Considering the close strategic friendship between the autocratic governments of Russia and China – including a shared enemy in the United States – who could have predicted that Vladamir Putin would give a televised interview to publicly speak against Chinese military action in its territory dispute with Taiwan?
Well, Jonathan Grady for one.
The MBA’ 23 from NYU Stern is accustomed to making predictions (or at least designing models that make predictions) that go against conventional wisdom. And 90% of the time, Grady says, the predictions turn out to be correct.
Grady is the founder of The Canary Group, a “leading game theory service that can forecast and shape complex bargaining environments for more winning outcomes.” In other words, Grady uses the famed game theory method created by Bruce Bueno de Mesquita – professor of politics at NYU and senior fellow emeritus at the Hoover Institution at Stanford – to forecast everything from complex geopolitical negotiations to corporate mergers and acquisitions. Grady’s unique approach to using his mentor’s model has garnered him attention from media outlets like the Wall Street Journal to C-suite executives of Fortune 500 companies.
Most recently, he worked extensively with CNBC on “the largest computation ever run by the Bueno de Mesquita model in its history — more complex than any projects undertaken for the CIA or corporate clients,” according to the business media outlet.
CNBC’S ‘THE QUAD’ PROJECT
In February, CNBC began an ambitious reporting project which sought to answer a question – “What is the future of the Quad?” – with the help of advanced game theory. “The Quad” is the nickname of the Quadrilateral Security Dialogue, an informal alliance between the United States, Australia, India, and Japan to counterbalance China’s growing economic, political, and military might.
CNBC deployed Grady to build the forecasting model. In between classes, MBA commitments, and running The Canary Group, Grady gathered detailed input information from 37 policy experts and former government officials, compiling a forecast that included almost 300 individual players from the four Quad nations, China and 10 other countries and territories. CNBC published its finished report, “The ‘Quad’ is on the rise in Asia-Pacific: Game theory has a prediction about its future,” on Sept. 23.
CNBC reporter Ted Kemp, who reported project with Grady’s model, explains the purpose of the report and some of its key findings in the video clip above.
As part of our Student Founder series, Poets&Quants recently caught Grady on some rare down time to talk Game Theory, forecasting the moves by Vladimir Putin, and finding time for his MBA.
The Canary Group’s motto is, ‘We forecast future events. Scientifically.’ Can you tell us a little more about what the company does?
The Canary Group is a strategic advisory service using a computationally intensive, data intensive algorithm developed by Bruce Bueno de Mesquita. It is over 90% accurate, per a declassified review from the CIA. As the CNBC article pointed out, the author of that review later wrote that the model – at the time of his writing – had been used at CIA over 1,200 times in over 70 countries across a range of issues. The model was 90% accurate at that large of a sample set. So this is really amazing technology, and I’m doing this as a Stern MBA student.
Something important to note: We have refined our tools since the CIA report that said it was 90% accurate. We are offering a service that is more sophisticated and superior compared to the tools the CIA used at the time.
I offer this service to all sorts of interested people, whether it’s predicting political events or working on corporate strategy, mergers and acquisitions, and that kind of thing. It’s been used by Fortune 500 companies. The value in this sort of approach is that, oftentimes, people do not want to make future predictions. And even when they do, the prediction is going to be very vague. We can offer ex ante, before-the-fact predictions for negotiations, telling clients how different individuals will play the issue out, and from that point, advise the client how to make the outcomes come out closer to her preferences.
You use the model developed by Bruce Bueno de Mesquita, the famed game theorist and model forecaster. How did you learn its methodology?
When I was an undergraduate at NYU, I took Bruce’s PhD class on conflict and game theory. It had been over 30 years since he last let an undergrad do something like that. Well, he taught me how to use the model, and I did all sorts of interesting things with it. Philip Tetlock (Leonore Annenberg University Professor in Democracy and Citizenship at Wharton School) ran a major IARPA forecasting competition, and I out performed a lot of people by saying, ahead of time, that the U.S. would not have troops in Iraq by the time negotiations began between the Obama administration and the Iraqi government. That was a very surprising result because Bush and former Bush administration officials were very confident that the deal would be made, and I was indicating a deal was not going to be made.
This is a model that Bruce developed over 40 years ago, and he used it to help with U.S. government assessments of different political events. As the Wall Street Journal put it, ‘this is the stuff of the Cold War.’ But it doesn’t have to be the stuff in the Cold War. We’re using game theory quite widely.
Something really valuable that has been offered for both clients today and the government back then is that we can forecast not just things that you think are going to happen but surprising events. Being forewarned is forearmed. Once you know that something could happen, you can take countermeasures.
OK, for laymen like myself, can you explain a little about Game Theory? What is it actually?
Game theory is a mathematical approach to decision making. What it ultimately tries to do is put pen and paper to our decision making incentives and work out what direction an individual is going to go. A lot of people, when they talk about decision making, don’t try to formalize it – which is unfortunate. On a basic level, what game theory does is look to explain people’s incentives and motives, work out their decision process, and see what can be done to influence and shape their decisions.
So what are you bringing to the Bueno De Mesquita model that is different from others who have used it?
I’m very artful in how I use this model. I was able to forecast, as an undergrad taking a PhD class, that the Egyptian government would form into a military junta. This was during the Arab Spring. Bruce wrote about that and it ended up being put in a book chapter calling for reforms to the way that we do assessments in foreign policy analysis. I’m really sophisticated with my use of the model, and I use the model as part of my business.
Did you form The Canary Group before or after starting your MBA?
It came a little bit before the MBA. I was always interested in data approaches, and I had even developed a way to infer monthly sales at Costco for a range of products from the model. There was something very particular about the custom model I made that I was able to watch stock prices for individual companies that shelf at Costco moving up and down, according to the data I had ahead of time. It was very interesting.
NEXT PAGE: Using Game Theory to predict world events & more on ‘The Quad’ project