Finding an Edge You Can Name

In 1967 a mathematics professor at the University of California, Irvine published a book with a title that sounded like every trading pitch before and since. Beat the Market: A Scientific Stock Market System, written by Edward O. Thorp with the economist Sheen Kassouf, was not a pitch. Thorp had already shown, in Beat the Dealer (1962), that blackjack, a game casinos believed fair, could be beaten by counting cards and sizing bets carefully. The new book made a narrower claim about a narrower market. Warrants, long-dated call options that companies issued on their own shares, traded on exchanges and over the counter with no accepted way of pricing them, and many were priced far above any reasonable value. Thorp and Kassouf's system was to sell the overpriced warrant and buy the common stock in a ratio that left the position nearly indifferent to small moves in the share price, then wait for the overpricing to decay, adjusting the ratio as the stock moved. It was a delta hedge, published six years before Fischer Black and Myron Scholes published their option-pricing formula (companion account).

Thorp traded the system as well as publishing it. He began managing money for individuals in 1967 in a small partnership that took 20% of profits, and in November 1969 he and James "Jay" Regan, an investor who had read the book and approached him, started Convertible Hedge Associates with $1.4 million. Thorp's team in Newport Beach, California built the models; Regan's office in Princeton, New Jersey raised money, traded and ran the business. Renamed Princeton/Newport Partners in 1974, the fund hedged warrants, convertible bonds and, after the Chicago Board Options Exchange opened on 26 April 1973, listed options against their stock. Assets reached $273 million at the end of 1987. The record most often quoted is about 19.1% a year gross and 15.1% net of fees from November 1969 to December 1988, with three losing months (secondary sources citing Thorp); other accounts give 18.2% a year to limited partners after fees or about 20% after fees. None is audited, and the companion keeps all three with their bases.

What makes the case the right place to begin this part of the book is not the return. It is that Thorp could say exactly where the return came from. Speculators bought warrants for leverage and overpaid for them; few professionals had a model that showed by how much; a hedger who did could collect the overpayment while carrying little market risk. He could also say what would end it, and it did end. Once Black–Scholes was published in 1973, the Cboe standardised options and traders programmed the formula into calculators, the mispricings that had paid Princeton/Newport shrank, and convertible arbitrage became a crowded hedge-fund strategy. Thorp moved on. A 1979–80 research project found a short-term reversal effect in stock returns, and in 1985 the firm began statistical arbitrage with Gerry Bamberger, formerly of Morgan Stanley, earning 25–30% a year on $30–60 million before returns fell to about 15% by 1987 (Thorp's own figures). The fund was destroyed in the end by something outside any model: federal agents raided the Princeton office on 17 December 1987 over tax trades and stock "parking" tied to Drexel Burnham Lambert, five Princeton-office executives were indicted in August 1988, the firm ceased operations that year, and most of the key convictions were reversed on appeal in 1991. Thorp was never charged.