[Note: to adjust for inflation, multiply amounts by 8.15]


There are so many myths connected with gambling that you couldn't explode them all if E equalled mc3 instead of mc2 Generally speaking these are harmless myths, resulting from ignorance or superstition. But one myth stems straight from the most advanced knowledge. It is: You can't beat the house.

The arguments run like this:

The house has a mathematical advantage on every bet you place (true).

Because of this mathematical advantage, the house stands to win about 1-1/2 to 15 cents in every dollar you bet (true).

There is no way to add up losses so as to show a profit (true).

Therefore anyone who bets against a gambling house will eventually lose (false; utterly, completely FALSE).

This myth pleases the reformers who must maintain that gambling, like crime, doesn't pay. It pleases the hoodlums (and never think they aren't hoodlums) who run the gambling houses because it keeps the reformers off their necks — and they know that never, never would it discourage the pigeons who patronize their emporia of chance. Morally, sociologically and mathematically it's an admirable myth. The only thing wrong with it is, it just ain't so.

By judicious betting and money management, it is possible not only to beat the house but to have a mathematical expectancy of doing so, even though you are betting constantly with the percentages against you.

To the mathematician this sounds as ridiculous as the chestnut about the merchant who loses money on everything he sells but makes it up on big volume.

The fact remains that a few thousand small-time gamblers, scattered throughout the gambling centers of the world, make their livings this way, have done so for years, and seem likely to continue to do so. Eventually most of them are due to go broke, mathematically speaking; but when you come right down to it, eventually most of the gambling houses go broke too.

I do not base these statements on superstition or iconoclasm. I rely on the same mathematical bases as the writers who warn all gamblers that they are suckers. Admittedly I am affected by the many small-time gamblers I have known. For thirty years I have seen them permeating the casinos, haunting the tables, content with their few dollars a day and always making them. They are the subsistence gamblers. They don't want to be Nick the Greek; they don't want to be the man who broke the bank at Monte Carlo; they just want to live. And, long ago, I observed that they do live — they live, or rather they subsist, in defiance of the mathematical probabilities that I was taught to hold holy. Since I was faced on the one hand with a fact I could not dispute and on the other hand a theory I could not refute, I had to find a reason.

It is a complex reason and must be explained in sections.




I said before that most of the myths of gambling stem from superstition or ignorance. Rather they stem from a misconception of what gamblers usually call the law of averages and mathematicians call the theory of probabilities.

The theory of probabilities is not capable of proof yet it is not open to argument or question. It is fact — the fact that makes Monaco a taxfree state, builds multimillion-dollar hotels in Las Vegas, and causes most gamblers to die broke.

Anyone who explores the theory of probabilities must start off with three fundamental truths:

1. Nothing that is mathematically possible is wholly improbable. You may have heard the story about the monkey and the typewriter. If you place an immortal monkey at a typewriter that can't wear out and have him punch the keys haphazardly throughout eternity, sooner or later he will type out all of Shakespeare's sonnets without missing a period, a comma, or an Elizabethan misspelling.

Perhaps no one has ever seen a thousand straight passes at craps or a thousand reds running at roulette, but if you could keep the game going enough billions of years such runs would become mathematically probable.

2. Statistics mean nothing — absolutely nothing. When they run true to form, they can be mildly reassuring to a mathematician; when they seem to defy the probabilities they can warn a roulette tourneur that his wheel is on the blink or a blackjack dealer that he is being cheated; but they are never conclusive. In calculating what is to come, you must totally disregard what went before.

3. The probabilities, or percentages, do not necessarily affect any individual. Suppose there are ten equal horses in a race. The proper odds against each horse are nine to one, but ten wise guys persuade ten chumps to accept eight to one, $8,000 to $1,000 each. The instant those bets are placed, theoretically and mathematically each of the chumps has lost $100 and each of the wise guys has won $100. But not one theoretical $100 ever enters into the result. What actually happens is that nine of the chumps lose their G's entire. One of the chumps wins $8,000 and is just as happy as he could have been if his bet had been a wise one instead of a stupid one. One of the wise guys loses $8,000 and winds up just as broke as he would have been if the odds had been against him instead of for him.

There is a parallel case in the gambling house that works on a 1.4% edge, as the best American and European houses do. Into each of seventy-three casinos walks a plunger who makes one bet covering the entire resources of the house. Thirty-seven of the houses will win and rejoice, thirty-six will lose and go out of business. The theory of probabilities has been gloriously fulfilled but thirty-six gambling houses are ingloriously bankrupt.

As long as a gambling house has the odds in its favor, it is mathematically due to make money. The question is, when?

Because nothing is mathematically improbable, a house pitted against unlimited capital over an indefinite period of time will ultimately run into a series of losses that will break it. The more limited the house capital, the sooner this is likely to happen.

The patriarch of the gambling business, the adventurous Franois Blanc, who created the casinos at Monte Carlo, went broke in his first two starts. He went broke because Prince Charles Bonaparte walked into his houses with more money than Monsieur Blanc had. Sooner or later the plunging prince had to clean out the prince of percentages.

The venerable Monsieur Blanc then enunciated the principle that is still the soul of the casino business: "The house must either (a) have more capital than any individual gambler, or (b) establish a limit so that superior capital can never be brought into play against it." Every gambling house since has adopted that principle.

But the principle works both ways. If a small-time gambler has enough capital so that his bets are very small compared with that capital, he can make a lot of bets before he loses his capital. The more bets he makes, the greater the likelihood that from time to time he will be ahead of the house before the house advantage asserts itself and takes his capital away.

I don't care what kind of betting system you propose, any mathematician worth his salt can show you a fallacy in it and prove that it stands to lose rather than win. Furthermore, he'll be right.

