A high school math and statistics teacher from Henderson, Nevada, spent nearly a decade taking tens of millions of dollars from Las Vegas casinos, not by cheating or hacking, but by exploiting a physical flaw in the roulette wheels themselves. Over approximately nine years, the operation he built extracted between $55 million and $62 million from some of the most powerful gambling houses on Earth, all without ever breaking a law. The man at the center of the story was not a professional gambler or a career criminal. He was a quiet, methodical teacher who had no real interest in gambling until an offhand comment from a student in the spring of 1979 sparked an idea.
A freshman complained about losing money at a rigged carnival wheel at the school fair. Instead of dismissing the complaint, the teacher told the class that a wheel doesn’t have to be deliberately rigged to be unfair. It just has to be physically imperfect. All machines are.
That observation led him to a concept known as wheel bias. Roulette wheels are precision instruments, but over thousands of spins, the metal dividers called frets wear unevenly, the axle develops micro variations in tilt, and the ball track develops wear patterns. None of these flaws are visible to the naked eye, but they can be detected through mathematics. The question wasn’t whether the bias existed.
It was whether anyone had the discipline to collect enough data to exploit it. The math teacher began visiting downtown Fremont Street casinos in 1979, not to gamble, but to observe. He would sit near roulette tables with a notebook and record outcomes, hundreds and then thousands of them. He did this for over a year before betting a single dollar.
He understood that 200 or 300 spins was statistical noise. He needed at least 8,000 recorded spins from a specific table to isolate genuine bias from random variance, and ideally far more. By the spring of 1981, he had recruited two assistants: a former student studying computer science and his own brother-in-law, a delivery truck driver who looked completely unremarkable in any setting. The rules were strict: no drinking, no unnecessary conversation, rotate positions every 90 minutes, and never bet on a wheel without at least 8,000 prior recorded spins.
They tested the system at $2 minimum tables for nearly two full years before making serious money. In 1982, they found their first genuinely profitable wheel at a mid-Strip property. A cluster of seven numbers was appearing roughly 11 percent more often than pure random distribution would predict. On a game where the house normally holds a 5.
26 percent edge, that 11 percent advantage reversed the math. The house was now playing against the players. The first year of serious betting yielded about $50,000 in profit, shared three ways. The second year brought $200,000.
By the third year, the operation had expanded to four additional wheels across three properties, and the sample size had grown to more than 700,000 recorded spins. As the money grew, the teacher became more careful. Team members tipped dealers generously, dressed neither richly nor poorly, drove ordinary cars, and rotated their appearances so no casino employee would see the same face too often. By 1985, the operation had grown to 11 people divided into three functions.
Scouts collected data and never bet. Bettors arrived at a wheel only after the bias had been confirmed. A single data analyst, a woman with a background in actuarial mathematics, never entered a casino at all. She worked from the notebooks, ran the numbers, and determined which wheels were still viable and which had been serviced or replaced.
Her accuracy was critical because bias does not last. Wheels get repaired, and frets get replaced. The money compounded. By 1986, the team cleared $1.
2 million in a single year from roulette alone. At its peak in 1988, the operation cleared $11 million across 14 different Nevada properties. The running total at that point was somewhere between $30 million and $40 million over seven years. The sheer volume of money eventually created its own problems.
One team member began skimming cash from the shared pool before it hit the collective ledger. The math teacher caught the thief using the same method he had used to catch a biased roulette wheel: pattern recognition and statistical variance. The betting logs didn’t match the winnings reports, and the variance was too consistent to be luck. Around the same time, the casinos began to notice unusual losses.
The Nevada Gaming Control Board received quiet private complaints from casino operators around 1988 and 1989 about statistically anomalous results at specific tables. The corporate surveillance era had arrived in Las Vegas, and cameras that had once focused on dealers and chip trays were now being pointed at the wheels themselves. MGM Grand and Caesars Palace independently began analyzing roulette betting patterns with early computerized tracking systems, and they found the tell. The tell was simple.
