Best Picture: Awards Daily Simulated Ballot Nominees!

Here’s Dr Rob’s Best Picture chart that I meant to post 12 hours ago before the day got away from me. By now you’ll either follow what’s happening with the redistribution of bouncing ballots from stack to stack or else nothing I can say in any intro will be of any help. There’s a really good discussion around the Directors Chart where Rob summarizes his methodology. After the cut is where you’ll find the full-size expandable BP chart too.

I struggled to get my head around the reason why 180 Actual Oscar ballots is a magic number that any movie needs to stay in play in the first round of counting. Marshall, Rob and myself worked through my confusion yesterday:

Marshall:
My observation – a film needs at least 3% of initial #1s in the first round (before any votes are redistributed) to have a shot at hitting the 5% required for the nomination

So that would mean roughly 180 #1 votes out of the 6000 cast in Round 1.

So you need 300 (5% of 6000) for the nomination after all vote redistributions have been completed (either from surplus or transfers from eliminated films). The way Dr. Rob and I perform preferential balloting, we consider this “Round 3.”

The 3% I’m referring to (~180 voters) is from the initial count of ballots (Round 1), where all films are sorted by the amount of #1 votes they received.

I’ll share a chart of multiple simulated ballot results (both Dr. Rob’s and mine) later highlighting this observation. But 3% looks like a pretty decent predictive hueristic.

Ryan:
Alright, so if I understand correctly,

According to Marshall’s findings and formulas, and now backed up by Rob’s simulation, a film needs at least 180 raw #1 votes in the first round of voting to remain in play.

Less than 180 ballots with a #1 vote, and it apparently becomes next to impossible to accumulate enough ballots to reach the other key number of 300 — that’s the minimal number needed after redistribution in subsequent rounds, the number needed to achieve the winning threshold for final inclusion .

a film only needs about 180 voters marking it as #1 to stay in the game
but a film does need 300 ballots in its “stack” after the accountants are finished dealing them out through all the subsequent rounds. Is that right?

Marshall:
You’ve got it.

Here’s a chart that should help show the 3% rule-of-thumb in action:

The only real exception to this heuristic has been Imitation Game. Yes, one needs to keep in mind the small sample size (only four years of data), but 3% does seem to be a decent rule-of-thumb given how the vote transfers tend to consistently be all over the place.

Rob:
Get ready for some math!

3% is an excellent target. If a film has 3% then the likelihood of a nomination goes up. This is predicated on the amount of surplus being redistributed—whether thats from a number of films triggering it or a landslide surplus.

This MM:FR, Spotlight, and Carol accounted for 47.4% of the ballots. Now (equivalent to) 20% of the votes were redistributed (or 282 votes), leaving approximately 27.3% remaining for the top 3. 3% of the initial round is equivalent to 5.7% of the non-top 3 films. Assuming that the redistributed votes follow the same distribution that the first round did (excluding top 3) means that a 3% vote will become a 5.7% vote. The Martian started with 42 votes which is 1 vote away from 3%. Using this scaling factor would mean 81 votes in the end. It received 78 votes in the end. This does support the 3% rule.

However, there are some issues here. This year there were no films that made the 9.09% in the first or second rounds without triggering the surplus. This would boost the large number of films above 3% up. If a film after round one meets the 9.09% rule but does not trigger the surplus then a lot more than votes are tied up with that title. If too many films meet the 9.09% rule then again too many votes are tied up. If too few meet exceed the 9.09% and trigger the surplus vote then not enough votes are redistributed.

The goal of 10 nominations would need a Goldilocks distribution: 1) A small number of films which meet the surplus requirement in a sizable way and 2) the rest of the field hovering around the 5% (down to 3%)

That happened this year.

Last year, Gone Girl did not meet the surplus vote but qualified in the first round. Boyhood had its votes redistributed at .68 weight. In the end equivalent to 24% of the votes were redistributed. What happened to throw things off is Grand Budapest and Whiplash. 1) They both met the 5% at 7.2% and 6.3% respectively but remained in play to draw a sizable portion (19%) of those redistributed votes. 2) In addition to syphoning off those redistributed votes they tied up the votes above the 5% that will ultimately net them a nomination. Combining these two factors, those two titles—which were going to get nominated anyways—tied up nearly 100 votes from the sub 5% films. This is another reason why we should hate Whiplash. If The Imitation Game didn’t have an unusual surge then there would have been only 8 nominations.

###

Links to PDFs of the complete set of ballot charts

Best Picture
Best Director
Best Actress
Best Actor
Best Supporting Actress
Best Supporting Actor
Best Original Screenplay
Best Adapted Screenplay
Best Cinematography
Best Editing