Roulette Roulette advantage play & cheating part 2

[cyph]

stack formations can be pushed to the agent with complete cover from all sides (Figsy 9 and 10)/J|

ilpn another version, the'dealer loads up (mucking regular checks with wheel checks) as many $5 and $25 checks into a dirty stack as he can during the mucking process. The stack is moved - to the bank, followed by other stacks. The goalpvto completely hide the valued checks within the bank of wheel checks. The. agent quickly Mows up with a buy-in. The dealer moves the checks out into a seven-stack formation, with the dirty stack nealdy hidden in the center, and pushes it off to the agent.

Mucking the dirty stack doesn't;have to be an open maneuver,,The dealer can muck the $25 checks first, and fake the motion of putting the greens into the bankroll before mucking the normal color.;|J

Here's a scam that went|undetected for , many years in a prominent Las Vegas Strip property. ; In this variation, the 'doorman steals from a dirty stack. With heavy play on the outside, the dealer's partner (the floorman) would stay clqse^nd occasionally help muck, which isn't uncommon. The dealer would sweep the layout and intentionally muCk one high denomination check to the bottom of the dirty stack, making it twenty-one checks high. The stack was pushed back into the apron area, but not flush with the bankroll. This was a signal for the floorman to come to the game. "He would move the top twenty checks into the bankroll as the other hand rested on the apron with open fingers, hiding the high denominnation check under his hand. This was then palmed as the boss walked away. Since the boss only appeared to handle the nonvalue roulette checks, and never the high denomination. ^checks, there was never any suspicion.
Overpay

Here's a classic overpay scam practiced for many years. One high roller bets many numbers, splits, and corners. Two or more members of the crew bet small, but seem to follow his money. After he bets, the small players come in with bets on top of his bets and the surrounding numbers. If the takeoff man hits a straight up number, someone turns the boss. The dealer then, apparendy, totals the payout. He cleanly sprays the high denominaton checks for surveillance to see, and pulls the appropriate stacks

 

[cyph]

img

 

[cyph]

needed to make the payoff (an overpay). He may even call out the payout, and turn slightly as if trying to find the boss for approval. Then, acting as if the payoff was approved—its difficult for the sky to see an approval from a boss standing fifteen feet away—the dealer pushes the payoff to the agent. When the smoke clears, surveillance may know exactly what the payout was, but it can be difficult to ascertain exactly how much was initially bet.

Immediately following the payoff, the bets are picked up, moved around, added to the players checks, and so on. In short, don't expect the “winning bet to be there very long, should you decide to take a closer look.

Conversion Scams

Conversion scams all follow the same premise; the cheater's;buy-in occurs with one dealer and supervisor, and the cash out with another dealer and supervisor. The cheater is attempting to cash out and color-up for an inflated valuer Ij

A common variation is for the cheater to buy in for purple, for example, during a busy game, and rathole some of his checks before cashing out. He later returns to the game, requests the same color, but buys in for a higher value per check. The hidden checks are now secretly added back to his ' playing checks, and here comes the color-up.

Or, one agent buys in for nickels and the value is properly marked..After a cash out, the dealer conveniently forgets to remove the lammer sitting on the fim.ofthe wheel. When a new boss arrives, another agent buys in for dollars, butasks^forthe same color just played by his partner. It now appears that his check value is higher, with each check now worth $5. The agent can stay and play, or opt for the sure thing by cashing out with the new supervisor.

Its just as likely that these scams could occur without the dealer's cooperation (obviously much 4| more dangerous), but either way, it's a perfect.example ofa’no-skillxno-preparation, brainless attempt at cheating the casino. Since many players leave the game with the roulette checks, the inventory is not always a constant. This opens the door for these scams.

Other Methods

The list continues, and many other methods exist. The most common are directed at paying the outside bets. The mechanics are similar to those explained in the blackjack and craps chapters. Sizing in high, picking up a losing bet to cap a winning bet, moving a losing column bet over to a winning column bet, and so on.

A dealer may palm a few checks during the mucking process and cap a check change, make it look like a check change when there was none, or appear to take checks from a player and bet them according to the players instructions, resulting in five free bets. The list goes on and on.

 

[cyph]

Orihe dealer may just be.fflilt^f^gr^ran^;int^i{m^ly,Wh&^agr^s^6 look the other wayji| as an agent or crew perpetrate any of the desctihed outside scimis/pf.past'postiri^^^Set switCT^^md such?SB$ ■BylFhad to pick^thVmost gutsy, blatant-scarrfof all time, I would have to^gowitmafnotspuv^ repeat-lof tli^l^^winmngmumbef. ^j^rm locked up the game^mter?afbig straight up numbers hit/ithe boss1-was turned, there wasMo1|mnfand the marker ju^Vtayed?omthe last winning The crey^^^^m^land Velebrated/^^if^itting«tne number on a repeat,* and the payoffs proceeaeaM normally ... alg^^^^n^Yes, "thisPctually;happened! ,rv-

GAFFING THE WHEEL

^iqj?gaff^a wheHfisfto fix it,‘or altet the trheness^faflegitimate wheebso/that certain pocketsj|>r.l| |¥e^ions'4r;^^pre’likely to catch th^ hounciii^ hall. ;

Early Wheels

A?lbng;time ;ago, 'casihos;'o1:casionally'uged'gaffed whbels-tdpnsure a healthy percer^^iiprh^earlMj wheels were ’drop ins', being dropped into a large hole cut into the table,with onlyfan inch dr two of J the wheel head protruding, above the layodt^Withvthis|jeMpdesigh^ the wheel was easier to gaff, and ftthfe;mdsf infamous of thfe,gaffedwhetylfvlpis the juice/wheefS Behind the'entire ball track was a series ! ofrelectromagnets, and the baUToused a small ste^coVel^If a player, bet on the single-zero/the dealer would simply wairfot'the ball ■to" pass the; single/zero, turh/thWjuice on—which literally stopped the ball in its track—and immediately turned the juice'off, allowing the ball to fell. This would eUminatef |any»,chance for the ball to work its way; around the wheelfback. to the single zero. These wheels couldl ■be activated by a hidden switch under the edge of the table,*pr by the dealer leaning on the apron. OneJ jpfithe most sophisticatedemethods had a playeriworkingffot|the house sitting next to the wheel. He controlled the juice from inside his pocket iwithVa'remote control device similar to a small garage door opener. Now you couldn't patcflthe dealer making a suspicious move sincefhe never had to look into the wheel until after the ball had landed.

