single first-place vote, they would be eliminated in the first pass—but by and large, having the most first-place votes is no guarantee of victory, and picking up a lot of seconds and thirds makes a real difference.
Strategic Implications
Because of the edge cases, there are theoretically options where it may make sense to vote against your own preferences (for instance, if you’re worried about your second choice overtaking your first choice and would prefer to see them eliminated early), but without a crystal ball and a vote that gets very close in some unusual ways, it’s hard to see cases where these strategic considerations would arise practically.
However, there is one little-mentioned fact about the voting system that makes a very large strategic impact: the last item on a person’s ballot is ranked ahead of any items left off the ballot. This means there is no mechanism by which to cast a vote of ambivalence. If you aren’t sure whether the thing you haven’t read is as good as the things that you have, your choices are limited to two: (1) guess, or (2) rank neither.
This becomes especially sticky if you rate something on your ballot below No Award. If you rank No Award at all, that counts as a vote for No Award above anything left off your ballot. So if you haven’t read everything but wish to rank No Award, you must either rank No Award ahead of something you haven’t even read (which I personally find distasteful), or you must include things you haven’t read on your ballot (which I also find distasteful, but perhaps slightly less distasteful).
How each voter approaches this conundrum is up to them, but it is almost certain to arise in the 2023 Novelette category for every voter who doesn’t read both Chinese and English, as only five of the six are currently available in English, and only three are available in Chinese. Personally, this will not be the first time I’ve voted in a category without reading all six finalists, and I’ve come up with my own standard for handling such situations. Anything I’ve rated 17/20 or higher (the score where I start rounding up to five stars) will be ranked on my ballot regardless of whether I’ve read everything. But if I haven’t at least attempted the entire set, I will leave the remainder of my ballot blank after ranking those five-star stories. I expect the average Hugo finalist to come down in the vicinity of 15/20, and I see no reason to prefer a 15/20 story that I happened to have read over a likely-worthy story that I have not. Even if there is a work I’d prefer to rank below No Award, I won’t rank No Award unless I’ve attempted everything.
My rules are my own, and I don’t wish to force them on anyone else. But I encourage other voters to understand the consequences of leaving works off the ballot and make informed decisions, rather than just ranking everything they’ve read. It’s important to remember: Not Ranked is not Neutral.