Oscar ballots are now in the hands of the accountants.
But what does that mean? What, exactly, are those guys (and gals) from PricewaterhouseCoopers doing with the choices that 6,687 potential voters have made in 24 different categories?
First of all, the Academy likes to talk about its move to online voting, but that doesn’t mean it’s moved to online counting. If so, everything could have been plugged into a computer, the count would be done by now and the PwC ballot team could take the rest of the week off.
But no, they still count the old-fashioned way — by hand, on sheets of paper. Which means that every online vote is first printed out, so the accountants have something tangible to work with.
That way, there’s a paper trail and the process is not susceptible to hacking — unless by hacking you mean taking an axe and breaking through the door into the secret room in the celebrated “undisclosed location” where the ballots are kept.
(That location, by the way, has been rumored to be everywhere from Chicago, which seems unnecessarily distant, to PwC’s downtown Los Angeles office, which would fall into the “hide in plain sight” strategy.)
Inside that location, the PwC team — leaders Brian Cullinan and Martha Ruiz, plus more than a dozen other staffers, none of whom see more than a portion of the ballots in any category — go to work stacking piles of paper on tables, because that’s how the preferential system of vote-counting works.
Preferential, or instant-runoff voting, or single transferable vote, can be difficult to understand, and I’ve tried to explain it in several different ways over the years. This year, to illustrate how it works, I’ve taken the 247 critics’ Top 10 lists compiled by Metacritic, and counted them the same way PwC will count Oscar Best Picture ballots.
Critics and Oscar voters are two very different breeds, and this is not meant to be a prediction of which films will be nominated for Best Picture, or how many will be nominated. (Spoiler alert: only six if we go by the critics.) Instead, it’s simply an illustration of how the system works for nomination voting in the Best Picture category.
Apologies in advance: This is going to get dense, and math will be involved.
For starters, Metacritic compiled 247 lists, but the critics who came up with 19 of those lists did not rank their choices. Those ballots were therefore useless in the preferential system, which requires voters to list their choices in order of preference.
One key thing to understand at this point: A voter can list five films, but he only has one vote. So except in the case of the Surplus Rule (we’ll get to that in a minute), he or she is listing five movies but only really voting for one of them.
I started with 228 usable lists. The first thing I needed to do was generate a magic number that will guarantee a nomination. To get that number, you take the number of nomination slots available (for Best Picture, that’s up to 10) add one (11) and divide the total number of ballots by that number: 228 divided by 11 is 20.73. You round up to the next whole number, which is 21. (If the number you get after dividing is already a whole number, you still go up one.)
That means that to guarantee a nomination, a film needs 21 first-place votes — because if 10 films got that many votes, it’d be impossible for an 11th to reach the same number.
With that number in mind, I separated the ballots into stacks based on the film ranked first on each ballot. A sizeable 52 different films got first-place votes, a tribute to the eclectic nature of film critics. Interestingly enough, those 52 included “Deadpool,” “Rogue One: A Star Wars Story” and “Moana,” but not “Hacksaw Ridge” or “Zootopia.”
At this point, any film that didn’t get at least one first-place vote was out of the running.
Only two of the 52 films reached the 21-film threshold in the initial count: “Moonlight,” which received 63 first-place votes, and “La La Land,” which received 35.
The Top 11:
“Moonlight” – 63 votes
“La La Land” – 35
“Manchester by the Sea” – 20
“Toni Erdmann” – 16
“Arrival” – 14
“Hell or High Water” – 7
“O.J.: Made in America” – 6
“Elle” – 5
“The Handmaiden” – 3
“Sieranevada” – 3
“Silence” – 3
Before the count went any further, “Moonlight” and “La La Land” were marked down as nominees. And they both triggered a wrinkle called the Surplus Rule, in which a film that gets 10 percent more votes than it needs has its excess votes redistributed.
Here’s how it works: “Moonlight” needed 21 votes to guarantee a nomination, but it got 63 votes. That’s three times as many votes as it needed — so in effect, “Moonlight” only needed one-third of its votes, and people who voted for the film “wasted” two-thirds of their vote.
To keep that from happening, I took every one of the 63 ballots on which “Moonlight” was ranked first, and gave the unneeded portion of the vote, .66, to the film ranked second on that ballot, or to the highest-ranked film that was still in the running.
Example: TheWrap’s Alonso Duralde put “Moonlight” first on his ballot. So one-third of his vote, .33, went to that film, and the other two thirds, .66 (because you always round down) went to the next film on his ballot. But he ranked the Sundance indie “Take Me to the River” second — and that film didn’t get any first-place votes, so it was out of the running.
He had “La La Land” ranked third — but as we know, that movie had already secured a nomination and didn’t need his vote. But his No. 4 choice was “The Lobster,” which was still in the running. So that film got the .66 of his vote that “Moonlight” didn’t need.