Then there was tne pin wheel’, also called a ’kickoff wheel’. Spring loaded pins were built into the head just behind the ball track closest to the players, and completely invisible from their vantage” point. If the dealer wanted to knock the ball off the track, the system was activated, and a smaUfpim would pop out, effectively deflecting the ball.

With these systems a ’dead ball’ was used to make the balls action more sure. This was a ball that exhibited weak bounce characteristics.

Eventually, the notoriety of such gaffs led to suspicion. To put the public at ease,'the wheel was] lifted. It now sits on .top of the table, raised and cleanly resting on three ;a<^ustable/feet, .Th^ entire wheel head is>visible to the, public.

 

[cyph]

Fixing the Wheel

Any employee with access 'to* the wheel can gaff it in a variety of ways/ making it more likely for the ball to fall into, or stay out of, a number, a sector, or even half the wheel. Perhaps a janitor on the graveyard shift during his routine cleaning, or an engineer with the duty of maintaining the wheels integrity, or if extra wheels are in storage as part of a rotation, they can be gaffed by management before they are moved to the pit.

To gaff the older style wheel, most methods targetted the pocket bottoms (the pads) and pocket walls (the frets)rem'ember hearing about cheaters taking a hair dryer and applying heat to 'certain sectors, trying to create a little give by attempting to soften the glue that holds these pads in place." Pocket pads with different compositions could be substituted and were designed to either deaden the ball’s bounce, or make it much more likely that the pocket will repel the ball (called ’bouncers'). In a few of the earlier European models, felt pocket pads were used and were easily [compromised by using more than one layer, or a thicker material.

As for the pocket walls, they have been changed for thicker walls, effectively reducing the available space in one or more pockets. They have been raised, making them more likely to act as a stop, whereas a lower wall would make it easier for the ball to pass over. With pliers and a thick cloth for protection to keep from scratching the metal, the pockets walls could be squeezed together, narrowing the entry space into the pocket and widening the space of the adjacent pockets.

In Beating the Wheel (Russell Barnhart, 1985), we "find the tale of a cheater by the name of Valdermier Grenic. His method of gaffing the wheel was particularly ingenious. He substituted the screws that held the pocket walls to the wheel head for screws with heads identical to the manufacturer s screws, but with a wider thread design. This had the effect of holding the pocket walls less firmly. If the ball hit a less secure pocket wall, there was less bounce, and more of a deadening action on the ball, making it more likely to fall into the walls adjacent pockets by following the path of least resistance. If the ball hit a secured pocket wall, it bounced more, as one would expect. .

Various gaffs were once sold through the crooked gambling supply houses. For example, the H.C. Evans Company, out of Chicago, offered 'wire bouncers', 'rubber bouncers', 'pasters', and 'cork plugs'. These gaffs were all designed to keep the ball in or out of certain pockets. They had this to say about their partitions, what we now call frets, or simply the pocket walls.

M^Partitions, when used in sets of eighteen will give the outside a margin of ten percent or better Our improved cushion partitions are the most perfect ever produced. They are made ofthesame metal and are the same thickness as regulation partitions... they are a perfect match .in everygf respect and impossible to detect. We also offer the old lead style partitions for operators ,whq^ prefer them ... in ordering, we need to know the make of wheel iti^wfichAo be usedy^..

 

[cyph]

'."In the a<ivertisement; |setsfof eighteen” is1 a^ref^ence ' ta a" higher- fret in everylothef positiotB : <3fhe ball was spun both eloclbvise?amf^ni5t^l^l^is^P^)riej direction, the higher frets woulallf! as stops to i^ypf the pocket in the othefrv^/eyety other pocket was the favoritejH

Modern Wheel Gaffs

?#Most modern^heelsifeature a solid number ring, making most of the better known methodsImt^H Yet; gwJropath- a^olid*.ririg, ‘consider how many plastic coatings, resins, silicons, and crtne^finisn^ ’could'b^adedftp the top, bottom, or sides of a pocket to'impact,the game inexactly the same^was j example, Varathan^i's;iiplutiip>tpajing thatvyo^^mbuild up, one layer at a tinted What^would be the effect of coating the redpockets qnce,;;andjthe' black pockets numerous times?! Ilffiap&tj and thickness would be added to a pocket wall a^qst’iny^ibly^jil^ne can cortceiye numerous! . ways, where the right product'couldrsubtly- alter^the true,probabilities^

The low profile cone is ubiquifqusa With .this design, the ball spends a lot more time bounciiig| ^^ahd|rollingon the cone. With more^confact comes the possibility; of gaffing the cone to inlluen£e^ ball's outcome. The^m&bbuld}be waxed in half only,(toicrea£^ slickeji

^surface as compared^the other side. ^mSw^uldj|Cutdown onthe balls, resistance, making-it morel likely-to, slide to the other half.

