The Pelayo Family and the Statistical Assault on Roulette

The Pelayo family’s story is one of the most remarkable — and legally vindicated — episodes in the history of advantage play. Unlike card counters who rely on conditional probability, the Pelayos exploited **physical imperfection**: the mechanical bias inherent in every roulette wheel. Their method was pure statistical bookkeeping, performed over hundreds of thousands of spins, and it worked so well that it took a Spanish Supreme Court ruling to secure their right to keep playing. For decades, casinos claimed roulette could not be beaten. Then one Spanish family walked into casinos with notebooks instead of computers, recorded tens of thousands of spins, and won so consistently that the dispute reached Spain's Supreme Court and decided too, in their favor.

---

## I. The Mathematical Foundation: Wheel Bias

Every roulette wheel is a physical machine. Manufacturing tolerances, wear, dust, and even the tilt of the table introduce **bias** — a deviation from the uniform 1-in-37 (European) or 1-in-38 (American) probability that each number should theoretically have.

The Pelayo method began with a simple observation: **not all numbers are equally likely on a given wheel**. Over enough spins, a wheel might show that certain numbers appear 10% or 20% more often than expected. If you can identify those numbers and bet only on them, you flip the house edge.

This is not a new idea. In the 1870s, Joseph Jagger, a Bradford textile worker, hired six clerks to record outcomes at Monte Carlo’s Beaux-Arts Casino. He discovered that one wheel favored nine numbers — 7, 8, 9, 17, 18, 19, 22, 28, and 29 — and won roughly $325,000 before the casino moved the frets and defeated his edge . His story inspired the song “The Man Who Broke the Bank at Monte Carlo” .

But Jagger relied on manual observation. The Pelayos industrialized the process.

---

## II. The Pelayo Method

### Data Collection

Gonzalo García Pelayo, a Spanish record producer and mathematics enthusiast, began recording roulette outcomes in Madrid in the late 1980s. The family — including his sons Iván and Vanessa and other relatives — dedicated themselves full-time, spending **eight hours a day, six days a week** at the casino, recording spin results across multiple wheels .

They collected data on a scale never attempted before. According to Harvard Business School case materials derived from their work, the Pelayo dataset included **70,340 spins across four wheels** — three real (imperfect) and one simulated (perfect) . The challenge posed in the case is to identify which wheel is the simulated one, demonstrating how clearly the biased wheels stand out under statistical analysis.

### Statistical Analysis

The family used the **chi-squared goodness-of-fit test** to determine whether a wheel’s outcome distribution deviated significantly from the theoretical uniform distribution . The chi-squared test compares observed frequencies against expected frequencies. For a wheel with no bias, the test statistic should be small. For a biased wheel, it becomes large.

Once a wheel showed significant bias, the Pelayos identified the **favored numbers** — those that appeared more frequently than expected. They then bet exclusively on those numbers, often placing multiple chips to cover them.

### Betting Strategy

The Pelayos bet heavily. In the early days in Madrid, they wagered up to **$150 per number**. On a good day, they took home **$5,000** . By the summer of 1992, before being banned from the Gran Madrid casino, they had won approximately **70 million pesetas** (roughly $500,000 at the time) .

They operated with strict discipline. Gonzalo instructed his family: “Don’t talk about a system to anybody. We’re just regular players who do this for fun. Leave the casino as soon as your shift is up. Don’t linger. Don’t have dinner or any type of relationship inside the casino. And, of course, only play the roulette you’re assigned and only the number you’ve been given. In other words, be efficient. This is a job” .

---

## III. The Casino Countermeasures

The Pelayos’ success inevitably attracted attention. According to NPR’s account, the family’s cover was blown by a personal indiscretion: Gonzalo’s nephew Heime was flirting with a dealer who happened to be the ex-girlfriend of the casino director. The director noticed Heime winning consistently and began investigating. Soon, the entire family was identified as a coordinated winning group .

The casino’s response was immediate and aggressive:

1. **Moving the wheels**: The family arrived one day to find the roulette tables had been relocated. Gonzalo, who knew each wheel by sight, simply found them and resumed playing .

