Which Movie Do the Stats Truly Favor for Oscar Night?

Guest post by Awards Daily stats wiz Claudiu-Cristian Dobre

As some of you already know, I have been, for quite a few years now, employing and simultaneously perfecting a system to predict the Academy Award for Best Picture via strong statistics based on major industry awards alone. Guild wins and nominations and Oscar nominations. The logic behind this is obvious: these being the groups that have by far the most voters in common with the one that decides the winners in all categories at the Oscars, the branches of the Academy itself (who decide the nominations) and the key guilds, which all feature many of the same voters – there is also BAFTA, but their tendency to at least somewhat favor local talent (not to mention their new procedures for determining nominations in a lot of the major categories) skews their results and stats just enough to bring into question their usefulness – it makes sense that, whatever clues from outside awards bodies may suggest, if the tale told by the wins and nominations from these industry groups is different, it is the one to which one should pay more heed. The percentages of correlation bear this out.

The idea behind this project is to use the strongest of these stats one can find to build a consistent method of reading their output and the story it’s telling, each year, based on the outcomes of previous Best Picture races, in order to then better predict the outcome of future races. Briefly, this is how the system in question works: a series of elimination rules (stats so strong that, barring any equally strong stats saying the opposite, can virtually guarantee that a certain contender pretty much cannot win Best Picture, under normal circumstances) whittles the field down to the likeliest few, after which a comparison of the performances by the remaining contenders in the major guilds and Oscar nominations, which I like to call the “branch count,” followed by tiebreaks, whenever necessary, determines the likely winner.

This method correctly identifies the eventual winner of all Oscar Best Picture races since 1990, when the Producers Guild first handed out their award – needless to say, a key factor in the system. (Some of those who will be reading this – such as Sammy, Idle Time or Greg Feasel – also have very interesting and logical systems for predicting Best Picture, along the same lines, but using some different stats, in places, not using others, and generally based on more recent results. However, I prefer to look at the entire run of the PGA, because I feel that the early years, before SAG and, later, before the reintroduction of the preferential ballot, also offer up very important lessons about how all of this works.) The long-term goal is to maintain this 100% accuracy level, so that none of the clues of the past, including new ones that might appear over time, are ignored when trying to predict each new race. The system has not been modified since the last time it, in its previous form, failed to predict the right winner, Green Book, in one of the most bizarre races of all time, and has, in the meantime, correctly called both Parasite and Nomadland. These were, no doubt, both less unusual races, much easier to call.

This year, given CODA’s unprecedented trajectory (Little Miss Sunshine is the oft-cited closest precedent, but even that was different in a number of ways), a situation has arisen which seems to warrant an anticipatory update of the system. Precisely because a very strong stat has come into play which had not quite been necessary to correctly “calculate” any of the previous winners, even though its influence could be felt in many other cases – most famously, Roma’s. (Which, however, also showed enough weakness in other ways, already covered by the system.) It should be specified that, in its current version, the system predicts The Power of the Dog to win Best Picture, because all other nominees fall to various elimination rules. (CODA, among other things, due to the stat that no Best Picture winner, since film editing was introduced as a category at the Oscars, nearly 90 years ago, has failed to be nominated in both that category and directing.) However, its failure to win any major guild awards besides the DGA indicates unusual weakness, enough to suggest that it, too, is historically weak and, thus, unlikely to win Best Picture.

It is, therefore, not surprising that the addition to the system of the stat I am about to present, which I had previously not paid enough attention to – as it had never been necessary before – would then make CODA the system’s pick and, thus, arguably, the statistical favorite to win the prize at this year’s ceremony. Paradoxically, no doubt, given its truly historical statistical obstacles – three nominations, compared to twelve for The Power of the Dog, no DGA nomination, the directing and editing snubs at the Oscars, and so on. (However, there is historical evidence that being ruled out by just one powerful stat is no better or worse than being ruled out by many, as long as the front-runner is also ruled out by at least one stat of comparable strength.

Not irrefutable evidence, it’s true, which may end up being the crux of the matter this year. We will know more very soon.) A situation would arise that has only come about once before, precisely the year of Green Book vs. Roma, in which all contenders displayed such weakness that the system deemed them too unlikely to win, yet one had to win… And the one that performed the best in terms of guild wins (Green Book was the only Best Picture nominee, besides Bohemian Rhapsody, that won at least two major guild awards – the PGA and SAG supporting actor – with the latter, however, displaying too much weakness in terms of nominations, and losing the count because of it.)

The stat in question is this: all Best Picture winners since Out of Africa won at least two major guild awards on their way to the big win on Oscar night. (So, the PGA, DGA, WGA, ACE or SAG, acting or ensemble.) In other words, it appears that, in order to become a Best Picture winner, a nominated movie must “prove” that it has significant support from more than one major branch. The Power of the Dog has failed to do so this year – it lost the PGA as a favorite, it earlier failed to win an acting award at SAG (despite Kodi Smit-McPhee having been the favorite up until that point), it failed to win the ACE (King Richard did, instead, in an upset) and was ineligible for the WGA, which does introduce some doubt (even though several Best Picture winners that were ineligible there in the past have all managed to win two or more big guild awards in spite of it), but, then again, it did lose the BAFTA, the Scripter and the Globe in screenplay (it only managed to win the Critics Choice), so, statistically, it simply would not have been the favorite to win the WGA, in any case, even had it been eligible. Thus, its only major guild win remains Campion’s DGA award. (There is a lone precedent for a movie that lost the PGA, ACE and all of its SAG races, as The Power of the Dog has, to win Best Picture at the Oscars, and it’s The Departed beating Little Miss Sunshine. The Departed, however, had an ensemble nomination at SAG and did win the WGA. But it’s maybe a good enough precedent to justify a possible win, in spite of this stat.)