Minuscule grooves cpuld.be.worked into the cone, "creating a bias for the ball to get caught in and directed to a-specific pocket. When Tsay^yroyked; Trefeq.to numerous chemical compound^ that contain solvents that are strong enough to, bre^down any protective surface, and this includes! lacquer, varnish, ure|hahe, polyurethane, or, epoxy One can; never lose sight of the fact that the roulette wheel is just apiece of furniture, and they'rqrefinished; all the time. If there is a desire to gaff any part of the wheel in this fashion, it isn’t that difficult. Not long ago, this gaff was detected. The grooves! were found directly on top ofthesectorlines on the cone. When the ball slowed down, it would occasionally get caught in the grooves and slide predictably into the connected pocket.; *

If the bowls inner lining wraps.(^to|thb;lower ball track (you’llsee a visible metal strip), or if a metal strip is used to hold the number ringsin;place,?either strip!cb)uldibe pried up enough to creates a tripping effect. Since the ball passes oyer ope or the other before reaching the pockets/;thish|o® force the balls trajectory upward, and making it unlikely to land in the connected pocket since jta gmoving away.

Its also possible, I’m told, to gaflF the spindle with abrasives, or grinding compounds. The goal! would be to wear but the bearing systems, creating a high^lpy ;spitt of the rotor,, which^v^s favoirsl those numbers pn the down side, :br- below the horizontal plane.

There have even been attempts by both the dealer and cheater to gaff the wheel during play.^j In one case, the dealer was attempting|tbi spread some cola ^fos^a predeterminedsector.of the number ring as he gave the rotor one last push. Apparently the goal was to create a sticky surface

 

[cyph]

across a few connecting numbers in an attempt to grab the ballpind. slow it down. I suppose that any wax, lotion, or paste that was easyito transfer could be used. ■

In another case,tone crew createu a disturbance and dumped’some metal filings into one sector of the wheel. The filings looked much like fine dust, and given the steady rotation'of the wheel, the foreign matter wastio.t^easylto see. Apparently, the goal was to manipulate the wheel from the outside, r possibly <foomr«underneayi:.the tableiyyith!some form of magnetic intervention/.£Fhe disturbance alerted surveillance to take a clo5je^.dppk^iand the filings were later discovered. What's interesting about the attempt, is that the ball would have remained perfectly legitimate, the; first place one would look if the play was suspected. ^

i^fThe.idea of gaffing the wheefin some way by hiding a metal element, one which would allow, y manipulation from a magnetic source, is interesting. After you read the section on visual prediction, you will immediately See why any system, magnetic or otherwise, with or without collusion, that equips cheaters^ith the ability to slow the wheel head down when desired, could prove to be devastating. |jSheaters could win at will—and for serious money.

^Finally, there’s a wonderful story about a cheater drilling a small hole in the wheel head during play. He was trying to recreate the infamous pin wheel, but from the outside. He stood close to the wheel holding a box (this was long before we had the protective; shields), and the box contained, a portable1 drill. Just; as he finished drilling the hole, [nej was caught. When the Gaming Control was contacted, they didn’t believe the casino executives, tpjiyhichjOne replied, "Okay, fine, but what would you like me to do with this hole in myCwhed^M

If we look at the manufacturer's brochures' for any wheel, the pitch is always the same: nontamper ^security seals, cheat proof, and advertised : tolerances;of a few one-thousandths of an inch on every component. No argument herei (The wheel is an incredibly* safe, precise gaming device, but to assume that it is made in a fashion immune to itamperingds a mistake. When one approaches the problem with larcenous intentions, the possibilities are endless,^ 3

Gaffed versus Biased

One of the most alarming aspects of the gaffed wheel is that some methods can produce a bias identical to those caused by normal wear and tear. Consider the following scenario. A player always betting the same sector makes a big score. iThe wheel is pulled off the casino floor for inspection. It is discovered that a few of the pocket walls have loosened, or the holding strip to the numbers is warped, or there is a small rise and fall to the rotor that favors half the wheel.; Who’s to say what actually caused the problem? Was it normal wear and tear? Was it tampering?

There have been more than a few significant scores throughout the world that were written off as a normal biased wheel play, when, in fact, collusion coupled with some very sophisticated methods of gaffing the wheel went undetected.

 

[cyph]

BIASED WHEELS

Mechanical; imperfeGtions^havefaIw^^p^^^^^^^p^^t^^^^^/rolilette1,<Itjstaftstwith*d^B

theory roulette wheels are ]S

. mad^frgM^^y^lumi^u^^^^^pLffdjro]fnpjM^^^t^ials^L^Mtha^all|^aterial^aVe^subiect^^^| kpxpansionV confeaction, ^arp^mgpand weakening when^exposeflltosthe eletrfent^Further^B^ justly plpP^^^^^^^^^^erpetual spin^hesslanimi^^flthe’ballufflO'the tracldprior ttf spinning]* and the fj constant ;kounqe^o^|^b^^Bffi^^yheet £■ v^rasn^|n^^ The wheel i^olwipusljS^byectito muffflH ^^^s^^^ii®HmlS^om^^^^utprise t^^n^cnamcaMSa^s likely

^-^Given tj^r^chanical^^ectspKtlife game,^^»go|o!cUDet that biased wheeltpIawKQef baclfftpM ' the;.gmiesiearl^^a?i2?^?'IIt^puld'’have made perf^^wi^fdr the pla^^ro So^l^^mechaniiaSB | Dia^aVaroossiols^almes^Wher^layers' hegan’tcf re^rd the 'winning numbers in,!an effort to realign ^^cl^mcalfpfofits)They:be^me^lafgyjiBPi wheel dockers', and probably are the earliest known advantage^!

