Edward Thorp and Beat the Dealer

Beat the Dealer (1962) is the book that turned blackjack from a casino game into a solved problem. Edward O. Thorp, then a young mathematics instructor at MIT, showed that the house edge in single-deck blackjack is not constant: it swings with the cards already dealt, and a player who tracks that swing and bets accordingly holds an advantage. His own site describes it as “the first book to mathematically prove that the house advantage in blackjack could be overcome by card-counting” (edwardothorp.com). Every counting system since, and much of the modern casino rulebook, descends from it.

The man and the machine

Thorp was born on 14 August 1932 in Chicago and raised in southern California. He took a PhD in mathematics at UCLA in 1958 and moved to MIT in 1959, where he had access to one of the most powerful computers of the day, the IBM 704. The Wikipedia biography records that he taught himself Fortran and used the 704 “to study winning probabilities and to devise card-counting schemes”, particularly for the end of a deck that is not reshuffled after every hand.

His starting point was the 1956 paper by Baldwin, Cantey, Maisel and McDermott, the Four Horsemen of Aberdeen, who had computed basic strategy on desk calculators. Thorp re-ran their work by computer, corrected the small errors forced by hand calculation, and then asked the question they had only touched on: what happens to the player’s expectation if specific cards are removed from the deck?

The answer was the discovery. Remove the fives and the player gains; remove the tens and the player loses. Because a single-deck game in 1960 was dealt deep without a shuffle, the deck’s composition drifted hand by hand, and the drift was observable by anyone watching the cards.

The Ten Count

The book’s system, the Ten Count, tracked the number of ten-valued cards and non-tens remaining and compared their ratio to the full-deck ratio of 16 tens to 36 others. When the deck became relatively rich in tens, the player bet more and took insurance; when poor, the player bet the minimum. Thorp sized bets with the Kelly criterion, which he had learned from a 1956 paper by John L. Kelly.

The Ten Count is correct but awkward: a player must track two totals and mentally divide. The Wikipedia article on the book notes that it is “harder to use and less profitable than later systems”. The point-count systems that replaced it assign +1, 0 or −1 to each rank, so the player tracks one running total; see our card counting guide for the Hi-Lo version. The second edition of Beat the Dealer in 1966 was “substantially updated” with computer results from that intervening work.

The bankroll and “Mr. X”

A mathematical proof was not going to persuade the public. Thorp needed a live test, and that required money. Once word of his results spread, he was approached by Emmanuel “Manny” Kimmel, whom the Wikipedia biography describes as “a wealthy professional gambler and former bookmaker”. Kimmel was a founder of the Kinney Parking Company, the business that later grew into Warner Communications; the Wikipedia article on Kimmel also reports FBI interest in his dealings with Newark racketeer Abner Zwillman and other organised-crime figures. Those connections are reported, not proved in a court record, and Thorp was not accused of any part in them.

Kimmel put up the venture capital, $10,000 according to Wikipedia, and Thorp set off for Reno, Lake Tahoe and Las Vegas. The biography records that “the team reportedly won $11,000 in a single weekend”. In the book Thorp disguised his backer as “Mr. X”. The Las Vegas Advisor piece on the first counters adds a complication for anyone reading the sources: a separate January 1962 trip, documented by Mickey MacDougall and Russell Barnhart, was bankrolled by William F. Rickenbacker, whom that account also calls “Mr. X”. The Kimmel trip is the one described in Beat the Dealer; the two should not be confused.

That 1962 trip is also a reminder of the conditions of the time. MacDougall, a former Nevada Gaming Control Board investigator, recorded dealers cheating against Thorp at several casinos, and Thorp called one technique, the high-low pick-up, “the most widespread single method of dealer cheating” in Nevada during the trip. The net result of that trip, after heavy losses in Reno, was a profit of $317 before expenses. The famous $11,000 weekend and the $317 trip are different events with different backers.

Publication and reaction

Beat the Dealer was published in 1962 and, per the Wikipedia biography, sold more than 700,000 copies and reached the New York Times bestseller list. Readers who had never thought of blackjack as beatable now had a strategy table, a counting method and a betting rule.

What the casinos did next is contested in the sources, and it is worth being precise:

Thorp’s own later editions describe a short-lived rule tightening that was reversed because it drove away ordinary players, and that is the account most books repeat. We could not verify the exact 1964 date and rule list against a primary source, so treat the specifics as commonly reported rather than established. What is not in doubt is the long-run outcome: more decks, shallower penetration, restrictions on doubling and splitting, and eventually continuous shuffling machines. Each change is covered in how casino blackjack rules evolved.

Casinos also simply refused service. Thorp was barred and took to disguises, including wraparound glasses and false beards, to keep playing. He also ran a winning baccarat team.

The wearable computer

Before the book came out, Thorp and Claude Shannon built what is generally called the first wearable computer, between 1960 and 1961, to predict roulette outcomes. One operator timed the wheel with a toe switch; the other heard musical tones through a hidden earpiece. The device was tested in June 1961 in Shannon’s basement laboratory. Nevada’s law against such devices, the biography notes, took effect on 30 May 1985.

After blackjack

Thorp left MIT for New Mexico State University in 1961 and moved to the University of California, Irvine in 1965, where he was professor of mathematics and later of mathematics and finance until 1982. With Sheen T. Kassouf he wrote Beat the Market (1967), applying the same expectation-and-bankroll thinking to warrants and convertible bonds. His hedge fund, Princeton/Newport Partners, ran from 1969 to 1989 using market-neutral derivatives hedging; a second fund, Ridgeline Partners, ran from 1994 to 2002. In 1991 he was an early sceptic of Bernard Madoff’s reported returns, seventeen years before the fraud collapsed. His autobiography, A Man for All Markets, appeared in 2017. He was an inaugural member of the Blackjack Hall of Fame in 2002.

Why the book matters for basic strategy and counting

Beat the Dealer did two separate things, and players still conflate them.

First, it confirmed and popularised basic strategy. The 1956 strategy existed, but its print run was 5,000 copies. Thorp’s table reached hundreds of thousands of readers and established that a player who simply follows the chart faces a small, calculable house edge rather than the several percent that guesswork costs. That is the part of the book that applies to every blackjack game, including six-deck shoes dealt today.

Second, it introduced counting. That part applies only when the cards you see are not returned to the pack before the next hand. Under the six-deck, dealer-stands-on-soft-17 profile our strategy chart assumes, basic strategy alone leaves a house edge of roughly 0.4%, and a count can shift that only where penetration and betting conditions allow. Thorp’s book proved a possibility, not a guarantee, and he said as much: the advantage is small, the variance is large, and casinos are free to change the conditions. They did.

Related reading

• The Four Horsemen of Aberdeen
• Card counting
• How casino blackjack rules evolved
• Famous blackjack players