Jagger, Garcia-Pelayo, and the Ritz prosecution that closed the loophole forever.
Biased wheels in history
Annabel Cavendish
Editor · 14 May 2026
Jagger and the Hypothesis
The beginning of the story is a Yorkshire textile businessman called Joseph Hobson Jagger, who around 1880 or 1881 took his borrowed money to the Casino de Monte-Carlo with a hypothesis: no mechanical roulette wheel can be perfectly balanced, and a sufficiently imperfect wheel would show predictable clusters of winning numbers over enough observations.
He spent roughly a month recording outcomes before placing a single bet. According to Wikipedia's Joseph Jagger article, which draws on the 2018 biography by Anne Fletcher and The Times, he identified nine numbers, 7, 8, 9, 17, 18, 19, 22, 28, and 29, that appeared with above-expected frequency. He bet them. He won over two million francs, approximately £80,000 at the time, equivalent to something in the region of £7.5 million today.
Popular accounts date this to 1873. The Fletcher biography and the Wikipedia article put the dates at around 1880 and 1881. The 1873 date appears to originate from secondary popular accounts without a primary source citation. The history of roulette is full of received wisdom that doesn't survive contact with the primary documents.
The casino's response: they rotated the frets, the metal dividers between number pockets, each night. The bias shifted to different numbers. Jagger found his wheel by a scratch on its surface; when the casino moved the wheels he tracked it down again. But the shifting bias eventually outpaced him and he withdrew with his winnings. Rotating frets rather than the spindle implies the defect was in pocket geometry rather than the central mechanism. The casino knew what it was fixing.
Hibbs and Walford: The Trigger for Modern Engineering
The American version of this story, and the one that directly triggered modern wheel engineering, involves two University of Chicago students in 1947. Albert Hibbs, studying mathematics, and Roy Walford, studying medicine, drove to Reno, spent roughly a week recording spin outcomes, identified the number 9 as appearing with statistically significant frequency on one wheel, and bet it almost exclusively. According to Roulette17's account of the story, they turned approximately $300 into $8,300.
The Milwaukee Journal and the St. Petersburg Times covered it at the time. A subsequent Life Magazine article in 1949 caused casino operators to take wheel quality seriously in a way they never had before. Jagger was the first documented case; Hibbs and Walford were the catalyst that made the industry respond. Wikipedia's Albert Hibbs entry confirms the key details.
Garcia-Pelayo and the Spanish Supreme Court
The most extraordinary and legally significant story belongs to a Madrid music producer named Gonzalo Garcia-Pelayo.
In 1991, Garcia-Pelayo noticed that no mechanical roulette wheel can be manufactured to perfect tolerances, and that the Gran Casino Madrid's wheels, if you recorded enough spins, would reveal systematic biases. He and his family, including his son Ivan and daughter Vanessa, spent months recording tens of thousands of spins by hand, then entered the data into a home computer and identified numbers appearing with above-expected frequency.
By summer 1992, less than a year after starting, the family had earned approximately 70 million pesetas. Their total across Spain and several European cities reached approximately 250 million pesetas, around £1.5 million equivalent, before the casinos identified the pattern and barred the family globally. As El Pais reported in their 2012 profile, individual cities contributed: 14 million pesetas in Vienna, 13 million in Amsterdam, 40 million in Lloret de Mar.
The Casino de Madrid sued for fraud. The Spanish Supreme Court ruled in the family's favour. As CasinoEdges.net confirms, the court found that Garcia-Pelayo had used ingenuity and computer analysis of publicly available information, namely the spin results recorded in plain sight at a public table. No device. No deception. No rule violation. The casino bore the burden of maintaining fair equipment. Exploiting a mechanical defect through patient statistical analysis was not fraud. It was intelligence. The ruling established that casinos cannot blame players for their own engineering failures.
After the ruling, Garcia-Pelayo was banned from casinos globally and his method became worthless: every serious casino adopted the electronic outcome-tracking systems he had done manually with a notepad.
Niko Tosa and the Ritz, 2004
The story that most directly shaped the current legal landscape involves a player known publicly as Niko Tosa and two colleagues, who played roulette at the Ritz Casino in London over two sessions in March 2004 and left with approximately £1.3 million. Scotland Yard arrested them at a nearby hotel that night. Nine months of investigation followed. The conclusion, confirmed by a Scotland Yard spokesman quoted in ABC News Australia, 7 December 2004: "All three persons that were arrested have been informed that no further action will be taken. All the money detained by the police has been returned."
