Numlock Awards: The Golden Experiment
Numlock Awards is your one-stop awards season newsletter, and it’s back! Every week, join Walt Hickey and Michael Domanico as they break down the math behind the Oscars and the best narratives going into film’s biggest night. Today’s edition comes from Walter
Hello! I’m back.
This week we’re going to get a monumental amount of clarity about Best Picture, with both the DGA and PGA nominations dropping Thursday. This will have a bit of a field-clearing effect, and will give us in pretty broad terms the movies that have any chance (movies with PGA noms) and are seriously in the mix (The four or five that will get both DGA and PGA noms).
I don’t want to talk about that yet! I want to talk about a fascinating experiment that is happening this year that will help us actually understand why it is that precursor awards are predictive.
Precursors are good at predicting the Oscars
If you’ve been following this newsletter or my Oscars work in general for the past several years, you know that this forecasting model is pretty much entirely based on the idea that there’s a whole bunch of awards that happen right before the Oscars, that these awards are often voted on by industry groups and critics’ groups, and that by looking at their track record and who they’re coalescing around in a given year you can get a pretty good sense of who’s going to win the Oscar.
It’s not perfect, but let’s be honest, it’s really effective: we can more often than not pick the winner in a given category, and when we get it wrong (as we definitely do sometimes!) it usually goes to the second-place entrant in the model output.
No matter which way you cut it, that tells us that this is working, that this model incorporates a productive set of inputs that reliably translates into projections about who is going to win an Oscar. It’s usually right, and when it’s not right, it’s close.
What it doesn’t tell us is why.
Wait, but why are precursors are good at predicting the Oscars?
In my various attempts at justifying our process, I’ve fallen back on a number of arguments. Here they are, in vaguely declining order of how important I think they are:
People who vote on precursor awards are the subjects of the same campaigns that Oscar voters are subjected to, so their votes in precursor awards can be good proxies for how effective different campaigns are.
People who vote on precursor awards have similar inherent tastes to the Oscar voters.
In many of the precursor awards, specifically the Acting, Directing, etc. Guilds, the very voters for those precursors are eventually Oscar voters.
Some precursor awards are massive television events, and can influence voter perceptions the same way that any large television event can.
Precursor awards give free minutes of air time to winners in the form of acceptance speeches, minutes those winners can use to construct an argument or case as to why they are deserving or charming.
Oscar voters look to precursor awards to get a sense for who has an actual chance to get a nomination and inform their ballots accordingly.
Cool. Obviously, not all awards have all of these qualities in similar amounts. The BAFTAs definitely have 1, 2, and 3 in spades; however, because of timezones, the BAFTAs don’t have the primetime TV audience that gets them 4 or 5. The SAG Awards put up decent TV numbers, and so are pretty great for all six. The number of Critics’ Choice Award voters with an Oscar ballot is functionally nil as far as I know, so they aren’t very good for 3, but they’re great for 1 and 6 and arguably 2 and 5. The DGAs are straight-up not televised, but regularly call Best Director, that’s classic 2 and 3.
This brings us to the main point of this post: the Golden Globes experiment.
The Golden Globes this year may be able to tell us why precursors are good at predicting the Oscars
The Globes shouldn’t matter. Before getting nuked by the publicists, they were 87 members of the foreign press who have no direct impact on the Oscars but nevertheless commanded one of the best time slots in television and miles of print coverage of their results.
I have long considered them dumb but sometimes useful. In fact, here are some cruel things I have said in the past about the Golden Globes, in print:
something to dangle “if you want to get Amy Adams to show up to your party”
I regret nothing.
Now, while the Globes don’t matter, they also do matter a little. Their predictiveness is above a random monkey with darts on the wall. We keep them in the model. In some categories, like Best Supporting Actor, they’re actually really useful!
