Exclusive Q&A with Ben Zauzmer (Author, OSCARMETRICS)

By Hyejee Bae AB '20

Ben_Zauzmer_picture.jpgBen Zauzmer AB '15 is a Baseball Operations Analyst for the Los Angeles Dodgers and also the author of Oscarmetrics: The Math Behind the Biggest Night in Hollywood. Originally from the Philadelphia area, Zauzmer studied Applied Math at Harvard with a focus in government, where he nurtured his trifecta passion for math, movies, and sports. During his freshman year at Harvard, he devised a statistical model that could calculate the chances a nominated movie would win the Academy Awards. When he’s not on the road with the Dodgers, he is busy utilizing his mathematical approach to forecast Oscar winners and sharing his statistical predictions on Twitter. In 2018, Zauzmer correctly predicted 20 out of 21 of the Oscar nomination winners.

Q. Let’s begin with a brief introduction. Starting from where you are from, to where you are now, could you give a short history of your life? 

A. I’m from the Philadelphia area and I graduated high school in 2011. From there I went to study Applied Math at Harvard with a focus in government. My Oscars work got started my freshman year during the 2012 Oscars, and I graduated in 2015. Shortly thereafter I moved out to LA to work for the Los Angeles Dodgers.

Q. What does your work at the LA Dodgers entail?

A. I work in Analytics. I was in the research and development department my first four seasons from 2015 to 2018. This means I did statistical modeling, trying to predict which players will be good, and then communicating this information to the general manager and other decision makers. This past season in 2019, I transitioned into the Baseball Operations department, which is the liaison role between the research and development folks and the players and coaches. I’ve been very fortunate. I’ve always been a huge baseball fan and obviously, I’m a math fan as well. So, I’m very lucky that I was able to find a job that combines them.

Q. Tell us the story of how you got into predicting the Oscars through statistics. Where did the idea come from?

A. The Oscar prediction work started my freshman year in college. I was sitting at the IOP, after an event there, getting some work done on my laptop after the event had cleared out. Being a big movie fan and a big Oscar fan, my mind started wandering towards the Oscars that year, which were rapidly approaching. Naturally, I found myself wondering who was going to win. As a budding applied math concentrator, I assumed that somebody out there must have used statistics to determine this, to figure out the probability of what nominee would win. It was something I had loved following in baseball, and something I loved following in politics with the 2012 election season coming up. Surely someone on the internet had done the same for the Oscars. But when I went on Google, I couldn’t find anything! So, I decided to do it myself. I practically parked myself in Lamont Library for a good month my freshman year, for the most of February. I gathered as much data as I possibly could, built statistical models, and threw the results of these models onto a quick WordPress website. That became the first year I predicted the Oscars. 

Q. Can you give a brief explanation of the math behind your prediction formula and give an example of how it works?

AI start by gathering data on previous years of the Oscars. And by data, I mean other award shows, which categories a film is nominated in, what the betting marks say, or what Rotten Tomatoes says. Really anything that I can put a number on I gather on all of the previous nominees. Then I use statistics to predict which of these predictors have done the best job of predicting each category in the past. If you’ve done a better job of predicting a category, you get more weight. Then I can take those weights and apply them to this year’s nominees, and this gives me the probability that each of this year’s movies will go on to win a category.

So as an example, take this year’s Big Picture race. You have Parasite, which won the Best Ensemble Award at the Screen Actors Guild. Once Upon a Time in Hollywood won the honor of the Critics’ Choice Award and one of the two Golden Globes. 1917 won the other Golden Globe, the Directors Guild, and the Producers Guild. All of these awards get different weights in the model based on how predictive they’ve been in the past. That gives me the chance that each film wins Best Picture. 

Q. Are there any aspects of your prediction formula that might be surprising to people? For example, are there any hidden biases?

A. For sure! I talk about this in the book, but I get asked a lot about the Box Office. People are asking me on Twitter or in person, “is the box office a good predictor of the Oscars?” And the answer is mainly no. There is a slight correlation, but it’s so slight that it’s practically irrelevant for prediction purposes. What I found is that at one point there was more of a correlation between popular films and the Academy Awards, but that correlation has all but disappeared in the last 15 years.

Q. Have you changed your prediction model every year or does it stay largely the same?

A. The formula structure has been essentially the same for the past three years. What does change are the new data sources, I’m always getting a new year’s worth of data. Even just one additional year can significantly change how much weight these factors get in some cases, because we’re not dealing with that much data to begin with. Many of these precursor awards only go back a couple decades. One year, where one guild gets one category right or wrong, can increase or decrease the amount of weight it gets by non-negligible margins. 

Q. How do non-mathematical factors, such as political or social movements, affect your predictions?

A. My model is able to capture non-mathematical factors because these factors actually affect all of the inputs as well. Say you have someone who is a sentimental favorite because they’ve never won an Oscar before, so perhaps Brad Pitt this year for Once Upon a Time in Hollywood. He has only won the Producer Award for 12 Years a Slave, but never as an actor. If you have someone like that who the entire industry is pulling for to win that Oscar, that can help them in the inputs as well. Something like sentimentality might sound like a non-mathematical input, but it can actually affect the mathematical inputs as well. 