With The Power of the Dog thus also “eliminated”, it comes down to the aforementioned “branch count” between all of the nominees, and CODA, with its four out of four major guild wins (PGA, WGA, SAG acting and ensemble), a performance that is also unprecedented, unsurprisingly compares favorably to all others. Against The Power of the Dog, specifically, it came out worse in directing and editing, where it was snubbed by both the guilds and AMPAS, but it outperformed it in picture (due to the PGA win) and acting (dominating SAG), with writing, based on the WGA results, as previously mentioned, at worst, tied, in which case the first tiebreak would come into play, which is the number of SAG wins. (There is ample historical evidence that close races are almost always decided in favor of the movie with a better showing at SAG.)

If The Power of the Dog nevertheless prevails – which is definitely possible, but, I do feel, not too likely, given the nature of the two movies and the level of passion they are each likely to elicit – proving that this stat can be overcome, then the system will have to learn this lesson, from this unprecedented situation, and revert to its old form. Which would, of course, since I have decided to perform this modification prior to the Oscars, this time around (when, in the past, I have always waited for that year’s outcome), represent a failure, but not a deal-breaker, by any means. In other words, all of these claims that “the stats are dead” are greatly exaggerated, as usual… As long as one of these two movies wins. Something else winning might complicate matters quite a bit.

Historical aside: Out of Africa, the last Best Picture winner which does not quite meet this requirement of having won at least two major guild awards (it lost the DGA, WGA and ACE, the only ones around in those days), would have been a strong favorite for the PGA win, had there been such an award available to win at the time, based on its Globe drama win and Oscar nominations lead. There is a very high correlation between those two factors and the PGA win. It probably would not have won SAG Ensemble, but might have won an acting award there, as it did collect a supporting prize at the Globes (a win it did not repeat at the Oscars). Also, with PGA giving it its top prize, it’s also possible the DGA result might have changed. These groups do influence one another, this much is pretty clear. There are many ways in which it might have won a second, in addition to the likely PGA win, had all of the guild awards of today been in play. Therefore, I don’t think this can be cited as a real exception. Also, most of the Best Picture  winners prior to the debut of the PGA Awards won at least two at DGA, WGA and ACE, anyway, and the ones that didn’t might easily have won the PGA and/or a SAG award, had those been around. It can thus be argued that, while this rule is 100% since 1986, anyway, it could have been 100% since at least 1950, had the PGA and SAG awards both begun their history as early as the DGA, WGA and ACE awards did.

All of the Best Picture winners since 1990 that won awards from no more than two of the five major guilds:

Nomadland (won PGA+DGA; lost ACE and SAG acting; not nominated for ensemble; not eligible for WGA)

Green Book (won PGA+SAG acting; lost DGA, WGA and ACE; not nominated for ensemble at SAG)

The Shape of Water (won PGA+DGA; lost WGA, ACE and SAG acting; not nominated for ensemble)

Moonlight (won WGA+SAG acting; lost PGA, DGA, ACE and SAG Ensemble)

Spotlight (won WGA+SAG Ensemble; lost PGA, DGA and SAG acting; not nominated by ACE)

12 Years a Slave (won PGA+SAG acting; lost DGA, ACE and SAG Ensemble; not eligible for WGA)

Million Dollar Baby (won DGA+SAG acting – two acting wins, in fact; lost PGA, WGA, ACE and SAG Ensemble)

Gladiator (won PGA+ACE; lost DGA and SAG, both ensemble and acting; not nominated by WGA)

Shakespeare in Love (won WGA+SAG – both ensemble and acting; lost PGA, DGA and ACE to Saving Private Ryan)

Braveheart (won WGA+ACE; lost DGA to Apollo 13; not nominated by PGA or SAG, in any category)

Unforgiven (won DGA+ACE; lost PGA+WGA to The Crying Game)

Driving Miss Daisy (won PGA+WGA; not nominated by DGA or ACE)

All other Best Picture winners in this span of time collected wins from at least three of the aforementioned major guilds. Interestingly, the only other movie this year that has won an award from at least two of the five groups in question is King Richard. This doesn’t mean it’s the only possible upset, or even that it is a possible upset. It’s just something to consider. Yes, its combination of wins (ACE+SAG acting), as can be seen, has not resulted in a Best Picture win before. But, then again, neither had the combinations achieved by Driving Miss Daisy, Braveheart (quite curiously, given how old those particular two guilds it won are), Shakespeare in Love, Gladiator, Million Dollar Baby, 12 Years a Slave, Spotlight (since it didn’t also win an acting award at SAG, like Shakespeare in Love had), Moonlight (since it didn’t also win ensemble) or The Shape of Water. So, firsts in this respect are most certainly not unusual…