2. **Changing the wheel axis**: This was more effective. By altering the tilt of the wheel, the casino changed the physical bias that the Pelayos had exploited. Gonzalo’s strategy was “thrown out the window” .

3. **Banning the family**: Ultimately, the Gran Madrid casino denied Gonzalo and his family access altogether .

Gonzalo attempted to fight back. He deliberately got himself physically ejected from the casino, hoping to provoke a legal confrontation. Two large men dragged him to the door and pushed him out . His plan initially failed, but the legal battle would take a decade.

---

## IV. The Legal Victory: 2004 Spanish Supreme Court

The Pelayo family spent years banned from Spanish casinos. They continued playing abroad — in Las Vegas, Australia, Austria, Denmark, and the Netherlands — and are estimated to have won **over 250 million pesetas** in total .

In 2004, the **Spanish Supreme Court ruled in their favor**, restoring their right to enter any casino in Spain . The court’s reasoning was decisive:

> **“It cannot be defined as cheating if the player does not intervene in any way to influence the result of the game.”** 

This is the legal crux. The Pelayos did not tamper with the wheel. They did not bribe croupiers. They did not use hidden devices. They simply **observed** and **recorded** outcomes, then used statistical analysis to identify the wheel’s natural bias. The court recognized this as **skill**, not fraud.

The ruling has been described as a landmark for player rights in Spain. It established that casinos cannot ban a player merely for being **too good** at a legal game .

---

## V. The Roulette Computer: A Different Approach

The Pelayos exploited **static bias** — the wheel’s tendency to favor certain numbers over time. A separate lineage of advantage players pursued a different target: **predicting the ball’s landing in real time** using physics and computation.

### Edward Thorp and Claude Shannon (1961)

In 1955, Edward Thorp conceived of a wearable computer to predict roulette outcomes . By 1960–61, working with Claude Shannon at MIT, he built the first functioning device. It was **cigarette-pack sized**, worn strapped to the waist, with **foot-switches hidden in a shoe** for input and an **earpiece** for output .

The principle was straightforward in concept but difficult in practice. The computer measured the **speed of the ball**, the **speed of the rotor**, and the **rate of decay** as the ball slowed. From these, it calculated the **most likely octant** (one-eighth of the wheel) where the ball would land. Testing in Shannon’s basement yielded an **expected gain of +44%** on bets in the favored octant . When they tested it in Las Vegas in 1961, predictions matched the laboratory expectation, but a hardware problem prevented sustained serious betting .

They kept the device secret until 1966, when Thorp published his methodology . The device is now in MIT’s museum .

### The Eudaemons (1978–1985)

A group of Silicon Valley college students, calling themselves **Eudaemonic Enterprises**, built smaller and cheaper roulette computers using microprocessor breakthroughs. Some were hidden **entirely in the hollowed-out soles of leather Oxford shoes** .

They tested their first computers in Las Vegas in **1978**, but recurring technical problems limited their practical use . Their most important contribution was not the hardware but the documentation. In 1985, member Thomas Bass published **_The Eudaemonic Pie_**, a detailed account of how they built the computer for surprisingly little money .

The book’s publication had a legislative consequence. In **1985**, the **Nevada legislature passed the first law making the use of predictive devices illegal** (Nev. Rev. Stat. § 465.075) . Legal scholars debated whether the law was a direct response to the book, but I. Nelson Rose noted that it was “no coincidence” that the law came at a time of renewed interest in player rights and the book’s publication . The original bill covered only card-counting devices; it was expanded to include all predictive gambling devices .

### Later Development: “Cerebral Clocking”

By the 2000s, some players had internalized the physics to the point where they could predict without a device. Bloomberg’s 2023 investigation into Niko Tosa — a Croatian gambler who won millions at the Ritz in London — found that he claimed to have **never used a computer** .

Tosa trained with a roulette wheel at home, watching the ball until he developed an intuitive sense of where it would land. Doyne Farmer, a physicist from the Eudaemonic group, told Bloomberg that “it’s conceivable that someone could do what we do without a computer, providing the wheel is tilted and the rotor is not moving too fast.” He compared “cerebral clocking” to musical talent, suggesting it activates similar brain regions dedicated to sound and rhythm .

