Introduction to How to Win at Roulette: Traditional Visual Prediction
Create several flappers that are bent by a different amount and swap in the correct flapper for a given position of the wheel. There's no reason to make the flapper physically interact with the wheel — you can slow the wheel down with code. In fact, for so simple a set.
By Arnold Snyder(From Blackjack Forum , October 2012)
© 2012 Arnold Snyder
- There was a trick to getting the car every time, which was patched before I was told. I used to be able to move the left stick down real slowly and bump the bottom to get a slow spin. That's gone too. So, I think you're right that Cheatstar has manipulated the wheel into their favor. I haven't won a car in a while.
- The more a wheel is used, the more worn down it gets—and the more it may privilege certain numbers. In 1873, Joseph Jagger found a wonky wheel at Monte Carlo and bet on the biased numbers.
- Whether the wheel is at an online casino such as LeoVegas, real casino, in another country or in your home makes no difference. It is much more difficult to design a wheel with random spins, than it is to predict spins with reasonable accuracy. The only way to make a wheel 100% random is by not using a real and physical roulette wheel.
[Note from Arnold Snyder: This article deals specifically with American roulette, which doesn't make much of a difference for the purposes of this discussion except when I talk about what exact number the ball will land in if it rolls x spaces, etc. You European players have it so made with your lower house edge at roulette and better roulette culture that I'm going to let you translate those numbers for the European wheel yourself.
For reference, the house edge at American roulette is 5.26%. The house edge with a single-zero European roulette wheel is 2.70%. If you're betting one chip per spin on either wheel, you need to hit on average once out of every 36 spins to take the house edge to zero. If you can hit once out of every 35 spins, your edge will be 2.86%; once out of every 34 spins, 5.88%; once out of every 33 spins, 9.09%.
Part II of this article has also now been posted: How to Win at Roulette, Part 2: Dealer Steering and Tell Play.]
How to Win at Roulette
In the three decades since I began publishing Blackjack Forum, I have run into a number of roulette system sellers, as well as a number of books and video or DVD products put out on legitimate to semi-legitimate to phony methods for beating roulette, and I've been taught methods in person by people who have proven they know how to beat roulette under certain limited conditions.
The going rate for these systems right now is in the $200-$300 range (without personal instruction), though of course many of the system sellers promise the moon to every player who buys one of them. The legitimate system sellers prescribe essentially the same method as the one described by Laurance Scott in his article 'Nevada Roulette' (Blackjack Forum Vol. XI #3, September 1991):
'First, you must find a wheel with a predictable ball fall-off point. Second, you must be skilled at identifying an exact point within each ball spin at which to make your prediction--generally three to four revolutions before the ball actually drops from the track. Third, you must make your prediction based upon a visual observation of the ball in relation to both the position and velocity of the wheel. Finally, you must place your bets on the layout.'
The traditional method of betting once a player had his prediction has been to bet a sector of roughly ¼ of the wheel, or nine consecutive numbers on the wheel in the area of where you predict the ball will land. A number of system sellers advocate betting some random subset of this sector to camouflage your play.
Although I learned to beat an old-fashioned deep-pocket wheel in this way at home, I experienced a number of problems with this method in actual casino play. Essentially, I found the modern game, as dealt in the U.S. on modern low-profile wheels, impossible to make any money at with these methods. This article will discuss the problems I experienced in detail.
This article will also discuss the solutions we developed to these problems, and hopefully leave readers in a position to begin to win at roulette (at least any who are willing to practice and develop the necessary skills.) This article will also discuss the debate on roulette dealer section shooting, or steering of the ball, and methods of winning at roulette that exploit 'dealer signature.' There's a lot to talk about, and I expect that this article will be published in at least three parts over the next few months.
For now, it's important for you to understand that the traditional method of winning at roulette, as described by Scott above, is a method that was devised on slow-spinning deep pocket wheels, of which there are very few remaining in play in U.S. casinos. Everyone I know who discovered this method learned it in the course of trying to develop a roulette prediction computer, with a mental framework for how to go about that based on Edward Thorp's chapter on roulette in his book The Mathematics of Gambling, which I highly recommend.
Let's start with some definitions, then I'll go further into the details of this method—exactly how you use it to gain an edge on roulette, and what conditions must exist for you to use it. After that, I'll discuss the problems I had using this method with modern wheels and game conditions. Then I'll discuss a better method for beating modern roulette.
Roulette Definitions
Rotor: The rotor is the spinning part of a roulette wheel. It contains the numbered pockets where the ball will land.
Stator: The stator is the stationary part of a roulette wheel. It contains the spindle on which the rotor turns, as well as the sloped track that holds the spinning ball and the apron where the deflectors are located. The ball must pass across the apron, and through the deflectors, to enter the rotor and a numbered pocket.
Deflectors: Deflectors are the vertical and horizontal obstacles on the apron leading to the rotor. They are typically thicker and wider in the center, and thinner at the ends. They are spaced evenly and symmetrically around the wheel, with alternating vertical and horizontal obstacles. Most casino roulette wheels have 8 deflectors; a few older wheels have 16.
Frets: Frets are the dividers between the numbered pockets on the rotor.
Apron: The slanted surface on which the deflectors are placed, located between the roulette wheel track and the spinning rotor.
Wheel Track: A slightly sloped track on the vertical side of the stator that holds the ball in its spin around the outside of the wheel before it drops onto the rotor and into one of the numbered pockets.
The First Step to Winning at Roulette: An Explanation of Ball Roll
Assume for a moment that you are trying to predict where a roulette ball will land on a stationary wheel. The dealer has launched the ball into its spin with a given amount of energy. If you knew exactly the amount of energy transferred to the ball with that launch, and exactly how much energy the ball dissipated on a revolution around the wheel track, you would be able to predict the numbered pocket under the point where the ball was going to fall off the track.
When the ball falls off the track its remaining energy is not sufficient to keep it spinning fast enough to sustain its position on the nearly vertical track, but there is still a given amount of energy remaining from the launch that must be dissipated before the ball can come to rest in a pocket. The energy level at the moment of fall-off will be essentially consistent from spin to spin (or close enough for the purpose of prediction). So from the exact point of fall-off, the ball will have a given amount of energy, essentially equal each time, to dissipate in its roll before it can come to rest in a pocket.
If the ball's path from the wheel track to the numbered pockets on the rotor is free of obstacles, the remaining launch energy will be dissipated in a pretty consistent way, meaning the energy will be dissipated through contact with the frets between the numbered pockets. Sometimes the ball will hit a pocket in such a way that it makes contact with a lower portion of a fret, and then flies out of the pocket in such a way that it will skip over the next fret or two before making contact with another fret. Other times the ball will skip in a more consistent way across the tops of the frets, lightly hitting each one in its path.
Either way, the distance the ball will roll across the numbered pockets will be remarkably consistent. When a ball flies across several frets without touching them, it avoids dissipating the energy it would have lost in contact with those frets. But in order for a ball to take flight over several frets, it requires a certain kind of contact with an initial fret when the ball has too much energy to stop—so much energy, in fact, that the ball is propelled out of the pocket into flight. The distance the ball then travels before stopping in a pocket is about the same as the distance it would have traveled if it had lightly touched and skipped across several frets.
You will see some rolls where stranger things happen—the ball hits a fret and bounces backwards, for example, until it hits another fret and resumes its forward roll—but even then the backwards bounce will use up an amount of energy that tends to take the ball to a predictable spot. In deep pocket wheels, you sometimes see a ball 'get stuck' in one of the early pockets, and vibrate away the rest of its energy within that pocket. But the only ball behavior that makes a roll distance truly random is when the ball either turns into a 'floater' or gets shot out of the wheel altogether.
A 'floater' is a ball that gets stuck in what looks like a magnetic field on the edge of the rotor, above the pockets. The ball gets held there, by whatever force, until it's slowed to way less than the normal speed at which it would normally be coming to rest in a pocket. We've even seen floaters come to a dead stop on the edge of the rotor and just hang there without ever dropping into a pocket, until the dealer gives up on it and gets permission from the pit boss to pick it up and spin it again.
Floaters occur on wheels that are spun at ridiculously fast speeds—the same speeds at which the ball tends to get shot out of a modern, low-profile wheel. They would be a great countermeasure against roulette predictors, except that they slow down the game too much. Nothing clears out a table of roulette squares like a dealer who consistently spins the wheel at speeds that produce floater after floater, interrupted by spins that shoot the ball out of the wheel.
