Feature The Gang That Beat Las Vegas Part VI
by Ian Thomsen, courtesy of The National Sports Daily
14 June, 2004
Part VI
Billy Walters claims that he gambled more than $500,000 of his own money each week with the help of computer information. He is just one of many big winners whose profits do not appear on the group's ledgers.
Even though he tried to gamble like the button-down brokers on Wall Street, Walters admits that he too fell victim to the occasional betting frenzy. During the Christmas holidays six years ago, Walters found he was betting hand over fist on Michigan in the Sugar Bowl against Auburn. It was one of those rare times when the tout services were opposing the computer on a major game. No matter how much Billy Walters bet on Michigan for the Computer Group and for himself, the line remained the same. The public kept laying money on Auburn, giving 4 ½ points.
"I kept betting on the game, and the line kept coming back, so I just kept betting,' Walters says. "I guess I got a little carried away. I had more than $1 million on that game. I literally bet my entire net worth on that game, and probably some additional."
Trailing 7-6 in the fourth quarter, Auburn took possession at its 39 with 7:44 left. If Auburn scored a touchdown to cover, Billy Walters would lose $1 million. But Auburn kicked a 19-yard field goal to win, 9-7, and Billy Walters is today a rich genius.
The Computer Wizard
One day Michael Kent, who was the centerfielder, got to wondering about his company softball team. How good were he and his teammates, really? When they destroyed a poor opponent by 15-4, was that as impressive as beating a good team by 6-5? His team had won a couple of league championships, but what had they really accomplished? All his life he had found answers to such questions in numbers, statistics. He simply had to find out what those numbers meant. What was the numerical definition of a good softball team?
His thoughts drifted naturally in this direction. Kent was a 27 year-old mathematician at Westinghouse in suburban Pittsburgh. Every day he worked with computers to help design a better nuclear sub. At night, he says, he began to formulate a computer program that rated the strength of his softball team. Each week he would update the statistics then feed the information into the high-speed Control Data computer at Westinghouse. His teammates were interested in this output of statistics - it was flattering to them - but Michael Kent ultimately was disappointed by the results. When the work was done, he had a printout listing his team's strengths and weaknesses. So what? He had given order to these numbers, but there was no application, no further use for them.
He says he began work on a more complex program. The game was college football. This time he could foresee a dollar sign in front of the numbers. The year was 1972. He recorded information from old NCAA football guides, which list the scores and statistics from the previous season. Then he visited the library, the old newspapers in particular, in order to see which teams had been favored each week, and by how many points. He examined the spreads and the stats, attempting to find a correlation. Which statistics, he wanted to know, were important in assigning a point spread?
He knew of only one way to find out. He began to write a program. The computer would ask hundreds of questions in algorithmic form, pinpointing strengths and weaknesses for each team. As his wealth of information grew, Kent learned that some strengths were more important than others. There was a value to first downs and there was another value to yards gained. Home-field advantage had a value. So did strength of schedule. So did success against common opponents. The list of questions went on and on, some so picayune that the average football fan might have laughed in the face of this stocky, bespectacled mathematician. Billy Walters believes that the program even accounted for the distance of the visiting team's road trip.
The hobby soon became his vocation. He began to test his model by placing bets with local bookies. He says he worked an average of two hours per night over the course of seven years, fine-tuning his football program and developing a similar program for college basketball, until one morning he walked into the plant and quit his job. He was very quiet about it. Only his closest friends were informed of his plans. He moved to Las Vegas in time for the 1979 college football season. For the last seven years he had been saving his money, to wager on football and basketball games. Still, when he looked in the mirror, it was a hard thing to believe, that the person staring back at him was a professional gambler.
When Michael Kent arrived in Las Vegas, he clearly was on his own. No gambler of note was depending solely upon a computer to analyze bets. Allow yourself to go broke because of machine? That was crazy thinking. But Michael Kent didn't know anything about Las Vegas common sense. He was from Pennsylvania. He wanted to know where he should do his laundry. On a daily basis he wanted to bet as many games as he could, whenever he perceived the slightest 1 ½ point advantage, and this was more crazy thinking. Common sense in Las Vegas said that you couldn't win big by betting a lot of games. You should concentrate on just a few games. That's what common sense said. Michael Kent didn't know about that either. Most of what he knew about his business was contained in a book called Theory of Gambling and Statistical Logic by Richard A. Epstein. Chapter 2 told him the percentage of his money he should bet, depending on how much he liked the game. The book was written in the language of numbers. Michael Kent wanted to meet this man Epstein.