iifDocumented gfioresj from?this|smt^y1start around &873^wnen>an engineerbwthe name of Pm8m^afeg^tw6^$32^0QQ^r^j&T^^Sa^Ni&tferQtfe^iSx^^ffl^layifes,!and teaml|were to follow: 1 Dr.-'Racihafdlfa^^^MontekCaadol^hejItdian's^Mdnte Carlo), The Argentineans ■( Argentxi^Thef* ’ Jones Boys (Reno)^AlbertiHibbs5$na|Eraw Walfoixh (Las Vegas)? and Billy Walters (Las Vegas), to 1 )$name^fe^ThMargest^c6re'known isxreditwl.to'the Billy Walters Team, wirihing'$3;8Q0,000 in il Las Vegas and Atlantic City m the late’L980s^p

Bias

Bias may exist due to warpage^ cracks, rotational imbalance, weighting, slope, swelling, warping, air pockets, 1 ror, as with htostrcas'tes; good old-fashioned wear and tear; These factors may lead to imperfections in the number strip, seams,''pocket walls; potket pads, the finish|(especially on the cone), the balls, and so on.

Whatever forte* is behind the manifestation of a bias, it appears to be a very slow and changing 3 process. Biases can remain'Strong and playable for long periods of time, but they can also lessen, surface in other parts of the wheel, or disappear without warning.

When we talk about biased wheel play, we are generally talking about biases found in the rotor a and:connecting parts. Thistirtcludes-the number strip, pockets, cone, and turret (the center vertical J fixture). Biases to the ball, ball track, apron (the lower ball track)] deflectors, and bowl may be important J to other strategies, but they have little value to the wheel docker. Even if you knew that the ball wpddjM always exit the track at a particular point, you still wouldn’t know where the ball will land. There is an exception to this logic.

First pointed out in The Biased Wheel Handbook (Mark Billings 3c Brent Fredrickson, 1995),yt 2 is possible to find a rotor spinning almost imperceptibly off its horizontal plane combined'ywith baMM

 

[cyph]

track bias to create a superduper bias* If the wheels down side is always positioned under the ball track bias, not only is the ball more likely to fall off at a certain point, its more likely to fall into the wheel on its low side.

It may also be true that the ball, rotor speed, and dealer are important variables. Some wheel dockers have claimed better results with certain ball compositions, certain rotor speeds, and certain dealers. When they collect: data, they also record the sizeofiball, composition, average rotor speed, and any relevant information about the dealer, or dealer's spin characteristics.

Finally, one would assume that some wear and tear is needed before bias can take shape. This is not always the'icase. On more than one occasion, a wheel, right out of the box, has exhibited strong bias. This may be due to the travelling temperatures (expansion and contraction of the wooden parts) as the wheeljis transported to the buyer. Sometimes the wheel is too tight when it first leaves the manufacturer, so some wheels need to be broken in. On one model, the number ring was secured to the'base with five evenly spaced screws. The screws may have been too tight, as the area directly above each screw,Was harder, producing more bounce than other parts of the wheel. The bias never presented a problem as software analysis detected the problem early,

.: An unusual element^oif bias wheel play is that the cause of bidtf is generally unknown. Its invisible, and only rarely can bias be seen or heard. In most cases,1 bias can only be detected by statistical analysis. .:

Collecting the Data

The strategy begins with the collection of dataJ'Teams will record spins for weeks, sometimes montnsg The team may break up into different shifts, with eacl^laying the role of a system player who records the numbers on a tallycard. The data xan|alsa be collected! secretly with the help of recording devices. In Europe and other places around the world, a list of the winning numbers over any time period is available at the players request.

Today, the collection, of data is easy, thanks to thejwinning number displays; rfThese displays typically show the last twenty winning numbers, and sometimes more, thus eliminating the necessity <of being on the property for each spin. For example, a team walks into Caesars Palace and, from the lobby, writes down the last twenty numbers on three different wheels. Then they go to the Mirage where the process is repeated, and then to Bally's, and so on. Forty-five minutes later, they are back in Caesars where the process starts all over again. It's now possible for a team to record thousands of spins on many tables, and in different clubs, over a single weekend. ,'C

Analyzing Data

How one collects the data is relatively unimportant, it's the quantity of data and how that information is analyzed that is paramount.

 

[cyph]

Sample size must be large enough to ensure a reliable statistical evaluation, and as a general $ rule, players are looking to collect. aboutTO,000 spinll At 30 sj5ins an hour; 720 a! day/this can take about two weeks collecting data theold-fashioned way.Thisnumber can be reduced;sighificantly ^ when either a sector breaks off (a^groupof connectingmumbersithat all exhibit bias) or visual confirmatiombfia'bias is made, such’as a rough spot • on:tne'wheel.

How many times does a.number have to.hit over 10,000'spinstopeideerhed playable? It : depends on the team, and ion thelplayer’s criteria for acceptable 'hit frequencies! For example/f in' Getting ffoe'Best;'<^J^(David:S]dahsky, 1982); the authordboks *for a numbers occurring with l/25j ffequehcy^wf|af5>P00’-;spimsa 1/27'frequency over a; 10,000-spin sample. - In The GaWwm

Gambler’ss.Guidejj(Allan Wilson, 1970), a 1/29 frequency for 5,000. spins andja $32 frequency?for| 10,000 spins is]recommended. .

On a straight up bet, a 1/36 frequency indicates a break-even game. For every 36 spins ;you: lose 35 bets and hitone straight up number to win>35 bets. A1/27 frequency is very strong. For every' $%27 spins, you lose 26 bets and hit one straight up number to win 35 bets. This is a net win of 9 bets, '; indicating a' 33% edge (9/27). -

The two most commonly u’seih statistical measures-for detecting wheel bias are standard! deviation and chi-square. Most gamers are intimidated wherf they hear statistical terminology, but fflets-take alobk. There's nothing to it!