As mentioned above, that's the other thing that truly randomizes ball roll: a wheel that's spinning so fast that the ball's first contact with a fret hurls the ball entirely out of the wheel into the eye of a passerby. It's a measure of the lack of roulette professionalism in Las Vegas that such an event is no longer uncommon, even in the supposedly elegant joints. Routine beaning of passersby would be another great countermeasure against roulette prediction, but again, it slows down the game too much, increases the risk of lawsuits by injured casino patrons, and tends to clear the squares off a table.
Some casinos, apparently determined to have their dealers spin their wheels at levitation speeds, actually install clear plastic shields over the tops of their wheels, so that the flying ball is stopped by the shield instead of beaning passersby. Unfortunately for the casino employees who are sweating the game, a hit on a shield is not a true roll randomizer, nor an effective countermeasure against roulette predictors. A hit on the shield simply absorbs a consistent amount of energy, and takes the remaining roll out of the ball. It's very similar to the effect of a ball hitting a particular spot on a vertical deflector. More on that below.
Thorp calls the behavior of the ball upon hitting the frets 'spatter,' and notes, correctly, that it has some randomizing effect, but not enough, according to his measurements (or ours) to deter him from developing a roulette computer. What we've found is that spatter tends to look a lot crazier than it really is, and that overall a fixed amount of energy tends to take a ball a fixed distance across the numbered pockets on a rotor, especially on the newer wheels with low frets that now dominate the game in Las Vegas and most of the U.S.
In his letter (Blackjack Forum, December 1991—link at the left) responding to Scott's 'Nevada Roulette,' Darwin Ortiz says that even a breeze blowing across the wheel can disturb the movement of the ball enough to negate predictions. We didn't find things like air movement or the weather (with a few rare exceptions) to be enough of a source of randomness in either the ball spin or roll to negate the ability to get an edge from prediction.
For example, we often practiced in a room with a big open window, and there was often a breeze in the room, and we didn't notice any impact of breezes of various strengths on our ability to predict where the ball would land. The ball has a certain amount of intense energy transferred to it by the dealer's action of launching it into its spin, and it has to dispel that energy. If it's pushing against a headwind during one portion of a spin, that headwind becomes a backwind when it reaches a different point on the wheel.
A good strong wind in a room probably has about the same effect on a ball's spin as a slightly tilted wheel—if a ball's about to drop off the track, it could make it drop a slight bit earlier. If a ball gets past the headwind and is about to drop, it might go a little farther than it would have because the wind is now a backwind. All of this is academic of course, since no one deals roulette games outdoors, and casinos are not really very breezy places.
For the moment let's just note that it takes a big force, like a sudden huge drop in atmospheric pressure due to the arrival of a big storm, to change the behavior of a roulette ball enough to force you to modify how you predict. In fact, I got a watch with a built-in barometer to keep track of these kinds of changes, because you can't see the arrival of a major storm when you're sitting inside a casino.
In summary, if you know where a roulette ball is going to fall off the wheel track, and you assume its behavior after leaving the track won't be affected by the deflectors on the apron, you can predict where it will come to rest in the numbered pockets—at least closely enough to get a good edge.
What about Differences between Roulette Balls?
If the only force affecting a roulette ball were gravity, its density, mass and size would not affect the spin or prediction. (That was the whole point of the Galileo Leaning Tower of Pisa experiment.) However, gravity is not the only force affecting spin and prediction, because the ball must travel on a track and through air.
There are well-established mathematical formulae to express all this, but essentially a larger ball is affected more by air resistance and drag, all other factors being equal, while a ball with greater density is affected less by air resistance and drag. For example, as you gain experience at roulette, you will notice that when denser balls are used, the ball speed that usually designates three spins until drop-off (with lighter balls) will instead signify five or six spins until drop-off.
In practice, what you do is calibrate your predictions for whatever ball the dealer is actually using, and recalibrate if the dealer changes balls. (There aren't that many types of roulette balls out there, so you will quickly get a feel for differences in their behavior if you get to play each type regularly enough, or can experiment with them at home.)
Scott talks about dealers switching balls mid-shift as a predictor countermeasure, but we never saw that. The few times we observed a dealer switching balls mid-shift, it seemed to be related to the dealer trying to hit a number that a lot of players had bet on. [Note: if a dealer is changing balls mid-shift to try to hit a number, it's a sure sign that the dealer is incapable of hitting anywhere near the desired number except through sheer luck.]
Sometimes, after a few spins, a new dealer who has just come to the table will select a different ball. The typical motivation seems to be dissatisfaction with the number of spins the dealer is getting from the original ball. But again, we seldom saw dealers change balls in the middle of a stint at a roulette table. That doesn't mean Scott didn't see it back at the time of his article. It just means we haven't seen it except on rare occasions in the past ten or so years in Las Vegas.
Now let's look at the deflectors and their effect on the roll.
How to Win at Roulette: Understanding the Deflectors
As Thorp observed in The Mathematics of Gambling, the greatest factor in making a ball's roll inconsistent from spin to spin is the obstacle course of deflectors (also called diamonds, or pins) that it must pass through on the apron before entering the rotor. Modern roulette wheels typically come with four vertical deflectors spaced evenly around the wheel, and four horizontal deflectors spaced evenly around the wheel between the vertical deflectors.
Both the vertical and horizontal deflectors are usually significantly thicker in the middle than at the top and bottom of each pin. If they're not significantly thicker, they are significantly wider. (On modern wheels, they're usually both.)
If a ball hits the thick middle part of a vertical deflector after leaving the track, it will dissipate most of its remaining energy in that hit. A ball that hits the middle of a vertical deflector typically rolls very few numbers once it reaches the rotor. The exact number of pockets it rolls will depend on the exact type of deflector and where exactly the ball hit the deflector. The roll may also be influenced slightly by other factors, like tilt.
If a ball makes its way across the apron without hitting any of the deflectors, it dissipates none of its energy on this kind of hit, so it takes a lot longer roll to use up all its remaining energy.
A ball that hits the bottom or top tip (the thin or narrow portion) of a vertical deflector dissipates a portion of its remaining energy in that hit, but not as much as if it hit the middle. It will roll further than a ball that hit the middle of the deflector, but not as far as a ball that hits no deflector at all.
Hits on horizontal deflectors use up energy in a way similar to hits on vertical deflectors.
As Thorp noted correctly about deflectors (The Mathematics of Gambling, p. 53): 'On average, perhaps half the time these have a significant effect on the ball.'
How the Deflectors Steer the Roulette Ball
How To Make The Casino Wheel Spin Slowly
I've just described a variety of ways that deflectors affect ball roll. You might think, because of this variety induced by the deflectors, that it would be impossible to get an edge predicting which number will hit, since it will depend on whether and how the ball hits the deflectors on the way to the rotor.
But in fact, the ways the ball interacts with deflectors tend to cluster in useful ways on any particular wheel. (In other words, you tend to see more of one or two types of hits or non-hits than any of the other theoretically possible types.)
On some wheels, the deflectors even perform a useful type of steering. For example, on one wheel in Las Vegas a ball will either hit the center of a vertical deflector and drop dead into the pocket directly below, or lightly cross the tip of that deflector, start moving down, and then curve back up before falling, so that the ball spends a lot more time on the apron that you'd normally see after a ball leaves the track. This allows more numbers to pass by below, especially on today's fast-spun wheels. The ball will then roll enough spaces to bring it back to the number that it would have fallen into if it had hit the middle of the deflector.
On another wheel, it either hits the middle of a vertical and rolls 5-6 numbers, or it hits the tip of the same deflector then rolls 24-25 numbers from the number that was under the deflector at the time of the hit. That is a terrific wheel to play if you broaden your concept of a 'sector' for betting purposes. (I will discuss optimal betting in Part II of this article.)
So, if the casinos all run out and change the deflectors and deflector positions on their wheels, will that kill the chance for advantage play? No, it'll just change how the ball is steered on any given wheel, and thus change the optimal numbers to bet.
How to Win at Roulette: More on the Drop
There are very few roulette wheels out there in real life casinos that don't have some kind of a drop (when I say 'very few' I mean I've never seen such a wheel, but I presume there could be a few out there somewhere in this wide world). By 'drop' I mean a place where the ball tends to fall off the track at a higher rate than would be expected if the wheel were perfectly level and perfectly machined.