Normal roulette players, even experienced ones, scatter their bets across the table. The math teacher’s team bet on specific number clusters, consistently, session after session, because the bias was in a specific section of the wheel. Once you knew to look for it, the pattern was unmistakable. In 1989, gaming agents approached the math teacher at a Reno casino.
He was polite and completely cooperative. He sat down with them and explained in detail the theory of wheel bias, the mathematical framework, sample sizes, variance, standard deviation, and expected value. He committed no crime. Exploiting wheel bias is not cheating.
It requires no devices, no accomplices inside the casino, and no manipulation of the equipment. It is pure observation. The agents knew it, and the math teacher knew they knew it. He walked out without even a warning.
But knowledge and tolerance are different things. The casinos were done tolerating. The end came on a Friday night in October 1991. The math teacher arrived at a Las Vegas Strip casino, one of the operation’s most consistently profitable properties.
Within 20 minutes, four men approached him from four different directions: casino security and a man in a gray suit who introduced himself only as a representative of the corporate ownership group. They escorted him to a back office where a manila folder sat on a table. Inside were hundreds of surveillance photographs of his face and his team members’ faces at dozens of properties, with dates, times, table numbers, session durations, and bet amounts. They had been watching for 18 months and had built a complete case file.
The man in the gray suit did not raise his voice. He explained, with complete professionalism, that the math teacher was no longer welcome on the property. His photograph and the photographs of every identified team member were being distributed that evening to every major casino operator in Nevada through a private industry security network. No law had been broken, so no one was involving law enforcement.
He was trespassed from the property, and within 72 hours, his photograph was distributed to every major casino on the Strip, Fremont Street, Reno, and Lake Tahoe. He was never charged with a crime and never would be, but he was finished in Nevada. He walked out of that back office, drove home to Henderson, and graded his students’ algebra tests that same night. According to someone close to the family, he seemed completely calm, like a man who had simply finished a project.
The final accounting, pieced together from gaming commission records, court filings in unrelated civil matters, and accounts from people connected to the operation, puts the total somewhere between $55 million and $62 million over approximately nine years. The math teacher did not live large with the money. It was invested quietly in real estate in Nevada and Arizona, modest municipal bonds, and a college fund for his children. He had built a $60 million operation on the principle of not being noticed, and he had no interest in being noticed once it was over.
Some team members scattered. The former student eventually moved to London and attempted a similar operation at British casinos with limited success before retiring. The brother-in-law moved to Arizona and opened a landscaping business. The skimmer moved to Atlantic City and had an undistinguished career in the gaming world.
The data analyst took her share and disappeared from every available record. The math teacher went back to his classroom and taught for seven more years before retiring in the late 1990s. Former students from those final years describe him exactly as students from the years before the operation did: patient, methodical, and genuinely excited about variance. The casino industry responded swiftly.
Within three years of 1991, every major Nevada casino had implemented mandatory wheel inspection protocols, regular fret replacement schedules, automatic table rotation to prevent bias exploitation, and computerized outcome tracking on every roulette wheel. The Nevada Gaming Commission issued formal guidance on wheel maintenance standards that became the industry template across the country. Wheel bias play is still legal in the United States today. A player can walk into a casino and attempt it.
But modern casinos swap wheels constantly, level tables with laser precision, inspect frets on regular maintenance cycles, and run computerized outcome analysis on every result in real time. Finding a genuinely exploitable bias in a modern casino is, for all practical purposes, nearly impossible. The precautions that make it impossible exist because a math teacher from Henderson spent nine years proving they were necessary. The house always wins because the house owns the math, unless the house gets complacent and starts trusting its own mythology more than its own machinery.
The math teacher took the house’s greatest weapon and turned it around. He walked into the house’s church, sat quietly for nine years, and walked out with $60 million, not because he cheated, but because he paid attention. Then he went home and graded some tests.