Organic Roulette Experiment

Introduction

In September 2016, I came across claims suggesting that Organic Roulette machines from Interblock were employing aggressive tactics to counter players who clock the wheel. As someone who champions fair gambling practices, I found this subject particularly compelling. Over a span of six months, I partook in numerous discussions about this matter and gathered extensive data from thousands of spins. On this page, I recount the findings of my investigation.

The Game

Organic Roulette machines can be spotted in various casinos throughout Las Vegas. They function similarly to traditional table games, allowing players roughly 30 seconds to place their bets. Typically, the ball is launched with around 1 to 5 seconds left for players to make their wagers. Once betting concludes, it takes a few additional seconds for the ball to settle. Subsequently, winning bets are paid, and the entire process starts anew. Overall, the entire cycle lasts just a little over a minute.

The Allegation

It is feasible to win at the game if there is sufficient time between the moment the ball is released and when betting closes. Players skilled at wheel clocking can analyze both the speed of the ball and the wheel, even before bets are finalized. They can assess the ball's speed, the wheel's rotation, and its position at the moment of the launch to predict where the ball might land on the wheel.

Let's refer to the individual who alerted me to this situation as Mr. S. He indicated that casual players would likely experience a relatively fair game, with a winning chance of 1 in 38 for each number. However, if certain warning signs emerged, the game would enter a state he termed 'threatened mode.' Some of the indicators that caught the attention of wheel clockers were:

• Significant bets. The definition of 'significant' was somewhat vague, but bets totaling $20 were considered sufficient.
• Last-minute bets. Specifically, the five seconds that elapsed between the launching of the ball and the closing of the betting window.
• Cluster bets. These involve placing wagers on a few adjacent numbers on the wheel.

If the game sensed a threat, it was rumored that the machine would deploy aggressive countermeasures designed to influence the ball's landing away from the specific section bet by the wheel clocker. The exact mechanisms behind this were unclear, but Mr. S had some hypotheses that he requested I keep confidential.

Mr. S provided me with his analysis and data, which undoubtedly piqued my curiosity. However, he asked me not to disclose any specific details about it.

Global Gaming Expo

Initially, during my involvement in this project, I attended the 2016 Global Gaming Expo in Las Vegas, where Interblock had a prominently displayed booth. With Mr. S's permission, I inquired about speaking with one of the engineers present. The sales representative I spoke with was able to summon an engineer, who introduced himself. I took a seat at a similar machine to the Organic Roulette, which also allowed a grace period of about five seconds post-launch for betting to complete. Our conversation unfolded as follows, and I’ll paraphrase:

• Shackleford: I can't help but notice that on this machine, the ball is released several seconds before betting concludes.
• Interblock: And?
• Shackleford: Do you not think this might make it susceptible to teams clocking the wheel?
• Interblock: We are aware of that concern.

His demeanor and tone suggested he was hesitant to provide further commentary, so I expressed my gratitude for his time and moved on.

You can interpret that exchange as you wish.

First Experiment

After receiving Mr. S's presentation, I was certainly intrigued by the ongoing situation. As an initial measure, I compiled 299 spins of data along with another participant I’ll refer to as Mr. C. We adopted a strategy of placing $2 bets on five consecutive numbers. Mr. S was uncertain if total bets of $10 would suffice to make the game feel threatened but believed it was a good compromise to avoid irritating the game while minimizing financial loss during our investigation.

Neither Mr. C nor I practiced wheel clocking; instead, we approached the betting systematically, positioning our bets around the winning number from two spins prior. The objective was to compare the actual winning position on the wheel against the center of our betting arc.

Graph 1 and Table 1 illustrate the occurrences of how much the position of our center bet varied in relation to the winning number.

When applying a chi-squared goodness-of-fit test, we generated a chi-squared statistic of 45.29 with 37 degrees of freedom. The likelihood of a random game yielding results skewed more significantly than this is 16.5%.

Table 2 outlines the outcomes of a Gaussian test on different arc sizes that were centered around our main betting number in the first experiment. For the five-number arc we wagered on, the anticipated number of wins on a fair wheel would be 39.3. However, we only recorded 29 wins. The chance of experiencing such underperformance in a fair game is merely 3.84%.

While our results fell below expectations, it could still have been attributed to mere bad luck. Nonetheless, I calculated that with sustained losses at that rate, gathering a larger sample could solidify a case for unfair play. Following numerous discussions and emails, it was determined that we would advance with a secondary experiment.

Second Experiment

To establish a strong case asserting that a casino game is operating unfairly, the experiment must be reproducible. After considerable chatting, it was agreed to test for insufficient wins by placing $5 on each number within a five-number arc. We decided to conduct this over 1,000 spins across three different wheels. Besides Mr. S, at least one witness, either myself or Mr. C, would always be present. Once we completed our play, we would analyze the data and decide on our next steps. If it appeared we were encountering a non-random game beyond a reasonable doubt, we intended to present our findings to the public.

As this story reaches its conclusion, we had amassed 1,204 spins of data. While we recognized that our results were slightly below expectations, Mr. S assured us that the games were primarily performing randomly, with 'threatened' mode only activating sporadically.

Based on our analysis of the data, we concluded that our losses did not justify continued play.

The upcoming sections provide the results and analyses from our second experiment. To begin with, the following graph and Table 3 indicate the frequency of offsets between the center of our betting arc and where the ball landed.

After performing a chi-squared goodness-of-fit test, we found a chi-squared statistic of 34.79 with 37 degrees of freedom. The probability of a random game leading to results that are this skewed is 57.32%.

Table 4 presents the results from a Gaussian test conducted on various arc sizes centered on the main betting number. All tested arc sizes returned results well within the expected range.

Combined Results

The subsequent data represent a synthesis of both the first and second experiments. To kick things off, the following graph and Table 5 illustrate the frequency of offsets between our betting arc's center and the actual ball landing position.

A chi-squared goodness-of-fit test yields a chi-squared statistic of 41.75 with 37 degrees of freedom. The probability of a random game showing results more skewed than this is 27.19%.

Table 6 includes the findings from a Gaussian test regarding various arc sizes aligned with the center betting number. All tests remain within reasonable limits for a game deemed fair.

The correlation coefficient calculated between the outcomes of the first and second experiments stands at 0.153853. In simpler terms, this indicates a minor correlation, which falls within the expected behavior of two random data sets.

Conclusion

I uphold a very rigorous standard of proof before I would assert that any game isn’t adhering to the standard probabilities of an equitable game. In this instance, that threshold was not achieved. Simply put, I find no grounds to claim that the Organic Roulette machines are operating unfairly.

Responses

I extended Mr. S the opportunity to review this page prior to its public release. He offered many insightful suggestions and corrections. Ultimately, he still expressed concerns regarding some of my phrasing. Thus, I provided him the chance to publish his response, which he did. It is titled An Alternative Perspective on the Interblock Experiment Results by Mr. S.

I do appreciate putting my own spin on things, so I released a counter-response titled, Rebuttal to Mr. S's Feedback on the Organic Roulette Experiment by Michael Shackleford.

Written by: Michael Shackleford