Standard Deviation

Standard deviation is^aj statistical measure that provides us'with a range .of what to expect (discussed in chapter -eleven,"Game Protection Mathematics"): When the result falls outside these ranges, we can deem the event to be unlikely to have occurred by chancb, and, therefore, more likely -to be the result of mechanical bias.

There are two kinds of standard deviation. The first is the theoretical deviation, and this is computed before you collect data. The second is the sample standard deviation, and this is computed with the data you collect. For our purposes, we’ll take a look at computing the theoretical standard deviation for a single number, which is the square root ofT/38 x 37/38 x]fcl, where N equals the sample size. We just multiply the probability of hitting a number (1/38) times the probability of missing the number (37/38) times the number of theoretical spins (N), and then take the square root of that product. On 10,000 spins, one standard deviation is 16, and three standard deviations are 48; that|^3 x 16.

The theory behind standard deviation states that we can expect our results to be within,three standard deviations about 99.7% of the time. Over 10,000 spins we can expect a specific number to hit 263 times on average (1/38 X 10,000) plus or minus three standard deviations. This is a range of 215 to 311, three standard deviations on both sides of 263, which is our expected result.*.: &

 

[cyph]

Once our parameters are set;,we;.are ready to collect data and quantify unusuaT results. After 10,000 spins," lets say that only one number falls outside our range/ Its the double zero and it hit 315 times. What does' this mean?

♦The total number of double ?|e|oM)yilhfall outside our range on only 0.3% of all samples (100% — 99?73o)V This means that we can expect the total number of double zeros to hit more than 311 times on 0Jl5%;bf all samples, and less than 215 on 0.15% ;of all samples. Our result should only occur once in 66^simples (W.0015')^motherwcjrds, if you collected 10,000 spins on 667 individual trip's,torily once would you expect the double ze^to hit more than 31T times. Would you now consider the number;to:be biased?|III

Now we get to the most important part of statistical interpretation. An unusual result doesn't ^provide4’indisputable proof of bias, nor does it guarantee that if a bias exists, it will continuefto exist. It just means that a particular sample is unlikely to be the result of chance fluctuation. In our example, hitting more than 311 double zeros over * 10,000 spins is almost a 700 to 1 shot|to have occurred by chance.

This result alone is enough for some players to go ahead and bet this number. Others may; look for a result outside four or five standard deviations as a stronger indicator ofii'unlikely to have occurred by chance".

The professional will often compare the theoretical standard deviation with the sample actual standard deviation, which requires a few more steps to compute.^The two measurements should converge over the long haul.

Chi Square

Chi Square is another statistical’measure often used for evaluating roulette data. Let's simplify the game and consider a roulette wheel with only ten numbers. Well spin 1,000 times and record the results, / The data collected is in the Sample row of the following table. What we would have expected for each number over 1,000 spins is in the Expected row. Looking at the results for the number 1, the bottom row is the result of subtracting the expected result from the actual spin result; for example, 91 - 100 = -9. We then square that number: -9 x -9 = 81. Finally, we divide by the expected result: 81/100 = .81. When we add up all the numbers in'the bottom row, you have chi square for this sample. That is, chi square equals 24.9.
Number 1 2 3 4 5 6 7 8 9 10 Totals
Sample 91 82 79 84 89 113 106 m 115 120; -it,000.
Expected 100 ...100 MWI mm 100 100. 100 100 | itioo;i 100 -Eooo
Chi Square .81 | 3.24 4.41 2.56 MEM 1.69 ,36.:; mmm 2.25 4 24.9

 

[cyph]

What's next? We look up the chi -square number in a statistical table found in-every reference •. 'book,'and it provides us with th® following probabilities; Using "nine degrees<of freedom'far chi'square term which always iridicafce|\(^^^S^nar^^^^al numberfoFfoumom^^e?see'that-there is a*{>%r chance ofexceedingT6.9, a 25^pn^^of:exceedihg^l9lO]p^%cKahcefof exceeding 21|CTaifd a 0!5 m chance ofr23T6^|p

Thus, there is a 1/200 cm^^e of “exceeding 23.6, The results are unlikelyTo havefoomeifrom •' our legitimate .wheel, one with expected frequencies'of 1/10. The same analysis’canbe conductedjpn a regulation wheel with?38/pbckets; only we would use '37 degrees of freedom.

Chi square is commonly used ifriurveillance analysis software programs, testing labs^and by many enthusiastic gamers trying to get a handle on statistically, evaluatingthe?game, but the measure hasfit&h'brtcbmings. This was lucidly pointed outm^dBiakd^^eliJiandbod^^S

„ tt^oesnmeVyoU‘whtch numbers W^lm^^ormo^ttitelliyou whetherWr, not a biased Bpwm thai'jails the test Wsuffciehtly biased for profitable play... chi squaregives'no credit %|j v1 i t jot sectoring. %If there ivereieighi numberspioith'^^^m.n^tneWi square would b^thp sametM ■ - whether th'ey werefdistributed^mdomly‘around'thejrowmoriwere'allcontiguou^^mt

As you can- sei^H^^tildcalfe^luatabns?indicators^ :dlt takes a fair .amount of^common sense tq|interpret,.theinmeaning.^It's;nopalways,!easy or intwtiveSH

There you have it, the basics of using^staridard deviation and chi square, the two most common] ^statistical ‘measures for quantifying tl^a^^iil^^^mfriot the only options for the piro, there are other distributions, some that eveffifook'"atiafcfrcular format of outcomes] but the lo^eP'always jthe Vame.''•First comes the 'statisticml^luationrthen comes the decision to play bW^^/bised;dn the results, and this decision changes with the player or team.