A typical roulette wheel has a dominant drop point with a secondary drop point that kicks in if the ball gets past the dominant drop point. If you look at enough wheels in enough casinos, you'll see all kinds of drops and secondary drops—even drops where the dominant drop point is at, say, twelve o'clock, and the secondary drop point is at six o'clock. But usually they're within a short distance of each other.
The reason you need a drop to predict roulette is because it's so difficult to estimate visually, even when you know the ball is roughly three spins from dropping off the track, exactly how much energy the ball has left in it. In other words, you might know from the look of the ball that it has slightly more than three spins to go but less than four, but on a perfect wheel it would be difficult to tell whether the ball had exactly 3 1/4 spins to go from your read point, 3 1/3 spins, 3 ½ spins, etc. If you don't know where the ball's final roll across the apron and numbered pockets is going to start, you can't predict where it will end.
A wheel with a drop takes care of that problem pretty well. The drop itself uses up the excess energy that might have caused a ball with borderline energy to go past the drop point if the wheel was perfect. A good drop gives you a ball falling off the track at a consistent place or places with a consistent amount of energy to spend on the roll to its landing place.
Wood warps and cracks, players and dealers lean on tables, floors are never perfectly level in the first place. In all the years we've spent dealing with roulette, we never saw any maintenance guy out there checking a wheel with a level.
A drop usually seems caused by a tilt in the surface on which the wheel is resting. The way you can tell that the drop is caused by tilt rather than an imperfection in the track of the roulette wheel is through the effect of the drop on the behavior of the ball.
When a drop is caused by a tilt, the ball tends to speed up and go further than you'd expect after the dominant drop point if it makes it past that point near the end of its spin. There will also be a predictable number of balls that don't quite make it up the hill to the drop-off point. You'll see virtually no balls fall off the track at the low point of a tilted wheel, and you'll see no really weird ball behavior before hitting the frets.
By contrast, if you deliberately create an imperfection in the track of a roulette wheel, and a ball makes it past that imperfection near the end of its spin, you'll see very strange ball behavior after the track fall-off point. In Las Vegas right now, at a popular Strip casino, the ball leaves the track reliably at the same point every spin but will often swoop almost back up to the track in a very drunken-looking way before proceeding across the apron. We know this drop is due to an imperfection in the wheel because we've seen the drop rotate around the wheel—we've played a six o'clock drop on this wheel and a nine o'clock drop, etc., and in all cases the weird drunken ball behavior was in play.
Wheels that develop a drop tend to keep a drop. There has been some debate about this in the literature. In his article, 'The Myth of Roulette Dealer Signature and ‘Sector Shooting'' (Blackjack Forum Vol. XII No. 2, June 1992—link at upper left) Steve Forte wrote:
'Consider the capricious nature of the wheel. It is an undeniable fact that the characteristics of a specific wheel that may theoretically make it beatable one day can change the next day, or even the next minute, making the game unbeatable. Even the same dealer, same equipment and similar measurements of ball and rotor speed will yield different results at different times.
I proved this phenomenon to myself some years back after spending 17 days on the road scouting, evaluating conditions, and recording data on hundreds of wheels.'
Laurance Scott states the same opinion in 'Nevada Roulette.' He cautions: '…there are inherent factors of the game that cause wheels to phase in and out of predictability,' and, 'I have scouted over 300 wheels and only a handful exhibit consistent behavior day after day.'
I recognize the variability that Forte and Scott are talking about, but in my view it's due to a very consistent factor in the game that requires a different approach to getting an edge than the traditionalists use.
Basically, in my experience most wheels that develop a drop keep the same drop for years, even decades, and thus retain the same predictability for years or decades. What varies, and causes the wheel to seem unpredictable to traditional wheel predictors, is the way the ball drops at the drop—in other words, the style of ball behavior at the drop point. But all drops produce this variability, and the proper response to it is to expect it at all times, on every wheel, and alter your betting pattern from the traditional approach to take account of it.
Let's say you find a wheel where the ball drops off between 12 o'clock and 1 o'clock 100% of the time. But it hits the vertical deflector at the start of the drop-off area roughly 1/3 of the time. It hits a tip of that same vertical deflector roughly 1/3 of the time. And it sails past that deflector to drop without any contact with a deflector roughly 1/3 of the time. Can you play that wheel or not? (Answer: I haven't given you enough information to judge.)
When you first walk up to that wheel to scout it, you may see the ball hit the middle of that vertical deflector 8 times out of 10, and start exulting in the consistency of the drop. Any wheel that was dropping at the same point, with the same deflector hit each time, 8 out of every 10 spins, would be a wheel you could murder.
What Laurance Scott and Steve Forte seem to be saying is that you actually require a wheel like that to win at roulette. If this were true, the only predictable wheels would be ones with a drop where it almost always—a high percentage of the time—either smacks the center of a vertical pin, or lightly nicks the top of a pin, or makes it from the track to the rotor without hitting any pins—one or another, without mixing it up.
But even if you come across a wheel that appears to be this consistent, how can you be sure that's the 'real' drop style instead of just normal variance in the drop style, like the normal variance in heads and tails results when flipping a coin? The answer is you can't without watching the wheel so long that your eyes start crossing and you're no longer fit to play.
People tell me they've actually played such wheels, and they really seem to mean it. But in over 10 years of learning roulette, scouting roulette, staring at roulette, cursing roulette, and playing roulette, we've never found a wheel consistent enough to meet thisdefinition, and we've never been able to get a wheel as consistent as this at home either, no matter what we've done to a wheel.
Instead, what we've seen in real life are wheels that give you 8 drops out of 10 with a consistent style (or even 16 out of 20), then 2 out of the next 10 with that style, then 6 out of 10, then 1 out of 10, then 4 out of 10.
If you're flipping coins, you can get streaks of heads or tails, and if a wheel has a drop, you'll get streaks of one style of drop at the drop-off, but sooner or later, and usually sooner, it will revert to its normal overall ratio of drop styles, which will be much more variable. In other words, you can get short-term streaks where you only see the ball hitting the pin or missing the pin a very high percentage of the time, but the streaks are unpredictable in length and don't last at roulette. A typical drop will hit about 50% of the time, as Thorp noted.
The drop is eternal; the variety of ball behaviors at the drop point comes from normal variance in results that has nothing to do with breezes or the weather or oil on the dealer's fingertips or any other external conditions (most of the time). Again, the best way to deal with this problem is usually by revising your idea of optimal betting, which I'll discuss in Part II of this article.
How to Win at Roulette: Introduction to the Rotor Speed Prediction Adjustment
For the purposes of understanding roulette prediction, I suggest you look at the ball and wheel this way:
A particular ball launched onto a particular round roulette wheel track with a given amount of energy means a given length of time for that ball to dissipate enough of that launch energy to fall off the track. For the purposes of roulette prediction, it's the speed of the ball that sets the time.
Now, instead of thinking of a roulette wheel as a round rotating object, think of the numbers on it stretched out in a straight line along a road, with the numbers repeating, in the same order, every time you come to the end of the set of numbers. Wheel speed sets the amount of road that will pass during the time the ball is on the track.
If a wheel is spinning faster than speed x, more of the road will pass during the ball spin than passed by at speed x. If a wheel is spinning slower than speed x, less of the road will pass during the ball spin than passed by at speed x.
Here's a very simple example of an adjustment for wheel speed estimation. Assume a wheel spinning at a moderate speed (by U.S. standards today), moving from right to left. You take a read of the numbered pocket below the drop-off point three spins before the ball will fall off the track there. The read is 13. The wheel goes through 2 ½ revolutions, i.e. 97 numbers pass the drop-off point during the time until the ball drops off the track. The ball thus drops off the track at 14. It hits no pins, rolls 12 numbers (this is a deep pocket wheel), and comes to rest on 34.
For the next trial, you again see a slowly spinning wheel, moving from right to left, but you judge that the dealer has spun the wheel 10% faster than on the first trial. (I'll get into how you make that judgment in Part II). So you judge that roughly 107 numbers will pass the drop-off point from the time you take the read three spins out until drop-off.Your read at the drop-off three spins out is the same—13. If your speed read is correct, the number 5 will be under the ball at the drop. The ball again hits no pins, and rolls exactly 12 numbers. This time it lands on the 25. From our starting rotor speed, a 10% increase in rotor speed, with all other conditions the same, has moved the landing point back 10 numbers.