The press reported a laser scanner concealed in a mobile phone, calculating the ball's decaying orbit. Tosa's own account, given in a Bloomberg interview in 2023, denied any technology and claimed the wins came from visual bias tracking. Which version is true is genuinely disputed. The detailed analysis at OnlineRouletteSites.org.uk covers both versions. What isn't disputed: the Ritz replaced its wheels after the incident, which is the casino's own acknowledgement that something was exploitable.
The case prompted the UK government to commission a review confirming that basic roulette computers were effective, and directly influenced the cheating provisions in the Gambling Act 2005. Section 42 of the 2005 Act broadened the cheating offence to cover any acts likely to affect the random element of a game. The window that existed in 2004 is closed.
Modern Countermeasures: Catching the Bias Before You Do
The era of the exploitable wheel at a serious venue closed when the Garcia-Pelayo case forced every major operator to implement electronic monitoring. The sensor-equipped wheels finished the job in the 2000s.
The TCS John Huxley Saturn contains Drop Zone Detection: sensors that flag deviation at the ball-drop stage, before any biased pocket affects visible results. The Cammegh Mercury 360 uses four in-rim sensors and an open data protocol for real-time bias detection. The Cammegh 2025 brochure recommends weekly spirit-level checks. A bias-tracking patent registered as US20190012877A1 describes an image-processing system that logs drop zones continuously and detects deviation before the winning-number distribution shows anything unusual. The casino finds the bias before you do.
What It Would Take Today
Even setting aside the detection systems, the sample size required to confirm wheel bias at statistical significance makes the exercise impractical at any modern venue.
It would take a minimum of approximately 7,500 to 10,000 spins on the same identified wheel to confirm bias at the level required for confident betting. At 60 to 80 spins per hour at a busy table, that's 125 to 167 hours of table time, on the same wheel, without that wheel being rotated, removed, or repaired. Drop Zone Detection systems flag anomalies at a smaller sample than 10,000. You would not accumulate actionable data before the casino's monitoring detected the same anomaly.
Key numbers
Case
Method
Era
Outcome
Joseph Jagger, Monte Carlo
Manual recording, pocket bias
c.1880-81
~£80,000 won (~£7.5m today)
Hibbs and Walford, Reno
Manual recording, pocket bias
1947
~$8,300 from $300
Garcia-Pelayo family, Madrid
Computer analysis of spin records
1991-1995
~250 million pesetas (~£1.5m equivalent)
Niko Tosa, Ritz London
Laser device or visual tracking
2004
~£1.3 million kept
Detection requirement
Value
Minimum spins to confirm bias (sector level)
~400
Minimum spins for statistical confidence (single pocket)
7,500-10,000+
Hours at 80 spins/hr to reach 10,000 spins
~125 hours
TCS Saturn: bias flagged before player can confirm
Welcome to what is, if I'm honest, one of my favourite lessons in this series.
I'm Annabel, and today we are talking about biased wheels: the history of people who found them, the fortunes that changed hands, and why the era of the exploitable wheel is definitively over at any casino running modern hardware.
The beginning of the story is a Yorkshire textile businessman called Joseph Hobson Jagger, who in approximately 1880 or 1881 took his borrowed money to the Casino de Monte-Carlo with a hypothesis: that no mechanical roulette wheel can be perfectly balanced, and that a sufficiently imperfect wheel would show predictable clusters of winning numbers over enough observations.
He spent roughly a month recording outcomes before he placed a single bet.
He identified nine numbers, seven, eight, nine, seventeen, eighteen, nineteen, twenty-two, twenty-eight, and twenty-nine, that appeared with above-expected frequency.
He bet them.
He won over two million francs, which was approximately eighty thousand pounds at the time and is equivalent to something in the region of seven and a half million pounds today.
Popular accounts date this to 1873.
The 2018 biography by Anne Fletcher, which is the most scholarly treatment of the story, and the Wikipedia article drawing on it, puts the dates at around 1880 and 1881.
The 1873 date appears to originate from secondary popular accounts without a primary source citation.
I mention this because the history of roulette is full of received wisdom that doesn't survive contact with the primary documents.
