Additional Thoughts on Models

Additional Thoughts on Models

After some additional thinking on the concept of models from yesterday’s post and reading an article on the mind-blowing performance of The Medallion Fund at Renaissance Technologies, I have some additional thoughts to add.

This was been the performance of the fund:

Additional Thoughts on Models

This is how the fund started:

Initially he bought and sold commodities, making his bets based on fundamentals such as supply and demand. He found the experience gut wrenching, so he turned to his network of cryptographers and mathematicians for help looking at patterns: Elwyn Berlekamp and Leonard Baum, former colleagues from IDA, and Stony Brook professors Henry Laufer and James Ax. “Maybe there were some ways to predict prices statistically,” Simons said in a 2015 interview with Numberphile. “Gradually we built models .”

At their core, such models usually fall into one of two camps, trend-following or mean-reversion. Renaissance’s system had a foot in both. Its results were mixed at first: up 8.8 percent in 1988, its first year, and down 4.1 percent in 1989. But in 1990, after focusing exclusively on shorter-term trading, Medallion chalked up a 56 percent return, net of fees. “I was confident that the models would work better,” says Berlekamp, who returned to academia in 1991 and is now a professor emeritus at the University of California at Berkeley. “I didn’t think they would be as good as they were.”

This spectacular performance over almost 30 years was due to creating a complex model that predicted the behaviour of movements in asset prices. It clearly works.

The Medallion Fund has as much genius running the investments as Long Term Capital Management did.

At the 2013 conference, Brown referenced an example they once shared with outside Medallion investors: By studying cloud cover data, they found a correlation between sunny days and rising markets from New York to Tokyo. “It turns out that when it’s cloudy in Paris, the French market is less likely to go up than when it’s sunny in Paris,” he said. It wasn’t a big moneymaker, though, because it was true only slightly more than 50 percent of the time. Brown continued: “The point is that, if there were signals that made a lot of sense that were very strong, they would have long ago been traded out. … What we do is look for lots and lots, and we have, I don’t know, like 90 Ph.D.s in math and physics, who just sit there looking for these signals all day long. We have 10,000 processors in there that are constantly grinding away looking for signals.”

In addition to language specialists, astrophysicists have historically had an outsize impact on the system’s success, according to people familiar with the firm. These scientists excel at screening “noisy” data. String theorists have also had a major role, and the Della Pietra brothers—who reunited with their former IBM bosses to work on equities—were the first of many with that background. The identical twins, now 56, have never strayed far from each other: They took an honors science program at Columbia University as high school students; attended Princeton as undergraduates, studying physics; and received doctorates from Harvard in 1986.

So it’s not the models forecasting complex systems per say that’s really the issue, but hubris mixed with large doses of leverage that undid one brilliant fund while the other has been able to keep a long, consistent streak of incredibly out-performance.

In the early days, anomalies were easy to spot and exploit. A Renaissance scientist noted that Standard & Poor’s options and futures closing times were 15 minutes apart, a detail he turned into a profit engine for a time, one former investor says. The system was full of such aberrations, he says, and the scientists researched each of them to death. Adding them all up produced serious money—millions at first, and before long, billions.

Long Term Capital Management’s model ran into problems because they employed immense amounts of leverage that blew up the operation when it ran into problems that the model predicted would essentially not occur, or at least had an ultra low chance of occurring in a lifetime.

Therefore, it seems leverage and hubris are the main detriments to one’s financial health, not so much models, especially if they are constructed and operated correctly.

Additional Readings