Casinos call this “cerebral clocking,” and it is nearly impossible to detect — there is no device, no signal, no evidence. Tosa was never caught with a computer. He was, however, beaten by casino security more than once .

---

## VI. The Modern Countermeasure: Bias Detection Systems

The Pelayo method — and the Jagger method before it — exploited **wheel bias**. Modern casinos have responded by developing systems to **detect bias in real time** and **correct it before players can exploit it**.

A 2020 patent (US20200211327A1) describes an **image processing unit** that uses a camera to track the roulette ball’s trajectory . It records the position where the ball reaches a specific circle around the deflectors, storing this data in a database. By analyzing the distribution of these drop positions over many spins, the system can **detect deviations from expected statistics** — i.e., bias — and generate an alert. The croupier can then **close the wheel for maintenance** before the bias becomes exploitable .

Another patent (EP3403249A1) describes a similar system that stores drop-zone data and compares it against expected distributions . The goal is **proactive** bias detection: catching the deviation **before** it affects the winning-number statistics that a player like Gonzalo Pelayo would be recording .

This is the modern arms race. Jagger defeated it by hiring clerks. The Pelayos defeated it by dedicating their lives to data collection. But a camera system that monitors every spin in real time and flags bias immediately is a different kind of opponent.

---

## VII. The Scientific and Legal Significance

The Pelayo case is scientifically significant because it demonstrates that **roulette is not purely random**. A wheel is a physical object subject to wear and manufacturing variance. The Pelayos did not cheat; they **measured**. Their method was essentially the same as an engineer testing a machine for performance deviations.

It is legally significant because the Spanish Supreme Court **recognized this distinction**. The court did not say “roulette is random, so you must be cheating.” It said: if the player does not influence the outcome, using statistical observation to bet on a biased wheel is not fraud .

This stands in contrast to the treatment of roulette computers. In Nevada, **using a predictive device is explicitly illegal** . The distinction is subtle: measuring a wheel’s historical bias (Pelayos) is legal; measuring a ball’s trajectory in real time with a device (Thorp/Eudaemons) is not. The first is “observation”; the second is “instrumentation.”

---

## VIII. Summary

| Method | Practitioner | Principle | Legal Status |

|--------|-------------|-----------|--------------|

| **Manual bias observation** | Joseph Jagger (1870s) | Record outcomes, find favored numbers | Legal; casino countered by moving frets |

| **Statistical bias analysis** | Pelayo family (1990s) | Chi-squared test on 70,000+ spins; bet only favored numbers | **Legal**; Spanish Supreme Court 2004 ruling |

| **Wearable predictive computer** | Thorp & Shannon (1961) | Measure ball/rotor speed; predict octant | **Illegal** in Nevada (1985 law) |

| **Microprocessor in shoe** | Eudaemons (1978) | Same principle, smaller hardware | **Illegal** in Nevada |

| **Cerebral clocking** | Niko Tosa (2000s) | Internalized physics; no device | **Legal but undetectable**; casinos use physical ejection |

The Pelayo family is the **purest example** of statistical bookkeeping as an advantage play method. They did not use technology in the sense of a device. They used **time, discipline, and mathematics**. And they won enough — and fought hard enough — to force a Supreme Court to declare that being too good at a legal game is not a crime.

---

**References**

1. Harvard Business Review, “The Pelayo Family Plays Roulette: The Prequel” 

2. BBC Radio Leeds, “The Bradford man who broke the Monte Carlo bank” 

3. IEEE Xplore, “The invention of the first wearable computer” 

4. UNLV Gaming Law Journal, “Roulette Computers” 

5. RTVE.es, “The Pelayos: the real story of a Spanish family” 

6. NPR, “Los Pelayos Beat The Wheel” 

7. Guinness World Records, “First wearable computer” 

8. Micro Cornucopia, “The Eudaemonic Pie” review 

9. Google Patents, US20200211327A1 

10. Wikipedia (IT), “Gonzalo García Pelayo” 

11. Wikipedia (archived), “Joseph Jagger” 

12. IEC Blog, “The first computer wearable” 

13. Sina Finance, “Edward Thorp autobiography” 

14. Bloomberg, “The Gambler Who Beat Roulette” 

15. Google Patents, EP3403249A1