Obviously when a wheel is moving from right to left, a move 'backwards' is further to the right. When a wheel is moving from left to right, a move 'backwards' is further to the left.
Now you're calibrated. If the wheel speed had been increased 20% from the speed of the first spin, you'd expect roughly 116 numbers to pass while the ball remained on the track. That means, assuming once again no deflector hit, and the same 12-number roll, that we'd expect the ball landing point to move back roughly 19 numbers (116-97 = 19).
This is a greatly simplified explanation of how you adjust for wheel speed. It's a very common-sense adjustment. If a wheel is going faster on one spin than another, you'd expect more of the wheel to pass during the time until drop off. So you'd expect the ball to land further back from the read number than you'd expect on a slower wheel.
Because a wheel is round, it should be obvious that at some point an increase in speed takes you back to your read point. A slightly greater increase in speed than that would take you to your read point and then slightly past it.
I'll give tips in Part II of this article on how to learn to make speed adjustments.
One skill you might get working on in the meantime is developing your ability to pick up speed clues. It's relatively easy for your eye to pick out the single-digit numbers on the rotor above the pockets—even on a wheel that's spinning very fast, the narrower single-digit numbers stand out amidst the wider double-digit numbers.
How To Make The Casino Wheel Spin Slow Motion
In fact, if you relax your eyes and get used to the effect of the single-digit numbers passing your field of vision, you'll find they tend to create a pulsing effect. At fast speeds, you may have to train yourself to pick up the beat of every second single-digit number. What you want to start to do is hear the rhythm of that pulse in your head as you watch a wheel.
You might try clicking your tongue against your teeth to give you the sound (don't let the dealer or pit boss hear you). The sound of that pulse is how you will train yourself to read wheel speed. Wear a watch with a second hand and start paying attention to the number of clicks you get in some fixed number of seconds so that you start to get a feel for the look and sound of a wheel that's 10% or 20% slower or faster.
I've already mentioned that the speed adjustment example above is greatly simplified. What adds a lot to the thrills and chills factor of predicting roulette is that other conditions will usually not remain constant. In our example, the ball never hit a deflector and always rolled twelve pockets. In real life, you'll never find such consistency. We'll talk more about that in Part II of this article.
First, let's look at the problems involved in trying to use the traditional method in today's game.
The Effect of Modern Low Profile Roulette Wheels on Results from the Traditional Method
One of the things many of the traditionalists seem to agree on is that modern roulette wheels tend to give truly random results. For example, in his article 'Nevada Roulette,' Scott says:
'Some casinos run a fair game. They use modern wheels which tend to yield truly random results. However, other casinos still use older style equipment which is quite beatable. Why would a casino use beatable equipment when modern non-beatable equipment is available?'
Scarne agrees. In his 1978 revision of his 1961 book, Scarne's Guide to Casino Gambling, he states emphatically that no roulette dealer can section shoot with accuracy on modern roulette wheels because the wheels randomize results too much.
But that's not what we saw. I can tell you definitively that modern wheels are every bit as beatable as the old-fashioned deep pocket wheels, and may even be more beatable.
It's true that old-fashioned deep pocket wheels tend to grab a ball earlier in its roll. The rest of the ball's energy seems to get dissipated vibrating in that deep pocket.
But that doesn't make lower profile wheels any less beatable. In fact, because lower frets interfere less with the natural roll of the ball in dissipating the last of its energy, results are actually more consistent with a low profile wheel than with a deep pocket wheel.
I believe what makes results on modern wheels appear more random to traditional roulette predictors is
- the increased speed at which dealers can spin modern wheels without causing the ball to fly out of the wheel tends to make rotor speed adjustments more difficult for predictors and prediction mistakes more likely, and
- betting a 9-consecutive-number sector on a modern wheel is, over the long run, not only not optimal—it's a losing proposition.
It's relatively easy to make adjustments for rotor speed when a dealer keeps that speed in a fairly narrow range.And dealers tend to be more consistent with their rotor speeds on old-fashioned, deep pocket wheels than they are on the low-profile modern wheels because the deep-pocket wheels tend to launch the ball out of the wheel at far lower speeds.
In other words, on a deep pocket wheel, a dealer quickly learns to keep wheel speed moderate to avoid spending half her shift chasing down balls launched into the aisles.
There is a point on modern low-profile wheels where even the lowest frets will hurl the ball out of the wheel and into the faces of passersby, but you don't get to that point until the wheel is much faster. So on modern low-profile wheels, dealers can and do use a much wider range of wheel speeds (mostly on the faster end of the range), and they tend to be less consistent in rotor speed from spin to spin.
It's simply harder to judge the exact speed of a wheel that's gone from really fast to supersonic than it is to judge the exact speed of a wheel that's gone from slow to slightly less slow.
Also, on the modern low-profile wheels, on which the final roll of a ball is much smoother than on a deep-pocket wheel, the ball is much more likely to roll the exact number of pockets dictated by the remaining energy when it falls off the track. It's almost impossible for the ball to get stuck in a pocket in the traditional sector and just vibrate the remainder of its energy away.
As a result, what we've found on modern low-profile wheels is that there's not only no edge on most of the numbers in a traditional 9-number betting sector, there's actually a strong house edge on most of those numbers—maybe not the full house edge of 5.26%, but not a player edge, and enough of a house edge to kill the player edge overall. Reading the wheel accurately, but betting it wrong, leads to dismal results.
On deep pocket wheels, by contrast, which are spun at much slower speeds, all of the numbers in the traditional sector had enough of an edge to make them worth betting. We believe the difference is due to the effect of errors in speed adjustments. On a slow wheel, if you're off a little, the ball will often still land in the traditional betting sector. On a fast wheel, being off a little takes the ball much farther away. Small mistakes kill you on fast wheels.
More Problems with the Traditional Method of Winning at Roulette
Laurance Scott used to put out a video with his system; maybe a similar dvd is included in the current package. It was essentially a video of a slowly spinning deep-pocket roulette wheel, filmed from straight overhead. My wife learned to predict that wheel within about 15 minutes, on a very small TV, from the opposite side of the room, where she was glancing up from reading in bed.
Great. She had a terrific edge on bedroom video roulette. When we looked for that game in a real casino, we never found it.
Another roulette predictor who demonstrated his abilities to us did so from a balcony overlooking a deep-pocket roulette wheel—and I agree, this is an ideal way to do it, with no heat from the dealer or wave-offs to worry about—except that you can't make any money doing imaginary betting from a balcony.
We asked this player to actually scout low-profile wheels with us and let us sit in on a play. I told him we couldn't beat them with his method and we wanted him to demonstrate. He never came through, even though he was soliciting my help in selling very expensive roulette prediction lessons to other pros. In other words, there was a potentially huge financial incentive for him to demonstrate his prowess on an actual modern game, but he's never made himself available.
In addition to finding no edge on most of the numbers in a traditional 9-number betting sector on modern wheels, and the overwhelming odds against finding wheels you can beat with the traditional method, here are the problems we ran into trying to use traditional roulette prediction methods on modern live games.
First, dealers wave you off long before you can bet if you are trying to predict from three to four spins out.
Create several flappers that are bent by a different amount and swap in the correct flapper for a given position of the wheel. There's no reason to make the flapper physically interact with the wheel — you can slow the wheel down with code. In fact, for so simple a set.
By Arnold Snyder(From Blackjack Forum , October 2012)
© 2012 Arnold Snyder
- There was a trick to getting the car every time, which was patched before I was told. I used to be able to move the left stick down real slowly and bump the bottom to get a slow spin. That's gone too. So, I think you're right that Cheatstar has manipulated the wheel into their favor. I haven't won a car in a while.
- The more a wheel is used, the more worn down it gets—and the more it may privilege certain numbers. In 1873, Joseph Jagger found a wonky wheel at Monte Carlo and bet on the biased numbers.
- Whether the wheel is at an online casino such as LeoVegas, real casino, in another country or in your home makes no difference. It is much more difficult to design a wheel with random spins, than it is to predict spins with reasonable accuracy. The only way to make a wheel 100% random is by not using a real and physical roulette wheel.
[Note from Arnold Snyder: This article deals specifically with American roulette, which doesn't make much of a difference for the purposes of this discussion except when I talk about what exact number the ball will land in if it rolls x spaces, etc. You European players have it so made with your lower house edge at roulette and better roulette culture that I'm going to let you translate those numbers for the European wheel yourself.