That’s not to say that my model can always capture every sort of social movement or every type of news article that affects the awards race, especially the articles and movements that don’t build up until a week or two before the Oscars, when a lot of these precursors have already concluded. This is the reason that math is not the end-all be-all of Oscar prediction, and why there are many talented and smart people out there that predict the Oscars in a more traditional way, without math. They go on their gut and their knowledge of the industry. That’s why both are valid methods.

Q. What are some of your most confident predictions for the top categories of the Oscars this year?

A. These predictions are based off the data we have at this moment, before this weekend’s Writers Guild and the BAFTAs. The frontrunner for Best Picture and Best Director is 1917. It’s not locked by any means, but currently it is in front. The acting awards are a little clearer this year. You’ve got Joaquin Phoenix in Joker, Renée Zellweger in Judy, Brad Pitt in Once Upon a Time in Hollywood, and Laura Dern in Marriage Story. They have all dominated awards season, winning Critics’ Choice, Screen Actors Guild, and the Golden Globes. Especially if they win BAFTAs this weekend, they would seem to be in very good shape. 

Q. If you had to incorporate some bias and choose Best Picture based on personal opinion, which would you choose?

A. It’s a hard race, it was a really strong year. But if I had to pick the Best Picture this year, I would lean towards 1917. The main reason is that I was just so impressed throughout. Every single scene, I was thinking to myself, “how on earth did they do that?” The ability to tell a compelling war story with all of the plot lines, sets, acting, and screenwriting is incredible. The film wraps it all together in a movie that never cuts, with eight-minute long shots, and weaves it all together from beginning to end. I couldn’t get over how they were able to do so much in a movie. 

So that would probably get my vote, but you could easily talk me into Jojo Rabbit, Parasite, or Joker. Those would be my next top picks.

Q. Are there any predictions for the Oscars this year that would surprise people? If so, why?

A. For the top categories, I really think it is the six that I listed, and then potentially Greta Gerwig for screenplay. I don’t think that any of those would be a major shock though. To get more technical about the model, the math is establishing the probability that each nominee wins. So almost by definition, the math is simply saying, “these are the favorites who are most likely to win.” Upsets are arguably when the math gets it wrong, when the lower probability wins the category. 

Q. I see that predicting the Oscars started out as a passion project. What spurred you to write a book? What was the writing process like coming from a more mathematical background?

Oscarmetrics_cover.pngA. The book started as an idea a year out from college. It just seemed like a logical next step to keep sharing stories about the Oscars through data, with people that enjoy the combination of those two things. Writing the book was actually a ton of fun, it allowed me to learn so much more about the Oscars. Oscar statistics allowed me to find all of these stories through data that I might have never known, stories that hopefully other people will find interesting as well. 

A challenge, but also an exciting part of writing the book, was trying to write a book that was not just a statistic textbook. My goal was not to write a book for mathematicians. My goal was to write a book for movie fans using statistics. So, there are a lot of parts of the book that use math to reach conclusions, but I don’t really get too deep into the weeds of the math. I stick with cursory explanations because I want the book to be interesting and engaging for people that have never taken a math class before but love watching movies.

Q. You seem to be a polymath, harboring a love for sports, math, and movies. How do you balance the love for these seemingly disparate subjects in your life?

A. I guess that is true, I appreciate that! I am very lucky that these two big passions of mine have seasons that dovetail nicely. Baseball season typically starts in mid-February with spring training, and if all goes well, goes all the way until the end of October with the World Series, as it did for the Dodgers in 2017 and 2018. Hopefully it will in 2020 again! Oscar season tends to pick up in November and then goes to around early February. That is very fortunate for me, because my job tends to be closer to 40-45 hours a week during the off-season, and then significantly more than that during the season. Whether the team is at home or on a road trip, it becomes a much busier job. The difference in timing is what makes the balance really doable, to have these two very different passions and to be able to dive myself into the both of them at different times.

Q. All of us currently on the Harvardwood 101 trip are seniors and soon-to-be graduates. As someone a few years out for college, what advice would you have for us moving forward in our own careers?

A. I would say, I get so much happiness every day when I wake up and do the things that I love. For me, that’s combining math and baseball, and combining math and movies. I would strongly encourage any Harvard graduates out there, and also graduates of any school, to do what they love. There are many industries that many people go into because they love them, and there are also many industries that people go into because they feel they should, is readily available, or is more lucrative. People have every right to make their career decisions on whatever they please. I would bet that they would wind up happier if they choose to do something because it is something they love. 

Q. What is your personal definition of success and what are your own goals moving forward?

A. In a broad sense, success to me is “achieving goals.” But those goals can come from all aspects of life. There are professional goals. I want to do my very best work for the Dodgers and I want to see us win a World Series. There are goals for projects outside of work. I want to continue doing this Oscar work, bigger and better every year. I’m excited that I got to write this book so that I was able to share the stories with more people beyond just 280 Twitter characters. There’s personal goals as well, in family life and with friends. All those categories factor into success for me. It’s not just a stereotypical definition of the word “success” that talks about career success only. I strive to be happy in all these different areas, and I’m very fortunate that thus far my life has been good, from Harvard to LA. If it happened to continue to follow along the same path, I would be very fortunate.

Q. You use the words “factor” and “categories” in your last answer. Even the way you characterize success is framed in a statistical model. 

A. I guess I can’t help it!

Calendar Programs Jobs & Career Development