CHAPTER 1
The Paradox of Dumb Money
âAs they say in poker, âIf you've been in the game 30 minutes and you don't know who the patsy is, you're the patsy.'â
âWarren Buffett (1987)
In the summer of 1968, Ed Thorp, a young math professor at the University of California, Irvine (UCI), and author of Beat the Market: A Scientific Stock Market System (1967), accepted an invitation to spend the afternoon playing bridge with Warren Buffett, the not-yet-famous âvalueâ investor. Ralph Waldo Gerard hosted the game. Gerard was an early investor in Buffett's first venture, Buffett Partners, and the dean of the Graduate School at UCI, where Thorp taught. Buffett was liquidating the partnership, and Gerard needed a new manager for his share of the proceeds. Gerard wanted Buffett's opinion on the young professor and the unusual âquantitativeâ investment strategy for which he was quietly earning a reputation among the members of the UCI community.
Gerard had invested with Buffett at the recommendation of a relative of Gerard's who had taught Buffett at Columbia University: the great value investment philosopher, Benjamin Graham. Graham had first published the value investor's bible, Security Analysis, along with David Dodd, in 1934.1 He was considered the âDean of Wall Street,â and regarded Buffett as his star pupil. Graham's assessment would prove to be prescient.
By the time Thorp met Buffett in 1968, Buffett had established an exceptional investment record. He had started Buffett Partners 12 years earlier, in 1956, at the tender age of 26, with initial capital of just $100,100. (Buffett joked that the $100 was his contribution.) By 1968, Buffett Partners controlled $100 million in capital, and Buffett's share of that was $25 million.2 For the 12 years between 1956 and 1968, Buffett had compounded the partnership's capital at 30 percent per year before his fees, which were 25 percent of the gain over 6 percent per year. Investors like Gerard had compounded at an average of 24 percent a year. Before taxes, each original dollar invested in Buffett's partnership had grown to more than $13. Each of Buffett's own dollars, growing at the greater prefee annual rate of 30 percent became before taxes over $23. By 1968, however, Buffett was having difficulty finding sufficiently undervalued securities for the partnership, and so had decided to wind it up. This had led Gerard to find a new manager, and Gerard hoped Thorp was the man. He wanted to know if Thorp's unusual quantitative strategy worked, and so, at Gerard's behest, Thorp found himself sitting down for a game of bridge with Buffett.
Buffett is a near world-class bridge player. Sharon Osberg, international bridge player and regular professional partner to Buffett, says, âHe can play with anyone. It's because of his logic, his ability to solve problems and his concentration.â3 Says Buffett, âI spend 12 hours a weekâa little over 10 percent of my waking hoursâplaying the game. Now I am trying to figure out how to get by on less sleep in order to fit in a few more hands.â4 Buffett presented a daunting opponent. Thorp observed of Buffett's bridge playing5:
Thorp was no stranger to the card table either. Before he figured out how to beat the market, Thorp wrote Beat the Dealer, the definitive book on blackjack card counting. William Poundstone recounts the story of Thorp's foray into card counting in his book, Fortune's Formula.6 In 1958, Thorp had read an article by mathematician Roger Baldwin, who had used U.S. Army âcomputersââwhich actually meant âadding machinesâ or the people who operated themâto calculate the odds of various blackjack strategies in an effort to find an optimal strategy. Over three years, he and three associates found that by using an unusual strategy they could reduce the house edge in blackjack to 0.62 percent. Amazingly, prior to their paper, nobody, including the casinos, knew the real advantage held by the house. There were simply too many permutations in a card deck of 52 to calculate the casino's edge. âGoodâ players of blackjack, other writers had claimed, could get the house's edge down to 2 or 3 percent. Baldwin's strategy, by reducing the house edge to 0.62 percent, was a huge leap forward. The only problem, as far as Thorp could see, was that Baldwin's strategy still lost money. He was convinced he could do better.
Thorp's key insight was that at the time blackjack was played using only one deck and it was not shuffled between hands. In the parlance of the statistician, this meant that blackjack hands were not âindependentâ of each other. Information gleaned in earlier hands could be applied in subsequent hands. For example, in blackjack, aces are good for the player. If the dealer deals a hand with three aces, the player knows that only one ace remains in the deck. This information would lead the player to view the deck as being less favorable, and the player could adjust his or her betting accordingly. Thorp used MIT's mainframe computer to examine the implications of his observation and found something completely counte...
