As a third-year law student at Harvard, I became fascinated with horse racing (or, to be more exact, with betting on the horses).
In an era when access to computers was severely limited—back then, the room-sized devices were usually found behind closed doors in temperature- and humidity-controlled areas—I discovered an outdated IBM 1620, sitting in the lobby of the graduate chemistry building, that was available to anyone on a first-come, first-served basis.
I’d been to the races once or twice and had read an explanation of the pari-mutuel betting system used at American tracks. In this system, all the money bet on a particular race is placed in a pool, to be divided among those holding tickets on the winning horses. (There are actually several betting pools on each race, including simple bets such as win, place and show; and exotic bets, with names like “quinella” and “trifecta.”)
To construct a simplified example, assume all bets on the first race at Upson Downs are simple $2 bets to win, and that 500 bettors have bought $1,000 worth of tickets. When betting closes just before the start of the race, it turns out that 100 bettors have put their money on Lucky Seven, who wins the race. The track first skims off 15 percent of the pool to cover operating costs, taxes paid to local governments and prize money for the horse’s owner, called the “vigorish,” leaving $850 in the pool for the winning horse. Each holder of a winning ticket then receives back $8.50 ($850 divided by 100), representing each bettor’s original $2 bet plus $6.50 of winnings.
What does this have to do with investing? As they say in the courtroom, “I’ll tie this up in a minute, Your Honor.”
I became interested in whether there was any discernable pattern in how bettors selected their horses that would enable a savvy bettor to make money under the pari-mutuel system. I wrote a program to analyze the daily race results at Suffolk Downs, the race track closest to Boston. Compiling several weeks’ worth of races in my database, I then analyzed the results, limiting myself to the published data and without any information on the horses or jockeys themselves. Those using this approach to analyze investment data are called “Quants”—short for quantitative.
Crunching the numbers, I discovered that, if one bet in the win pool on the second-favorite horse in races having at least eight horses running, one could come out ahead despite the track’s stiff 15 percent rake-off. Over the many races, all bets in the place and show pools proved to be losing propositions, and those betting on the favorite to win came out seriously worse off, regardless of how many horses were running.
Although this analysis took many hours to complete, being a starving student, I never actually went to Suffolk Downs to test my system. But since both the number of horses running is known well in advance, and the pari-mutuel odds on each horse in a race is widely available in real time at the track right up until the betting windows close, I was confident that the system could be implemented as a practical matter. I filed my newfound knowledge away for later use.
Translating my learning into investment jargon, the race track—or at least this race track—was an “inefficient market.” Had it been efficient (as the stock market is alleged to be), factors such as whether or not the horse was the favorite, or what its odds were or how many horses were in the race wouldn’t have made any difference. All betting systems should, in the long run, have produced a 15 percent loss due the vigorish deducted by the track from each betting pool.
I told myself that, after I finished school, I’d test the system in real time at some race track after conducting a similar study of that track’s results. But not having much of a stomach for gambling, I’ve never done so.
Instead, I later became interested in the stock market, and I’ve found the takeaway principal from my fantasy race track betting highly useful here: It matters what you pay for the stocks you buy.
As at the track, stock favorites are unlikely to pay off in the long run. The fact that certain stocks are “favorites” usually means other people are overpaying for them.
Of course, the stock market differs from the race track in that there’s a positive expected return from owning stocks that approximates the growth in corporate earnings—roughly about 7 percent a year since the end of World War II—plus the current dividend. Thus, the index investor who takes positions in all stocks can expect about a 9 percent positive return over the long run, compared to the negative 15 percent return from betting on every horse in every race at the track.
Why should it matter what one pays for a stock? Because, like bettors at the track, every investor has a different opinion of what information is important. And to a greater or lesser degree, every market participant allows a variety of human follies to influence his or her investment decisions. Some of these follies derive from how humans are hardwired to think and behave. Others are cultural, so not all markets can be expected to follow precisely the same pattern of “inefficiency” or, more correctly, “irrationality.”
To invest successfully, it’s important to study each market in which you intend to participate, so you can try to learn what strategies are likely to improve your returns.
Translating my learning into investment jargon, the race track—or at least this race track—was an “inefficient market.” Had it been efficient (as the stock market is alleged to be), factors such as whether or not the horse was the favorite, or what its odds were or how many horses were in the race wouldn’t have made any difference. All betting systems should, in the long run, have produced a 15 percent loss due the vigorish deducted by the track from each betting pool.
I told myself that, after I finished school, I’d test the system in real time at some race track after conducting a similar study of that track’s results. Buy not having much of a stomach for gambling, I’ve never done so.
Instead, I later became interested in the stock market, and I’ve found the takeaway principal from my fantasy race track betting highly useful here: it matters what you pay for the stocks you buy.
As at the track, stock favorites are unlikely to pay off in the long run. The fact that certain stocks are “favorites” usually means other people are overpaying for them.
Of course, the stock market differs from the race track in that there’s a positive expected return from owning stocks that approximates the growth in corporate earnings—roughly about 7 percent a year since the end of World War II—plus the current dividend. Thus, the index investor who takes positions in all stocks can expect about a 9 percent positive return over the long run, compared to the negative 15 percent return from betting on every horse in every race at the track.
Why should it matter what one pays for a stock? Because, like bettors at the track, every investor has a different opinion of what information is important. And to a greater or lesser degree, every market participant allows a variety of human follies to influence his or her investment decisions. Some of these follies derive from how humans are hardwired to think and behave. Others are cultural, so not all markets can be expected to follow precisely the same pattern of “inefficiency” or, more correctly, “irrationality.”
To invest successfully, it’s important to study each market in which you intend to participate, so you can try to learn what strategies are likely to improve your returns.