For reference, the house edge at American roulette is 5.26%. The house edge with a single-zero European roulette wheel is 2.70%. If you're betting one chip per spin on either wheel, you need to hit on average once out of every 36 spins to take the house edge to zero. If you can hit once out of every 35 spins, your edge will be 2.86%; once out of every 34 spins, 5.88%; once out of every 33 spins, 9.09%.
Part II of this article has also now been posted: How to Win at Roulette, Part 2: Dealer Steering and Tell Play.]
How to Win at Roulette
In the three decades since I began publishing Blackjack Forum, I have run into a number of roulette system sellers, as well as a number of books and video or DVD products put out on legitimate to semi-legitimate to phony methods for beating roulette, and I've been taught methods in person by people who have proven they know how to beat roulette under certain limited conditions.
The going rate for these systems right now is in the $200-$300 range (without personal instruction), though of course many of the system sellers promise the moon to every player who buys one of them. The legitimate system sellers prescribe essentially the same method as the one described by Laurance Scott in his article 'Nevada Roulette' (Blackjack Forum Vol. XI #3, September 1991):
'First, you must find a wheel with a predictable ball fall-off point. Second, you must be skilled at identifying an exact point within each ball spin at which to make your prediction--generally three to four revolutions before the ball actually drops from the track. Third, you must make your prediction based upon a visual observation of the ball in relation to both the position and velocity of the wheel. Finally, you must place your bets on the layout.'
The traditional method of betting once a player had his prediction has been to bet a sector of roughly ¼ of the wheel, or nine consecutive numbers on the wheel in the area of where you predict the ball will land. A number of system sellers advocate betting some random subset of this sector to camouflage your play.
Although I learned to beat an old-fashioned deep-pocket wheel in this way at home, I experienced a number of problems with this method in actual casino play. Essentially, I found the modern game, as dealt in the U.S. on modern low-profile wheels, impossible to make any money at with these methods. This article will discuss the problems I experienced in detail.
This article will also discuss the solutions we developed to these problems, and hopefully leave readers in a position to begin to win at roulette (at least any who are willing to practice and develop the necessary skills.) This article will also discuss the debate on roulette dealer section shooting, or steering of the ball, and methods of winning at roulette that exploit 'dealer signature.' There's a lot to talk about, and I expect that this article will be published in at least three parts over the next few months.
For now, it's important for you to understand that the traditional method of winning at roulette, as described by Scott above, is a method that was devised on slow-spinning deep pocket wheels, of which there are very few remaining in play in U.S. casinos. Everyone I know who discovered this method learned it in the course of trying to develop a roulette prediction computer, with a mental framework for how to go about that based on Edward Thorp's chapter on roulette in his book The Mathematics of Gambling, which I highly recommend.
Let's start with some definitions, then I'll go further into the details of this method—exactly how you use it to gain an edge on roulette, and what conditions must exist for you to use it. After that, I'll discuss the problems I had using this method with modern wheels and game conditions. Then I'll discuss a better method for beating modern roulette.
Roulette Definitions
Rotor: The rotor is the spinning part of a roulette wheel. It contains the numbered pockets where the ball will land.
Stator: The stator is the stationary part of a roulette wheel. It contains the spindle on which the rotor turns, as well as the sloped track that holds the spinning ball and the apron where the deflectors are located. The ball must pass across the apron, and through the deflectors, to enter the rotor and a numbered pocket.
Deflectors: Deflectors are the vertical and horizontal obstacles on the apron leading to the rotor. They are typically thicker and wider in the center, and thinner at the ends. They are spaced evenly and symmetrically around the wheel, with alternating vertical and horizontal obstacles. Most casino roulette wheels have 8 deflectors; a few older wheels have 16.
Frets: Frets are the dividers between the numbered pockets on the rotor.
Apron: The slanted surface on which the deflectors are placed, located between the roulette wheel track and the spinning rotor.
Wheel Track: A slightly sloped track on the vertical side of the stator that holds the ball in its spin around the outside of the wheel before it drops onto the rotor and into one of the numbered pockets.
The First Step to Winning at Roulette: An Explanation of Ball Roll
Assume for a moment that you are trying to predict where a roulette ball will land on a stationary wheel. The dealer has launched the ball into its spin with a given amount of energy. If you knew exactly the amount of energy transferred to the ball with that launch, and exactly how much energy the ball dissipated on a revolution around the wheel track, you would be able to predict the numbered pocket under the point where the ball was going to fall off the track.
When the ball falls off the track its remaining energy is not sufficient to keep it spinning fast enough to sustain its position on the nearly vertical track, but there is still a given amount of energy remaining from the launch that must be dissipated before the ball can come to rest in a pocket. The energy level at the moment of fall-off will be essentially consistent from spin to spin (or close enough for the purpose of prediction). So from the exact point of fall-off, the ball will have a given amount of energy, essentially equal each time, to dissipate in its roll before it can come to rest in a pocket.
If the ball's path from the wheel track to the numbered pockets on the rotor is free of obstacles, the remaining launch energy will be dissipated in a pretty consistent way, meaning the energy will be dissipated through contact with the frets between the numbered pockets. Sometimes the ball will hit a pocket in such a way that it makes contact with a lower portion of a fret, and then flies out of the pocket in such a way that it will skip over the next fret or two before making contact with another fret. Other times the ball will skip in a more consistent way across the tops of the frets, lightly hitting each one in its path.
Either way, the distance the ball will roll across the numbered pockets will be remarkably consistent. When a ball flies across several frets without touching them, it avoids dissipating the energy it would have lost in contact with those frets. But in order for a ball to take flight over several frets, it requires a certain kind of contact with an initial fret when the ball has too much energy to stop—so much energy, in fact, that the ball is propelled out of the pocket into flight. The distance the ball then travels before stopping in a pocket is about the same as the distance it would have traveled if it had lightly touched and skipped across several frets.
You will see some rolls where stranger things happen—the ball hits a fret and bounces backwards, for example, until it hits another fret and resumes its forward roll—but even then the backwards bounce will use up an amount of energy that tends to take the ball to a predictable spot. In deep pocket wheels, you sometimes see a ball 'get stuck' in one of the early pockets, and vibrate away the rest of its energy within that pocket. But the only ball behavior that makes a roll distance truly random is when the ball either turns into a 'floater' or gets shot out of the wheel altogether.
A 'floater' is a ball that gets stuck in what looks like a magnetic field on the edge of the rotor, above the pockets. The ball gets held there, by whatever force, until it's slowed to way less than the normal speed at which it would normally be coming to rest in a pocket. We've even seen floaters come to a dead stop on the edge of the rotor and just hang there without ever dropping into a pocket, until the dealer gives up on it and gets permission from the pit boss to pick it up and spin it again.
Floaters occur on wheels that are spun at ridiculously fast speeds—the same speeds at which the ball tends to get shot out of a modern, low-profile wheel. They would be a great countermeasure against roulette predictors, except that they slow down the game too much. Nothing clears out a table of roulette squares like a dealer who consistently spins the wheel at speeds that produce floater after floater, interrupted by spins that shoot the ball out of the wheel.
As mentioned above, that's the other thing that truly randomizes ball roll: a wheel that's spinning so fast that the ball's first contact with a fret hurls the ball entirely out of the wheel into the eye of a passerby. It's a measure of the lack of roulette professionalism in Las Vegas that such an event is no longer uncommon, even in the supposedly elegant joints. Routine beaning of passersby would be another great countermeasure against roulette prediction, but again, it slows down the game too much, increases the risk of lawsuits by injured casino patrons, and tends to clear the squares off a table.
Some casinos, apparently determined to have their dealers spin their wheels at levitation speeds, actually install clear plastic shields over the tops of their wheels, so that the flying ball is stopped by the shield instead of beaning passersby. Unfortunately for the casino employees who are sweating the game, a hit on a shield is not a true roll randomizer, nor an effective countermeasure against roulette predictors. A hit on the shield simply absorbs a consistent amount of energy, and takes the remaining roll out of the ball. It's very similar to the effect of a ball hitting a particular spot on a vertical deflector. More on that below.
Thorp calls the behavior of the ball upon hitting the frets 'spatter,' and notes, correctly, that it has some randomizing effect, but not enough, according to his measurements (or ours) to deter him from developing a roulette computer. What we've found is that spatter tends to look a lot crazier than it really is, and that overall a fixed amount of energy tends to take a ball a fixed distance across the numbered pockets on a rotor, especially on the newer wheels with low frets that now dominate the game in Las Vegas and most of the U.S.