“La La Land” also triggered the surplus rule, but it needed six-tenths of each of its 35 votes to reach the magic number of 21. So the unneeded .4 of each “La La Land” vote went to every one of its voters’ No. 2 choice, or the highest ranked film on that ballot still in the running.
The “Moonlight” and “La La Land” surplus redistribution pushed “Manchester by the Sea” past the magic number and made it the third nominee. It also added significantly to the totals of “Paterson” (which got six partial votes from “Moonlight” and one from “La La Land”), “Toni Erdmann,” “American Honey” and “The Handmaiden” (each of which got five from “Moonlight” and one from “La La”) and “O.J.: Made in America” (five from “Moonlight,” none from “La La”).
With “Moonlight,” “La La Land” and “Manchester by the Sea” locked in as nominees, here were the new, post-surplus standings:
“Moonlight” – *NOMINEE
“La La Land” – *NOMINEE
“Manchester by the Sea” – *NOMINEE
“Toni Erdmann” – 19.7
“Arrival” – 14.8
“O.J.: Made in America” – 9.33
“Hell or High Water” – 8.46
“The Handmaiden” – 6.7
“Elle” – 6.46
“American Honey” – 5.7
“Paterson” – 5.36
At this point, it was time to redistribute every ballot whose first choice had less than one percent of the vote (i.e, less than 2.28.) More than half the films, 31, fit this criterion.
Along with the eight films listed above that hadn’t been nominated, 10 other films remained in the running with more than one percent of the vote. So on each of the ballots for the 31 films that didn’t make the cut, the first choice was eliminated, and a full vote shifted to whichever of the 18 remaining films ranked highest on the ballot.
Of course, some of the ballots didn’t include any of the films that were still in the running, either because the voter didn’t indicate enough choices or because all the movies they listed either had already received nominations or had been eliminated from contention. (Note: Because Oscar voters only get five nominating slots, I didn’t look beyond the No. 5 film on each critic’s list.)
Seven ballots didn’t have any of the 18 remaining films in their top five, so they had to be discarded. If Oscar voters get too esoteric in their choices, their vote doesn’t count at all.
After the other ballots were redistributed, “Toni Erdmann” got six more votes, pushing it past 21 votes and making it a nominee. “Arrival” and “American Honey” got three more, seven different films got two more and three got a single new vote.
This was the final count, and the new standing:
“Moonlight” – *NOMINEE
“La La Land” – *NOMINEE
“Manchester by the Sea” – *NOMINEE
“Toni Erdmann” – *NOMINEE
“Arrival” – 17.8
“Hell or High Water” – 11.32
“O.J.: Made in America” – 10.33
“American Honey” – 8.7
“The Handmaiden” – 8.7
“Elle” – 8.46
“Paterson” – 7.36
At this point, the magic number changed. The Academy rule that’s designed to produce between five and 10 nominees says that at the end of the count, any movie with more than 5 percent of the vote is a nominee, and any movie with less than that is not.
Five percent of 228 is 11.4, which means that only “Arrival” qualified. But remember those seven ballots we had to discard because there was nowhere for their vote to go? That reduced the total number of ballots to 221, which reduced the magic number to 11.05 … which meant that “Hell or High Water” sneaked in under the wire to become the sixth and final nominee.
And that’s it. The critics only gave us six nominees, which is two fewer than this process has ever produced with the Academy. That’s likely due to the particular nature of critics, who can spread their affection among a broader range of often obscure movies, making it harder to amass enough love to secure a nomination.
By the way: The vast majority of the voters, 155 out of 228, helped the No. 1 choice on their ballot secure a nomination, although almost two-thirds of those also helped a film lower on their ballot. Another 40 helped a film get nominated with their second or third choice, a dozen did so with their fourth or fifth choice, and less than 10 percent of the ballots with ranked choices did not contribute to any film’s victory.
In almost every category except Best Picture, the redistribution does not stop after one round, and the 5 percent rule never comes into play. Instead, the film with the fewest No. 1 votes is eliminated, all of its ballots are redistributed to the highest-ranked film on that ballot that’s still in the running, and the process is repeated, over and over, until only five nominees remain. (Or three nominees in the Best Makeup and Hairstyling category.)
In a way, that process is more straightforward, more time-consuming and also more forgiving. The way the Best Picture category works, a film really needs a lot of No. 1 votes in the first round to get within striking distance of the magic number almost immediately; with only one round of redistribution, there’s very little chance to move up.
But in other categories, in which the count can go on for 10 rounds or more, there’s far more chance to pick up steam, and votes, over a longer period of time.
That’s what PricewaterhouseCoopers staffers will be doing for much of the week, until they reveal the slate to the Academy the night before nominations are announced.
We’ll find out on January 24 if Academy voters are more generous than film critics. Chances are, they will be.