INSPECTION

Now its time to learn the basics of wheel inspection.

Visual

Always start with a visual overview inspection. Observe the table, legs, layout, bowl, ball track, deflectors/canoes, rotor, pockets, frets, cone, and turret. You're looking for any irregularities, obvious wear and tear, or damage. Does the rotor noticeablyjand quickly slow down?fIs there dirt and dust buildup in the,pockets? Are the pocket pads worn? Spin the ,\yheel, and let the light shine off the wheel at different angles. Are there any streaks, grease marks;'or rough spots?

This first test addresses the general condition of the wheel; the other tests are designed to look

 

[cyph]

at each component'more ■ closely.

Detection: k Obvious signs of irregularities,; wear and teat> or damage;!?^

Table and Bowl

With a quality carpenter S'level/checkthe table across the center and ends, as its subject to much abuse from player contact.* Next, place the level across the bov^l and take multiple readings, north/southf? east/west, and sojon. Make thelnecessary adjustments,-as its important for the wheel and table to be Bpjyel before other tests are conducted. %

Detection: Table and wheel imbalance.

Turret

If’Spin the wheel artd'observe the jtutret.n Doesyt remain;rigid and vertical during the spin? Does it wobble? You can verify balance by placing a level omtop of the turret. Take a few measurements with the rotor in different positions.

Detection: Turret/spindle imbalance.

Ball Track

Once the table and wheel are level, it's time to inspect the ball track. This is one of the most important variables in any strategy based on calculation of the spin. The important factors to bear in mind are:

1. It takes very little to alter the ball's exit point.

2. Even with routine maintenance, contaminant buildup is inevitable (smoke, dust, dirt),

3. In regard to being level, there is often no correlation between the table, wheel, and ball track. It's common for the table/wheel to read level and significant ball track bias to still exist.

For a quick test, rub your fingers around the entire ball track and feel for any irregularity. Since most ball track biases are generally too subtle to detect by touch, follow up with about ten to twenty spins and observe the exit points. If the ball exits from the same area of the track repeatedly,, this is a reasonable sign of a ball track bias. With mixed results, continue for another ten to twenty spins, looking for a pattern. If you still get mixed results, the track is probably okay, as biases tend to show up quickly. As a general guideline, the wheel is deemed exploitable when the ball consistently exits the

 

[cyph]

psame half of the wheel Tfitfcmfthe

One can also listen for ball track;fh^^^0th«p^[lWcmtinuousIy; rattles^ or loUgly^sciflatj^^B ” 'throughout the spin, this may indicate a^lessj than smboth^^fad^ighoife the first'few revolutions, as Bat is^^^^^F^l^l^fl^^^en^e-hiitiirallv.when' first rele^a^p

:’’f<'PindU.y,'inspect ^eflow^v ball track surface, and look for inconsiStency*or irregularity^ You want to see and feel a'smooth, consistent surface.

Detection: Ball track bias.

Canoes/Deflectors

After exiting the ball track, the canoes (the diamond, shaped deflectors) are the next obstacle in the balls path. Ensure that all canoes/deflectors are securely fastened. If a deflector is loose, the balls action will deaden upon contact Also ensure thatpiflfcanbes' are properly aligned and are of the same length, width, and shape. A visual-inspection will suffice.

Detection: Canoe/deflector bias.

Number Strip

After the ball exits the lower ball track and strikes or clears a canoe, it reaches the number strip and the metal guide that typically , secures the number strip, depending on the model. Spin the wheel slowly as your fingers Hgftdj^buch'ra^ for indentations, bubbles, warpage, or buildup of

dirt and dust under the strip. Also pay particular attention to the strips seams, if any. Make sure that they are not separating or raising.

Detection: Number strip bias.

Rotor

One of the most important tests ensures that the rotor is spinning evenly on both the horizontal and vertical planes. Place a finger at the point where the number strip almost contacts the bowl. Spin the wheel slowly and feel for the levelness of the spin. If you feel a change in the pressure against your finger or see a discernable rise and fall, then the problem must be addressed. This is especially dangerous when the same section of the wheel is always rotating above the horizontal plane.

As the rotor spins, observe the distance between the rotor and bowl. Does this distance vary? This is the first test often conducted by the biased wheel player.

 

[cyph]

Finally, observe the rotors decay rate. Does it noticeably slow down during a normal spin? If so, the bearings may need maintenance or replacement. Toclay s bearing systems" are remarkably efficient^ causing Very little decay within a single spih.

Detectioni^Kotor imbalanced 1 Pockets
Tojinspect the pockets thoroughly, many areas must be

•‘examined. For the purpose of a routine equipment evaluation, it is not necessary to examine all the pockets. The following tests should be performed on about ten randomly selected pockets, and if the results are consistent, the tests are adequate. If discrepancies surface, a moire thorough inspection may be appropriate. In the case of'suspect play, a thorough inspection of all pockets is mandatory.
Fret Height - To check that the height of all frets is consistent, sit a small cross level on top of two frets; if you check more than two at one time, you risk the chance of geting a level reading when, in fact, only the outermost frets have similar heights. Move over one fret each' time, and repeat the test for three to five connecting pockets. Then take a few samples from different parts of the wheel.
Fret Thickness - To check fret thickness, use the inside jaws of a dial caliper. These calipers come|inj many different models (vernier, dial, digital), and , ] are generally accurate to within 1/1,000", but only if the same part of the fret is measured each time.

To measure the fret at a point,closest to the cone, use the cone as a guide. When checking the fret at a point closest to the number strip, use the number strip as a guide. To check the middle of a fret, drop a solid metal object into the pocket and push it flush against the inside or outside; a thick, hexagonal nut works fine. Keep the caliper flush with this object while measuring..,^ As in the case of detecting crooked dice, a good rule of thumb suggests that any discrepancy s exceeding 5/1000" should be further investigated.

