What did Jeff Seder know about American Pharoah?

Can big data transform financial markets and unlock the secret to predicting the stock market?


The Sunday Times 29 July 2018

Download Article

That the pedigree of a horse is an important determinant of how good it will be, has been a long- cherished truism of the horse racing fraternity. But the story of how American Pharaoh came to greatness bears examination.

Known then simply as Hip No. 85, American Pharaoh was one of 152 one-year old horses for sale during the 2013 Saratoga Springs yearling sale in upstate New York. He was drawing attention for all the wrong reasons. He had a tiny scrape on the back of a front ankle that was beginning to swell which seemed to be putting off prospective buyers as his sales price continued to decline. His owner, Ahmed Zayat, was looking to sell the horse and replace him with some others.

Tasked with evaluating all of the horses and recommending which ones to purchase, Zayat hired data scientist Jeff Seder, whose approach to equine appraisal was very contrary. His advice to Zayat was that there was only one horse worth buying at the auction, and Zayat already owned him. “Sell your house; don’t sell this horse.” Seder counselled. Zayat took his advice and bought his own horse at auction for $300,000 (as he was unable to withdraw him from sale).

Eighteen months later, American Pharaoh, thrilled the horse-racing world with his run to the Triple Crown, winning the Kentucky Derby, the Preakness and the Belmont. It was the first time that feat had been accomplished in 37 years.

What did Seder know about horse no. 85 that nobody else seemed to?

In all of Seder’s research, analysing tens of thousands of racehorses over the past three decades, he’s concluded that the pedigree of a horse is wildly overhyped. Breeding is a very weak predictor of how good a horse can be, but a very good predictor of its price he argues.

Seder had used his analytic, data-driven approach and identified a list factors that are dominant in predicting a successful racehorse. But what drove his enthusiasm for no. 85 was the size of his heart; specifically his left ventricle. American Pharaoh’s was in the 99.6th percentile. The data screamed that he was a one-in-a-million horse. The physical reality screamed otherwise.

For a column devoted to investment issues, this topic looks misplaced. American Pharaoh’s story is very similar to the Oakland A’s baseball story as lionised in the Michael Lewis novel “Moneyball”. And the two are analogous to investing in many ways – discovering inefficiencies in pricing and profitably exploiting the difference.

The investment world is grappling with the explosion of data, high performance computing and predictive analytics. In the same way that Seder is trying to uncover undervalued thoroughbreds through studying the data, a vast industry of economists, investment managers and data scientists, are attempting to tie capital market returns to predictable variables. 

Data from satellite imagery, automated drones, people-counting sensors, container ships’ positions, bank transactions, social media, and online search queries are now being used by Wall Street firms to generate data-driven investment theses. Data analytics can even anticipate the level of economic activity in certain areas by calculating the length of the shadow being cast by trucks and tankers.

The sort of discrepancy between price and value that Seder was able to exploit can only co-exist in a market where evaluation methods are poor. Up to now, it’s been hard to argue that there’s much of capital market activity where data methods are sub-standard. Big data may well move the goalposts and that standard to a much higher plain.

Horse racing is predominantly a question of biology. The stock market is a field that is far more complex as it is both biological and psychological. The mere belief that a certain variable affects stock market performance can itself influence its performance – a phenomenon George Soros termed ‘reflexivity’. The size of American Pharaoh’s left ventricle doesn’t change in response to our discovering it. It’s an objective reality. Subjective realities dominate financial markets. Our collective thinking is what moves stock markets and those beliefs directly affect the underlying fundamentals and vice-versa. This makes the job of exploiting inefficiencies infinitely more difficult.

There aren’t any immutable laws in investing. It’s not a hard science like physics or chemistry. So we can’t hang our hat on any particular approach and be guaranteed success. One principle I’m confident that hasn’t changed since time immemorial is, that return is linked to the price paid. Those that forget this principle usually end up victims of stock market wealth transfer –from the overly optimistic to the patient and disciplined.

As good as Seder’s approach has been, there are lots of things that won’t show up in any kind of statistical analysis. A horse’s spirit, a desire or willingness to compete are intangible factors that impact performance but elude measurement. And so it is with investing. Formulas very often fall short of capturing the complexities of capital markets.

The stock market’s equivalents of American Pharaoh are there to be discovered. Data will undoubtedly help, but they are not a panacea. For the moment at least, there’s still room at the investment table for humans.

Gary Connolly is Managing Director of iCubed. He can be contacted at gary@icubed.ie or on twitter @gconno1. iCubed Training, Research and Consulting, trading as iCubed, is regulated by the Central Bank of Ireland.

Contact us today to discover how we can help you build your own range of multi-asset portfolios

Tel: +353 1 6724264

Email: info@iCubed.ie

Sign up for iCubed Newsletter

Receive investment information from iCubed on a quarterly basis.
"Everything we do at iCubed is designed to give financial advisers the best chance of investment success."