In his letter (Blackjack Forum, December 1991—link at the left) responding to Scott's 'Nevada Roulette,' Darwin Ortiz says that even a breeze blowing across the wheel can disturb the movement of the ball enough to negate predictions. We didn't find things like air movement or the weather (with a few rare exceptions) to be enough of a source of randomness in either the ball spin or roll to negate the ability to get an edge from prediction.
For example, we often practiced in a room with a big open window, and there was often a breeze in the room, and we didn't notice any impact of breezes of various strengths on our ability to predict where the ball would land. The ball has a certain amount of intense energy transferred to it by the dealer's action of launching it into its spin, and it has to dispel that energy. If it's pushing against a headwind during one portion of a spin, that headwind becomes a backwind when it reaches a different point on the wheel.
A good strong wind in a room probably has about the same effect on a ball's spin as a slightly tilted wheel—if a ball's about to drop off the track, it could make it drop a slight bit earlier. If a ball gets past the headwind and is about to drop, it might go a little farther than it would have because the wind is now a backwind. All of this is academic of course, since no one deals roulette games outdoors, and casinos are not really very breezy places.
For the moment let's just note that it takes a big force, like a sudden huge drop in atmospheric pressure due to the arrival of a big storm, to change the behavior of a roulette ball enough to force you to modify how you predict. In fact, I got a watch with a built-in barometer to keep track of these kinds of changes, because you can't see the arrival of a major storm when you're sitting inside a casino.
In summary, if you know where a roulette ball is going to fall off the wheel track, and you assume its behavior after leaving the track won't be affected by the deflectors on the apron, you can predict where it will come to rest in the numbered pockets—at least closely enough to get a good edge.
What about Differences between Roulette Balls?
If the only force affecting a roulette ball were gravity, its density, mass and size would not affect the spin or prediction. (That was the whole point of the Galileo Leaning Tower of Pisa experiment.) However, gravity is not the only force affecting spin and prediction, because the ball must travel on a track and through air.
There are well-established mathematical formulae to express all this, but essentially a larger ball is affected more by air resistance and drag, all other factors being equal, while a ball with greater density is affected less by air resistance and drag. For example, as you gain experience at roulette, you will notice that when denser balls are used, the ball speed that usually designates three spins until drop-off (with lighter balls) will instead signify five or six spins until drop-off.
In practice, what you do is calibrate your predictions for whatever ball the dealer is actually using, and recalibrate if the dealer changes balls. (There aren't that many types of roulette balls out there, so you will quickly get a feel for differences in their behavior if you get to play each type regularly enough, or can experiment with them at home.)
Scott talks about dealers switching balls mid-shift as a predictor countermeasure, but we never saw that. The few times we observed a dealer switching balls mid-shift, it seemed to be related to the dealer trying to hit a number that a lot of players had bet on. [Note: if a dealer is changing balls mid-shift to try to hit a number, it's a sure sign that the dealer is incapable of hitting anywhere near the desired number except through sheer luck.]
Sometimes, after a few spins, a new dealer who has just come to the table will select a different ball. The typical motivation seems to be dissatisfaction with the number of spins the dealer is getting from the original ball. But again, we seldom saw dealers change balls in the middle of a stint at a roulette table. That doesn't mean Scott didn't see it back at the time of his article. It just means we haven't seen it except on rare occasions in the past ten or so years in Las Vegas.
Now let's look at the deflectors and their effect on the roll.
How to Win at Roulette: Understanding the Deflectors
As Thorp observed in The Mathematics of Gambling, the greatest factor in making a ball's roll inconsistent from spin to spin is the obstacle course of deflectors (also called diamonds, or pins) that it must pass through on the apron before entering the rotor. Modern roulette wheels typically come with four vertical deflectors spaced evenly around the wheel, and four horizontal deflectors spaced evenly around the wheel between the vertical deflectors.
Both the vertical and horizontal deflectors are usually significantly thicker in the middle than at the top and bottom of each pin. If they're not significantly thicker, they are significantly wider. (On modern wheels, they're usually both.)
If a ball hits the thick middle part of a vertical deflector after leaving the track, it will dissipate most of its remaining energy in that hit. A ball that hits the middle of a vertical deflector typically rolls very few numbers once it reaches the rotor. The exact number of pockets it rolls will depend on the exact type of deflector and where exactly the ball hit the deflector. The roll may also be influenced slightly by other factors, like tilt.
If a ball makes its way across the apron without hitting any of the deflectors, it dissipates none of its energy on this kind of hit, so it takes a lot longer roll to use up all its remaining energy.
A ball that hits the bottom or top tip (the thin or narrow portion) of a vertical deflector dissipates a portion of its remaining energy in that hit, but not as much as if it hit the middle. It will roll further than a ball that hit the middle of the deflector, but not as far as a ball that hits no deflector at all.
Hits on horizontal deflectors use up energy in a way similar to hits on vertical deflectors.
As Thorp noted correctly about deflectors (The Mathematics of Gambling, p. 53): 'On average, perhaps half the time these have a significant effect on the ball.'
How the Deflectors Steer the Roulette Ball
How To Make The Casino Wheel Spin Slowly
I've just described a variety of ways that deflectors affect ball roll. You might think, because of this variety induced by the deflectors, that it would be impossible to get an edge predicting which number will hit, since it will depend on whether and how the ball hits the deflectors on the way to the rotor.
But in fact, the ways the ball interacts with deflectors tend to cluster in useful ways on any particular wheel. (In other words, you tend to see more of one or two types of hits or non-hits than any of the other theoretically possible types.)
On some wheels, the deflectors even perform a useful type of steering. For example, on one wheel in Las Vegas a ball will either hit the center of a vertical deflector and drop dead into the pocket directly below, or lightly cross the tip of that deflector, start moving down, and then curve back up before falling, so that the ball spends a lot more time on the apron that you'd normally see after a ball leaves the track. This allows more numbers to pass by below, especially on today's fast-spun wheels. The ball will then roll enough spaces to bring it back to the number that it would have fallen into if it had hit the middle of the deflector.
On another wheel, it either hits the middle of a vertical and rolls 5-6 numbers, or it hits the tip of the same deflector then rolls 24-25 numbers from the number that was under the deflector at the time of the hit. That is a terrific wheel to play if you broaden your concept of a 'sector' for betting purposes. (I will discuss optimal betting in Part II of this article.)
So, if the casinos all run out and change the deflectors and deflector positions on their wheels, will that kill the chance for advantage play? No, it'll just change how the ball is steered on any given wheel, and thus change the optimal numbers to bet.
How to Win at Roulette: More on the Drop
There are very few roulette wheels out there in real life casinos that don't have some kind of a drop (when I say 'very few' I mean I've never seen such a wheel, but I presume there could be a few out there somewhere in this wide world). By 'drop' I mean a place where the ball tends to fall off the track at a higher rate than would be expected if the wheel were perfectly level and perfectly machined.
A typical roulette wheel has a dominant drop point with a secondary drop point that kicks in if the ball gets past the dominant drop point. If you look at enough wheels in enough casinos, you'll see all kinds of drops and secondary drops—even drops where the dominant drop point is at, say, twelve o'clock, and the secondary drop point is at six o'clock. But usually they're within a short distance of each other.
The reason you need a drop to predict roulette is because it's so difficult to estimate visually, even when you know the ball is roughly three spins from dropping off the track, exactly how much energy the ball has left in it. In other words, you might know from the look of the ball that it has slightly more than three spins to go but less than four, but on a perfect wheel it would be difficult to tell whether the ball had exactly 3 1/4 spins to go from your read point, 3 1/3 spins, 3 ½ spins, etc. If you don't know where the ball's final roll across the apron and numbered pockets is going to start, you can't predict where it will end.
A wheel with a drop takes care of that problem pretty well. The drop itself uses up the excess energy that might have caused a ball with borderline energy to go past the drop point if the wheel was perfect. A good drop gives you a ball falling off the track at a consistent place or places with a consistent amount of energy to spend on the roll to its landing place.
Wood warps and cracks, players and dealers lean on tables, floors are never perfectly level in the first place. In all the years we've spent dealing with roulette, we never saw any maintenance guy out there checking a wheel with a level.
A drop usually seems caused by a tilt in the surface on which the wheel is resting. The way you can tell that the drop is caused by tilt rather than an imperfection in the track of the roulette wheel is through the effect of the drop on the behavior of the ball.