Pocket Width - To check the distance between frets, use the outside jaws of the caliper. Follow
Fifr'in Srrifpr

 

[cyph]

the same procedure as described for fret thickness for accurate measurements;-

Pocket Depth - To check pocket depth, use a dial depth gauge. Rest the gauges T-base onithe top of two' frets, keepingrtne gauge flush with either|||

;the inside dpiufside!o|jp the depth at theljfcemer of -the pocket, use a solid objectifor guidance, aptaring that the gauge is positioned over-the center'of the pocket each ‘time. -,'\r
Bounce - 'Drop andpcatch* was a term used by the oldtimers when testing the bounce characteristics of the pocket pads. Holding a roulette ball about two inches above the pocket, drop and catch the ball. Sample different ^

* sectors of connecting pockets in different areas of the wheel.

Detection: Inconsistency in fret height, fret thickness, pocket width, pocket depth, and any pocket pad bias.

Cone

r.Thanks£to,the low profile design, the ball; spends: mof^time bouncing^mthe cone than ever before^ If the ball doesn't stick to a pocket with little or no bounce, its likely to bounce on the cone before eventually landing.

As the wheel spins slowly, lightly touch the cone up near the turret. You are feeling for^any bumps, cracks, crevices, warpage, rough spots, slick; spots, and so on. Gradually work your finger down closer to the pockets. Pay special attention to the sector lines. •

Minor irregularities on the cone can be significant. This is especially true at the moment the-ball slows down, is gently houncing, and is ready to fall.

Detection: Cone bias.

Inside Bowl and Rotor Bottom

Follow the manufacturers instructions to carefully remove the rotor. First, inspect the general condition of pe bowl lining. Second, if any part of the bowl lining edge is visible with the rotor in the normal game position, it will lay in the ball's path just before contacting the number stripi Rub your finger around the edge aud feel for any high spots, warpage, or other irregularities, .
Fig. 13 - Standard depth gauge

 

[cyph]

Finally, if the’rotoribase^i^ewi^d with a security seal system, check and confirm all sesds/ /

' Detection:.cBias'to the inside bowl casing or rotodbase seal system. .

Ball

Inspect the ball for any chips) cracks, or other irregularities/and then run a magnet over the tops of the roulette balls to detect the presence of steel cores. '

^p^jon/Magnetic roulette balls and ball bias.

Software

There areimany statistical software packages! currently offered by roulette manufacturers. Data ^collected via the infrared readers and sent to the winning number displays can be integrated with a full-fledged analytical software. These packages provide numerous statistical measures for interpreting unusual results that can be presented in a variety of histograms, pie graphs) and other visual aids. If you currently [utilize such a program, the software can.be used in conjunction with the physical tests justidescfibed fori confirmation of bias. .

iDetection: Statistical aberrations which m^indicate|mechanical bias. ^

Manufacturers Guidelines

Finally, all manufacturers will provide you. with complete guidelines for the cleaning and inspection of their wheels. Also, many offer specialized tools for taking pocket wall, width and depth measurements, as well as other measurements. For the best way to proceed, always check with the manufacturer for advice on the best way to conduct an evaluation.,

ROULETTE COMPUTERS

Back in 1955, long before his landmark book on blackjack, Professor Edward Thorp and a few friends began brainstorming the possibility of beating roulette. Thorp envisioned the use of a small concealed computer device with timing switches. The computer would make predictions and transmit them by radio to the bettor. In Beat the Dealer (first edition of 1962), its interesting how the professor teased the reader:

Will the rapid development of modern science and technology continue to produce winning systems for various other games?

 

[cyph]

Then in the revised edition of BeafcthbDealer, (>1967)^he tips his mitt:

There'are seyeral/^pMQnMdip^^^iw^m^Ts'S^wfhetnmffouM^mWo&lem. wheels whethcnj&m ^iorlmtthey aremef^^^M

TheimofessO^^n^^jK^^P^hat hisJmethpd’yieldeaW44^¥advantage in practice,%but p robllte^ witnltn^lbci.f ohics foshtSf ffomiMd^P^^lul^y^be c^inosJHj

In the. ^lwlQjEt,Os%prior tojthe antidevice laws, I met a few players'Haimingpxb ha|g so^H limited swcessliwith rroule^^omputers, ^but iVe never met anyone; who hasfclaimed ftcg cruslnthe .-as odct because the percentage advantage for the playb^was potentially^

Many talented groups have struggled with roulette computers, and some df the early developers® of the blackjackFcomputer attacked foe problem, but with litde success. I remember- one team abandoning! the project and looking to sell all of their equipment and research out; of frustration. Then therefwas? another team with another unit. I, along with my partner, witnessed many hours of demonstration^! Theirl’device was an impressive,jsfate, of the "art electronics package with the’capability to, build databases oforoulette'wheels all tover;the world*’ The;price tag was $50,00)0.)- Iivoted>yes, but myj partner saidmo; he was^i&ernmth^m^e^lsn^i^^^feldme to ^t betsdowrtft^the machmeWas spitting outpredictions in the last two to thre^ revolutions. ; In retrospect, he was right. The project just seemed to fizzle out.