The casino's response to Jagger is instructive.
They rotated the frets, the metal dividers between number pockets, from one set of positions to another each night.
The bias shifted to different numbers.
Jagger, who had been marking his wheel with a scratch to identify it and tracking its specific tendencies, could no longer keep up with the moving target, and he withdrew with his winnings.
The countermeasure was to rotate the frets, not the spindle.
This matters: if the bias had been in the spindle, rotating the frets would have been useless.
The casino's choice of countermeasure strongly implies the defect was in the pocket geometry, the relative height or weight distribution of the frets, rather than in the wheel's central mechanism.
They knew what they were fixing, even if they didn't announce it.
The American version of this story, and the one that directly triggered modern wheel engineering, involves two University of Chicago students in 1947.
Albert Hibbs, studying mathematics, and Roy Walford, studying medicine, drove to Reno, spent roughly a week recording spin outcomes, identified the number nine as appearing with statistically significant frequency on one wheel, and bet it almost exclusively.
They turned approximately three hundred dollars into eight thousand three hundred dollars.
The Milwaukee Journal and the St.
Petersburg Times covered it at the time.
A follow-up Life Magazine article in 1949 caused casino operators to take wheel quality seriously in a way they never had before.
Jagger was the first documented case; Hibbs and Walford were the catalyst that made the industry respond.
The most extraordinary and legally significant story belongs to a Madrid music producer named Gonzalo Garcia-Pelayo, who in 1991 noticed that no mechanical roulette wheel can be manufactured to perfect tolerances, and that the Casino de Madrid's wheels, if you recorded enough spins, would reveal systematic biases.
He and his family, including his son Ivan and daughter Vanessa, spent months recording tens of thousands of spins by hand, entered the data into a home computer, and identified the numbers that appeared with above-expected frequency.
By the summer of 1992, less than a year after they started, the family had earned approximately seventy million pesetas, equivalent to around four hundred and twenty thousand euros.
Their total take across Spain and several European cities reached approximately two hundred and fifty million pesetas, around one point five million euros, before the casinos identified the pattern and barred the family globally.
The Casino de Madrid sued for fraud.
The Spanish Supreme Court ruled in the family's favour in 1994.
The judges found that Garcia-Pelayo had used ingenuity and computer analysis of publicly available information, namely the spin results that had been recorded in plain sight at a public table.
No device.
No deception.
No rule violation.
The casino, the court found, bore the burden of maintaining fair equipment.
Exploiting a mechanical defect through patient statistical analysis was not fraud.
It was intelligence.
The ruling is cited in European gambling law and established that casinos cannot blame players for their own engineering failures.
Then there is the Ritz in 2004, and a man known publicly as Niko Tosa.
In March of that year, Tosa and two colleagues played roulette at the Ritz in London over two sessions and left with approximately one point three million pounds.
Scotland Yard arrested them at a nearby hotel the following night.
Nine months of investigation followed.
The conclusion: no further action would be taken and all money detained by police was returned.
Scotland Yard confirmed this publicly in December 2004.
The press reported a laser scanner concealed in a mobile phone, calculating the ball's decaying orbit.
Tosa's own account, given in a Bloomberg interview in 2023, denied any technology and claimed the wins came from visual bias tracking and last-second sector betting across three players.
Which version is true is genuinely disputed.
What isn't disputed is that the Ritz replaced its wheels after the incident, which is the casino's own acknowledgement that something was exploitable.
The case prompted the UK government to commission a review confirming that basic roulette computers were effective, and directly influenced the cheating provisions in the Gambling Act 2005.
Modern wheels are a different matter entirely.
The TCS Saturn's inclinometer fires an error at the same zero point two degree tilt threshold that academic research identifies as the exploitation boundary.
Cammegh runs weekly spirit-level checks and recommends them in their published brochure.
A bias-tracking patent, registered as US20190012877A1, describes an image-processing system that logs ball drop zones in real time and detects statistical deviation before the winning-number distribution shows anything unusual.
The casino finds the bias before you do.
The era of the exploitable wheel at a serious venue closed when the Garcia-Pelayo case forced every major operator to implement electronic monitoring.
The sensor-equipped wheels finished the job in the two thousands.
What remains is the history, which is genuinely extraordinary, and the mathematics, which hasn't changed at all.