And this year, we might actually learn why. The differentiating factor has long since been the enormous television audience of the event. Let’s go back to the list of reasons why a precursor might matter, and run through the assets that the Golden Globes have historically had:
Voters subject to same campaigns ✅ TRUE
Similar inherent tastes to the Oscar voters ❓MAYBE
Voters are eventually Oscar voters ❌ FALSE
A massive television event ✅ TRUE
Free air time to winners ✅ TRUE
Oscar voters look to them ❓MAYBE
This year, they lost their time slot. Nobody showed up! It was a hotel dinner during the worst of Omicron. They had no host and no press. This is what that looks like this year:
Voters subject to same campaigns ✅ TRUE
Similar inherent tastes to the Oscar voters ❓MAYBE
Voters are eventually Oscar voters ❌ FALSE
A massive television event ❌ FALSE
Free air time to winners ❌ FALSE
Oscar voters look to them ❓MAYBE
So naturally, we’ve got a little experiment on our hands. The Globes matter a little bit, but will they this year? As with all experiments, it’s typically best to write out your hypotheses beforehand and write up how you will interpret given bands of results. Let’s do that. I’ll focus on Best Supporting Acting nominees to give tangible examples.
Looking at the supporting actress nominees, the SAG and Globes differed only on Cate Blanchett (SAG nominated) and Aunjanue Ellis (Globe nominated). The Critics’ Choice nominated Ann Dowd and Rita Moreno, but not Blanchett or Ruth Negga.
In the supporting actor category, they differed more, with SAG nominating Bradley Cooper and Jared Leto while the Globes nominated Jamie Dornan and Ciarán Hinds. The Critics’ Choice also had J.K. Simmons, and nominated Dornan and Hinds over Ben Affleck and Cooper.
Outcome 1: The eventual Oscar nominees don’t really resemble the Globes nominees but do resemble the nominees of other awards shows
This would be a situation where Bradley Cooper, Jared Leto, J.K. Simmons, Cate Blanchett, Ann Dowd and Rita Moreno pick up a few nominations at the expense of Ellis, Dornan or Hinds. It remains to be seen how the Critics’ Choice — which is picking up a better time slot — or the Guilds come down on winners, but if their winners do well while the Globe winners don’t that’ll be this outcome.
In the situation where the Globe winners don’t really seem to rise to the top of the pack as much as we’d seen before, this could be a signal that the mojo of the Globe was really packed into the time slot and the ability for winners to come off as incredibly charming to an audience of millions on television. It would mean that the “Golden Globes voters are good at picking winners” theory has been losing power.
Outcome 2: The eventual Oscar nominees resemble the Globes nominees in ways that don’t resemble the nominees of other awards shows
This would be where Hinds, Dornan and Ellis get nods, and the Golden Globes somehow maintained their predictive power even despite being stripped of their time slot. It’d also be this if Globe winners The Power of the Dog and West Side Story have legs in Best Picture while not being recognized at other awards shows.
This situation would bolster the Globes as a proxy for Academy taste, sort of the Georgia Primary of awards season in that it’s a small group but because the small group reflects broader tastes they’re worth paying attention to. I do not think this will happen.
Outcome 3: Mixed bag
This seems likeliest. Our problems in predicting the Oscars are always ones of sample size, and this experiment is no different.
My bet is that the Globes kind of fall apart in a couple categories — the lead acting and the Best Pictures were never really their strengths, and the campaigns for the top prizes likely avoided the voters like skunk roadkill — but that the Globes are useful in a few others.
On some of the edge categories like score, like Animated Feature, like song and in the Supporting Actor categories where people can run on their reputation and where the campaigning might be a little less direct and and where character actors who really don’t get too many bites at the Oscar apple might have done a little more campaigning? Maybe the Globes keep their fastball.
This would bear out the “Globes voters as a good test audience for success of campaigning.” If their preferences in categories where they were actually campaigned bear out, while their preferences in categories where they were ghosted do not, that underscores that their main utility is as a bribe-able focus group rather than a telecast or an assembly with actual useful taste. And that’s where my personal hunch has long been.
What’s it all mean?
If the Golden Globes are “right,” in a couple categories outside the big ones, then that’s evidence their usefulness is because they’re a focus group for how effective an Oscar campaign is. If they’re good across the board, even in big categories where nobody in their right mind would have campaigned to them, it might even be a sign they have taste.
If the Golden Globes are “wrong” everywhere, then it’s a sign that it was the telecast all along, and that the first 87 names in the phone book could replace the Hollywood Foreign Press Association membership rolls and, as long as they got to keep the time slot and producers, would play functionally the same role.
We’ll find out in two months. Have a great week!