When a drop is caused by a tilt, the ball tends to speed up and go further than you'd expect after the dominant drop point if it makes it past that point near the end of its spin. There will also be a predictable number of balls that don't quite make it up the hill to the drop-off point. You'll see virtually no balls fall off the track at the low point of a tilted wheel, and you'll see no really weird ball behavior before hitting the frets.
By contrast, if you deliberately create an imperfection in the track of a roulette wheel, and a ball makes it past that imperfection near the end of its spin, you'll see very strange ball behavior after the track fall-off point. In Las Vegas right now, at a popular Strip casino, the ball leaves the track reliably at the same point every spin but will often swoop almost back up to the track in a very drunken-looking way before proceeding across the apron. We know this drop is due to an imperfection in the wheel because we've seen the drop rotate around the wheel—we've played a six o'clock drop on this wheel and a nine o'clock drop, etc., and in all cases the weird drunken ball behavior was in play.
Wheels that develop a drop tend to keep a drop. There has been some debate about this in the literature. In his article, 'The Myth of Roulette Dealer Signature and ‘Sector Shooting'' (Blackjack Forum Vol. XII No. 2, June 1992—link at upper left) Steve Forte wrote:
'Consider the capricious nature of the wheel. It is an undeniable fact that the characteristics of a specific wheel that may theoretically make it beatable one day can change the next day, or even the next minute, making the game unbeatable. Even the same dealer, same equipment and similar measurements of ball and rotor speed will yield different results at different times.
I proved this phenomenon to myself some years back after spending 17 days on the road scouting, evaluating conditions, and recording data on hundreds of wheels.'
Laurance Scott states the same opinion in 'Nevada Roulette.' He cautions: '…there are inherent factors of the game that cause wheels to phase in and out of predictability,' and, 'I have scouted over 300 wheels and only a handful exhibit consistent behavior day after day.'
I recognize the variability that Forte and Scott are talking about, but in my view it's due to a very consistent factor in the game that requires a different approach to getting an edge than the traditionalists use.
Basically, in my experience most wheels that develop a drop keep the same drop for years, even decades, and thus retain the same predictability for years or decades. What varies, and causes the wheel to seem unpredictable to traditional wheel predictors, is the way the ball drops at the drop—in other words, the style of ball behavior at the drop point. But all drops produce this variability, and the proper response to it is to expect it at all times, on every wheel, and alter your betting pattern from the traditional approach to take account of it.
Let's say you find a wheel where the ball drops off between 12 o'clock and 1 o'clock 100% of the time. But it hits the vertical deflector at the start of the drop-off area roughly 1/3 of the time. It hits a tip of that same vertical deflector roughly 1/3 of the time. And it sails past that deflector to drop without any contact with a deflector roughly 1/3 of the time. Can you play that wheel or not? (Answer: I haven't given you enough information to judge.)
When you first walk up to that wheel to scout it, you may see the ball hit the middle of that vertical deflector 8 times out of 10, and start exulting in the consistency of the drop. Any wheel that was dropping at the same point, with the same deflector hit each time, 8 out of every 10 spins, would be a wheel you could murder.
What Laurance Scott and Steve Forte seem to be saying is that you actually require a wheel like that to win at roulette. If this were true, the only predictable wheels would be ones with a drop where it almost always—a high percentage of the time—either smacks the center of a vertical pin, or lightly nicks the top of a pin, or makes it from the track to the rotor without hitting any pins—one or another, without mixing it up.
But even if you come across a wheel that appears to be this consistent, how can you be sure that's the 'real' drop style instead of just normal variance in the drop style, like the normal variance in heads and tails results when flipping a coin? The answer is you can't without watching the wheel so long that your eyes start crossing and you're no longer fit to play.
People tell me they've actually played such wheels, and they really seem to mean it. But in over 10 years of learning roulette, scouting roulette, staring at roulette, cursing roulette, and playing roulette, we've never found a wheel consistent enough to meet thisdefinition, and we've never been able to get a wheel as consistent as this at home either, no matter what we've done to a wheel.
Instead, what we've seen in real life are wheels that give you 8 drops out of 10 with a consistent style (or even 16 out of 20), then 2 out of the next 10 with that style, then 6 out of 10, then 1 out of 10, then 4 out of 10.
If you're flipping coins, you can get streaks of heads or tails, and if a wheel has a drop, you'll get streaks of one style of drop at the drop-off, but sooner or later, and usually sooner, it will revert to its normal overall ratio of drop styles, which will be much more variable. In other words, you can get short-term streaks where you only see the ball hitting the pin or missing the pin a very high percentage of the time, but the streaks are unpredictable in length and don't last at roulette. A typical drop will hit about 50% of the time, as Thorp noted.
The drop is eternal; the variety of ball behaviors at the drop point comes from normal variance in results that has nothing to do with breezes or the weather or oil on the dealer's fingertips or any other external conditions (most of the time). Again, the best way to deal with this problem is usually by revising your idea of optimal betting, which I'll discuss in Part II of this article.
How to Win at Roulette: Introduction to the Rotor Speed Prediction Adjustment
For the purposes of understanding roulette prediction, I suggest you look at the ball and wheel this way:
A particular ball launched onto a particular round roulette wheel track with a given amount of energy means a given length of time for that ball to dissipate enough of that launch energy to fall off the track. For the purposes of roulette prediction, it's the speed of the ball that sets the time.
Now, instead of thinking of a roulette wheel as a round rotating object, think of the numbers on it stretched out in a straight line along a road, with the numbers repeating, in the same order, every time you come to the end of the set of numbers. Wheel speed sets the amount of road that will pass during the time the ball is on the track.
If a wheel is spinning faster than speed x, more of the road will pass during the ball spin than passed by at speed x. If a wheel is spinning slower than speed x, less of the road will pass during the ball spin than passed by at speed x.
Here's a very simple example of an adjustment for wheel speed estimation. Assume a wheel spinning at a moderate speed (by U.S. standards today), moving from right to left. You take a read of the numbered pocket below the drop-off point three spins before the ball will fall off the track there. The read is 13. The wheel goes through 2 ½ revolutions, i.e. 97 numbers pass the drop-off point during the time until the ball drops off the track. The ball thus drops off the track at 14. It hits no pins, rolls 12 numbers (this is a deep pocket wheel), and comes to rest on 34.
For the next trial, you again see a slowly spinning wheel, moving from right to left, but you judge that the dealer has spun the wheel 10% faster than on the first trial. (I'll get into how you make that judgment in Part II). So you judge that roughly 107 numbers will pass the drop-off point from the time you take the read three spins out until drop-off.Your read at the drop-off three spins out is the same—13. If your speed read is correct, the number 5 will be under the ball at the drop. The ball again hits no pins, and rolls exactly 12 numbers. This time it lands on the 25. From our starting rotor speed, a 10% increase in rotor speed, with all other conditions the same, has moved the landing point back 10 numbers.
Obviously when a wheel is moving from right to left, a move 'backwards' is further to the right. When a wheel is moving from left to right, a move 'backwards' is further to the left.
Now you're calibrated. If the wheel speed had been increased 20% from the speed of the first spin, you'd expect roughly 116 numbers to pass while the ball remained on the track. That means, assuming once again no deflector hit, and the same 12-number roll, that we'd expect the ball landing point to move back roughly 19 numbers (116-97 = 19).
This is a greatly simplified explanation of how you adjust for wheel speed. It's a very common-sense adjustment. If a wheel is going faster on one spin than another, you'd expect more of the wheel to pass during the time until drop off. So you'd expect the ball to land further back from the read number than you'd expect on a slower wheel.
Because a wheel is round, it should be obvious that at some point an increase in speed takes you back to your read point. A slightly greater increase in speed than that would take you to your read point and then slightly past it.
I'll give tips in Part II of this article on how to learn to make speed adjustments.
One skill you might get working on in the meantime is developing your ability to pick up speed clues. It's relatively easy for your eye to pick out the single-digit numbers on the rotor above the pockets—even on a wheel that's spinning very fast, the narrower single-digit numbers stand out amidst the wider double-digit numbers.
How To Make The Casino Wheel Spin Slow Motion
In fact, if you relax your eyes and get used to the effect of the single-digit numbers passing your field of vision, you'll find they tend to create a pulsing effect. At fast speeds, you may have to train yourself to pick up the beat of every second single-digit number. What you want to start to do is hear the rhythm of that pulse in your head as you watch a wheel.