In 1985f^p Eudaemonic Piedetailedthe'plightof a group of scientists and their efforts’to build a roulette comput^j^ne'story ofifersa ghm^ajn^the enormous research and development needed ;?to solve theprediction puzzle, ahdtwereomeithetmhiyi,rechnological;bbstacl^®b appfeciatdthe techmcldj i flayorpf,theB book, here s^aiquickisyhopsis detailing some^of the research and developmentjhat went into their quest to develop the perfect predicfor. ^,

They set up ^^ea^i wheel wjth|eight , stations of infrared sensitive devices that wouldi measure velocity ^^nlijrne the bMlj^issed a station. A camera with stroboscopic flash was used to photograph the ball's ^bounce and scatter. The goal was to collect data on every conceivable variable in the game. After an extraordinary amount of data had been collected, the program was written and packed into a small computer much like the blackjack computers we know today. The algorithm would calculate coefficients of friction, wind,resistance, drag, rate of deceleration, and position,, and time of drop off for the ball. It knew in advance the speed, distance, and path of theh^S jtrajectoOT downjj^tq, the lower ball track and pockets. It tracked tire speed and relative position,,of the rotor spinning below* The goal was to predict one of eight sectors as the best bet. They considered lasers and radar for inputting the data, but decided on microswitches operated by the toes.. They used an, eye/toe .coordination device for improving human reflexes. For output, they considered polarized eyeglasses, ultrasonic technologies, light emitting diodes built into wristwatchesi and radio waves of every conceivable frequency.

 

[cyph]

Tlieir first Computers achieved only.*modest resdlts,’Ifo- improvemetiKwaV needeX^TvIbre * research and development was' required. Then, byStheitfown accounts, they conquered the games? & complexities, and, thanks to the advances in miniaturization, the computer andpower source (a battery pack) could now be completely hidden in the sole of a shoe! But the story doesn’t have a happy ending: Ji More technological problems 'surfaced. ^After five years of research^by sonpofthe smartest? ? people ontheTplanet, they beat the game for.'sn^timoneyjd^t ihe technological war, and failed to set the world on fire: Thii983, they put an4Htih^Gam^i^pnM5?p^

Investors Wanted for Computer System to Beat Roulette

Inll986,a mystefiohs book, titled TheRomedlPr6jedby Y. Gambol, appeared out of nowhere. Ijp&isappeared aln^ostx^^&t. Scheduled'to be released and sold for $i00fto $200^ro tlie';De^of my recollection), a funny thing happened. A^^Bkhown’comput^team from the East^^^^idnt want the information published, they made the author an offer .he couldnt refuse and the book never pfjjnade.it to the retail outlets. The.work detailed awery sopnisticate<®lgorithm for beating the game.

The book was a gem, and provided far more;insight into/^®mechahics^id intr iaine|jaffecting prediction gpihanpiKylother source.

£/ In London, 1985,jtwo gamblers were arrested fdrithe^™of a small electronicjdeyicel In airl ; lexcerpt from a’British newspaper, was .the follpirang commentatwlra

Las Vegas Roulette Secret is Out

^ . has( invented1 a formulaSy which*,h^^nmlw^Kjn vfrhtch third of the roulettepyheel the hall nj,

fr.jW.ill drop. iHeiis'ibbl&wffissesythejpin/Spefdffi of the balljofill in the gaps

i^his calculation. He never actually^at^n^fy^kj^tfp^^jdie result -sectorfr>ne,\two or^O.

. ; three — thebettipfawM

The players were arrested and detained for two days. After persistent pleading of their innocence,

. they were given the chance to demonstrate and explain their strategy.-They did, and all charges were dropped. It appears that a form bfrvisual prediction may-have cleared them. Despite electronical assistance in some form, perhaps only a signalling device, the authorities were comfortable with the t legal aspects of their strategy. The article claimed that the players were hugely successful/all over tne world.

A number of other roulette computers have surfaced overseas. Roulette Mastering Rotronis are two products out of Germany selling for $30,000. The Scorpion^ a combination roulette and blackjack computer, came out of Australia. <

 

[cyph]

Vr'ISCo the'left is the frontpage of a professionally designed twentyruv<l|| page skies brochure for the Rotronis cbmputer^Ari interesting feature^ of this product is*the keypad. When you purchase this unit/ you are L! i^veh&TOueithat enables sohnany predictions, say 2,000, and then the • unit locks up. At ithis time^you must call the seller, pay an additional fee, andrpei^anew code to continue using the device^The front pkge^ ^caption reads as translated: "

’ r Chance has been beaten. Rotronis is the unique hightech solutioh’for the | physically calculable permanent success at the game of Roulette'.jm

Roulette "computers are still around. In early 2003, a $4,500 unit ■ was being sold on eBay with no takers publically sharing an interest. The unit was Psion palmtop computer with two external input switches and a pea-sized induction loop receiver. The owner claims to have invented -! the perfect system, further asserting:

fFig. ^14 - Roulette'c&mMter 4 * * ^It’s'impossible to get'iladght1 wth the equipment I supply.

f from,Germany

More than one player has taken a step back frbm&mill-fledged calculating, miniature computer and opted for a partial electronic approach. Sin; 1983J Beat'Roultitetyith a New Patented Discoverj/i(Scott' Lang) presented a novel timing system utilizing an ordinary digital wristwatch set'in a stopwatch mode. Although the validity of his system remains questionably it may have sparked other developments (see VisuahPfe'diction).

How They Work

Basically, these units all work in a similar fashion. There are five essential pieces of information that must be input: rotor speed, ball speed, where the ball exits the track, where the ball first contacts-the^ rotor, and the winning number.

. This information, and much more, are collected before the play to build a database from and generate actual prediction. The process continues during play to both expand and update the database.

Here's a simple methodology for the typical unit during play. Envision a Small device with both an A and B switch. The operator must first time rotor speed. As the single zero, for example, passes a reference point the A switch is hit, and when the single zero passes the same reference point again, the A switch is hit again. The B switch, a control function, is now hit to alert the machine to