You might try clicking your tongue against your teeth to give you the sound (don't let the dealer or pit boss hear you). The sound of that pulse is how you will train yourself to read wheel speed. Wear a watch with a second hand and start paying attention to the number of clicks you get in some fixed number of seconds so that you start to get a feel for the look and sound of a wheel that's 10% or 20% slower or faster.
I've already mentioned that the speed adjustment example above is greatly simplified. What adds a lot to the thrills and chills factor of predicting roulette is that other conditions will usually not remain constant. In our example, the ball never hit a deflector and always rolled twelve pockets. In real life, you'll never find such consistency. We'll talk more about that in Part II of this article.
First, let's look at the problems involved in trying to use the traditional method in today's game.
The Effect of Modern Low Profile Roulette Wheels on Results from the Traditional Method
One of the things many of the traditionalists seem to agree on is that modern roulette wheels tend to give truly random results. For example, in his article 'Nevada Roulette,' Scott says:
'Some casinos run a fair game. They use modern wheels which tend to yield truly random results. However, other casinos still use older style equipment which is quite beatable. Why would a casino use beatable equipment when modern non-beatable equipment is available?'
Scarne agrees. In his 1978 revision of his 1961 book, Scarne's Guide to Casino Gambling, he states emphatically that no roulette dealer can section shoot with accuracy on modern roulette wheels because the wheels randomize results too much.
But that's not what we saw. I can tell you definitively that modern wheels are every bit as beatable as the old-fashioned deep pocket wheels, and may even be more beatable.
It's true that old-fashioned deep pocket wheels tend to grab a ball earlier in its roll. The rest of the ball's energy seems to get dissipated vibrating in that deep pocket.
But that doesn't make lower profile wheels any less beatable. In fact, because lower frets interfere less with the natural roll of the ball in dissipating the last of its energy, results are actually more consistent with a low profile wheel than with a deep pocket wheel.
I believe what makes results on modern wheels appear more random to traditional roulette predictors is
- the increased speed at which dealers can spin modern wheels without causing the ball to fly out of the wheel tends to make rotor speed adjustments more difficult for predictors and prediction mistakes more likely, and
- betting a 9-consecutive-number sector on a modern wheel is, over the long run, not only not optimal—it's a losing proposition.
It's relatively easy to make adjustments for rotor speed when a dealer keeps that speed in a fairly narrow range.And dealers tend to be more consistent with their rotor speeds on old-fashioned, deep pocket wheels than they are on the low-profile modern wheels because the deep-pocket wheels tend to launch the ball out of the wheel at far lower speeds.
In other words, on a deep pocket wheel, a dealer quickly learns to keep wheel speed moderate to avoid spending half her shift chasing down balls launched into the aisles.
There is a point on modern low-profile wheels where even the lowest frets will hurl the ball out of the wheel and into the faces of passersby, but you don't get to that point until the wheel is much faster. So on modern low-profile wheels, dealers can and do use a much wider range of wheel speeds (mostly on the faster end of the range), and they tend to be less consistent in rotor speed from spin to spin.
It's simply harder to judge the exact speed of a wheel that's gone from really fast to supersonic than it is to judge the exact speed of a wheel that's gone from slow to slightly less slow.
Also, on the modern low-profile wheels, on which the final roll of a ball is much smoother than on a deep-pocket wheel, the ball is much more likely to roll the exact number of pockets dictated by the remaining energy when it falls off the track. It's almost impossible for the ball to get stuck in a pocket in the traditional sector and just vibrate the remainder of its energy away.
As a result, what we've found on modern low-profile wheels is that there's not only no edge on most of the numbers in a traditional 9-number betting sector, there's actually a strong house edge on most of those numbers—maybe not the full house edge of 5.26%, but not a player edge, and enough of a house edge to kill the player edge overall. Reading the wheel accurately, but betting it wrong, leads to dismal results.
On deep pocket wheels, by contrast, which are spun at much slower speeds, all of the numbers in the traditional sector had enough of an edge to make them worth betting. We believe the difference is due to the effect of errors in speed adjustments. On a slow wheel, if you're off a little, the ball will often still land in the traditional betting sector. On a fast wheel, being off a little takes the ball much farther away. Small mistakes kill you on fast wheels.
More Problems with the Traditional Method of Winning at Roulette
Laurance Scott used to put out a video with his system; maybe a similar dvd is included in the current package. It was essentially a video of a slowly spinning deep-pocket roulette wheel, filmed from straight overhead. My wife learned to predict that wheel within about 15 minutes, on a very small TV, from the opposite side of the room, where she was glancing up from reading in bed.
Great. She had a terrific edge on bedroom video roulette. When we looked for that game in a real casino, we never found it.
Another roulette predictor who demonstrated his abilities to us did so from a balcony overlooking a deep-pocket roulette wheel—and I agree, this is an ideal way to do it, with no heat from the dealer or wave-offs to worry about—except that you can't make any money doing imaginary betting from a balcony.
We asked this player to actually scout low-profile wheels with us and let us sit in on a play. I told him we couldn't beat them with his method and we wanted him to demonstrate. He never came through, even though he was soliciting my help in selling very expensive roulette prediction lessons to other pros. In other words, there was a potentially huge financial incentive for him to demonstrate his prowess on an actual modern game, but he's never made himself available.
In addition to finding no edge on most of the numbers in a traditional 9-number betting sector on modern wheels, and the overwhelming odds against finding wheels you can beat with the traditional method, here are the problems we ran into trying to use traditional roulette prediction methods on modern live games.
First, dealers wave you off long before you can bet if you are trying to predict from three to four spins out.
The average roulette dealer in Vegas or Atlantic City or Mississippi or anywhere else in the U.S. seems to see his job as spinning the wheel very fast, giving the ball as long a ride as possible, and then waving the players off as early as he can get away with so that he can stare off into space and not deal with the whole thing for as long as possible. By the time the ball is actually coming to rest in a pocket, everyone has been sitting at the table motionless for so long they've forgotten they're playing a game.
I've heard it's different in Europe, where there's a longstanding roulette culture, where the tables are surrounded by crowds covering the felt until the last possible moment, and everyone's eyes follow the ball like a greyhound at the races, openly trying to calculate where it will land. But we were told it was different in Canada too, and we found conditions in Vancouver ten years ago about the same as in the U.S.
Often the only way to keep the dealer from waving you off5, 6, 7, 8 or 9 spins out is to keep betting yourself—chunking out the chips continuously throughout the spin until you finally have your read and have bet it—because as soon as you pause at all, that wave-off will surely come. On several occasions we had dealers remove our bets from the layout and even fling chips back at us if we were racing to bet against the wave-off.
Even when we ran into a tolerant dealer, perhaps on a crowded table, who actually allowed us to bet three to four spins out, she wouldn't continue to be so tolerant if we started to win. The wave-offs just got earlier and earlier.
And chunking out camo chips at the house edge continuously throughout a spin, when your read won't come until just a few spins out, is very expensive. It would take an enormous spread between your bet size on your camo chip chunking and the bet you place on your predicted numbers to cover that expense.
Second, it's not like dealers haven't seen these methods before. We know both system sellers who say they actually play this way, and professional players who try to play this way. Plus we have run into perhaps a dozen more players at the tables who appear to be students of the system sellers.
Add to that all the crazed European players who have a little roulette predicting skill from sheer time at the tables (though not enough to get an overall edge), and what you get is a bunch of roulette dealers and pit bosses who are surprisingly quick to get suspicious. There's a truly remarkable amount of heat at roulette considering how few people out there are actually capable of beating the game.
If you try to bet a 9-sector number after taking a read three spins out, you look so obviously like one of the people running around trying to do this that you tend to run into countermeasures even if you're not good enough to beat the wheel.
How to Win at Roulette: The Master
We found the traditional method of winning at roulette so painful, tedious, awkward and ill-paying that we would have given up on the game if we hadn't started running into The Master at the roulette tables. The Master never bet late in the spin. The Master never looked like he was watching the wheel. The Master never bet a sector of nine consecutive numbers, or random numbers chosen from within such a sector either.
The Master also never got heat from the dealer or pit, and we never saw him get up from the table without a nice win. And The Master did all this without the use of a computer.
In Part II of this article, I'll talk more about The Master, and go into more detail about how to win at modern roulette. ♠