But he is right only if the gambler following the system sits there at the table and continues to bet for an indefinite period. This doesn't happen. Players do leave the table and while they're away they spend their winnings. The house can't win back the money the bettors have already spent.

Remembering that all things are possible, this must be stated: If a million gamblers took about $20,000 apiece and interminably made small bets against gambling houses, three or four unfortunate souls would go broke within a month or two. They would be balanced by a scant handful of others who would never lose as long as they lived. A few would last only a year or so, a few would last only two or three years, and so on. But the majority would last ten or twelve years and from time to time would be ahead of the game until that inevitable last time when the house would break them.

The two things that knock the bettor off eventually are the house advantage and the long losing streak. Obviously it behooves the bettor to pick a game in which the percentage is least against him and a betting system that keeps his bets low while maintaining his chance of winning until that long losing streak comes along.

This brings us to the subject of betting systems, which deserves a section of its own.




Innumerable are the systems that have been devised to beat the game. The process has been going on so long that what seems newest is probably oldest. Some of the systems were old when the pyramids were young. If you are looking for a good system, don't waste time. Look them over; take your pick. Mathematically they are all the same.

The probabilities are fixed, stable, unassailable. They cannot be altered or modified by any betting system. Measured against the immensity of infinite time, which dwarfs even light-year figures, one system is not and cannot be better than another system; it is not and cannot be worse than another system; it is not and cannot be better or worse than no system at all.

But betting systems are not measured against the immensity of infinite time. They are measured against the picayune period of one short lifetime or an even shorter segment of it. By this criterion, one system can surely be more satisfying emotionally than another system.

The reductio ad absurdum of "no system at all" is to take all the money you have and bet it on one chance. In the best available games (craps in the United States, roulette on the Riviera), you have a 49.3% chance of winning and a 50.7% chance of losing. If you take the same amount of money, divide it, and make a thousand bets with it, you stand to win 493 of them and lose 507. Mathematically it is all the same thing. Subjectively there is all the difference in the world. Sudden death (the one-bet system) makes a man rich or poor in less than a minute. Very few persons are willing gamblers to that extent. Slow death keeps him alive for a long time, and while there's life. . .

The more bets you can make with your available capital, the more likely it becomes that from time to time you will be ahead of the game. The subsistence gambler lives on these times.

There is an old saying, "A player without a system is seldom without a dollar." The player who has a system is usually trying to borrow a stake; whereupon, he is confident, he will clean up.

The player without a system isn't better off because he has no system; he is better off because he does not place childish reliance in a system. The player with a system isn't worse off because he has the system; he is worse off because he is trying to clean up. To clean up, he has to place relatively big bets. The lasting power of his capital is reduced accordingly.

Keeping all these things in mind, consider the most popular systems:


THE MATURITY OF THE CHANCES. This is the most prevalent and yet the most fallacious of the systems. It is based on the popular misconception of "the law of averages" combined with the belief that "things even up in the long run." The player knows, for example, that the odds are about thirty to one against a run of five reds at roulette. He waits until red has shown up four times consecutively, and then he bets against it, thinking "the streak is due to end."

This misconception is found in the strangest applications. Leo Durocher had a .250 hitter walked to get to a .330 hitter; the .330 hitter popped up and Durocher's strategy was vindicated. "I was playing percentages," he later explained smugly. "The .250 hitter averages one hit in four times at bat; he had been up three times without a hit and he was due. The .330 hitter averages one hit in three times at bat; he had had his hit." Durocher was just as wrong as the system player. The probabilities are never affected by what has gone before. It's very unlikely (1,023 to 1 against) that a coin-tosser will toss ten straight heads; but if he's already tossed nine straight there is as much chance that the run will continue to ten as that it will end at nine.

The player who depends on the maturity of the chances doesn't reduce his expectancy of winning, but neither does he increase it. He just wastes a lot of time keeping records.


MARTINGALES, or DOUBLING UP. This is the simplest of the systems. You bet only on the even-money chances, such as red or black at roulette, or pass or don't pass at craps. Every time you lose a bet you bet twice as much as the next time. Regardless of how long you lose, every time you win a bet you have netted one unit on the series.

For example, you bet one and lose; you bet two and lose; you bet four and lose; you bet eight and win. You have put up fifteen chips and you get back sixteen; you have lost three bets and won one and still you are one unit ahead.

This system is thwarted by two factors: First, the geometric progression is so rapid that in the course of a year or less a run will occur that exhausts any reasonable capital. For example, every year or so there should be a run of fifteen straight losses. Even if the rules permitted it, you would be betting 32,768 to win that basic one chip — with less than half a chance of winning. (And on the first fourteen losses, you would have bet 32,767, making your total investment 65,535 to win one.) Second, the house has a limit that makes it impossible to double up indefinitely. In some houses the limit will be 1,000 (covering only a run of ten, which should happen about once in two weeks); in some houses the limit is 2,000 (covering only a run of eleven, which should happen about once a month). When you reach a limit of 1,000 and can't double any more, you have lost 1,022 chips. When you reach a limit of 2,000 and can't double any more, you have lost 2,046 chips. It takes months of chipping away at those lost chips to bring you back to even.

Nevertheless, at least 90% of all the subsistence gamblers select the martingales system. They simply limit their doubling up. They will double up through a run of four losses, betting 1, 2, 4, 8. At this point they have lost 15 chips. They pocket this loss and begin the series again with a one-chip bet. (Some take five straight losses, at which point they have lost 31.) For the ones who follow the four-loss system, if they place 1,000 bets a week then in the course of a month of normal results the long losing runs will have cost them about 3,800 chips, the normal runs will have gained them about 4,060 chips, and if the chips are worth a dollar apiece they have $260 to live on during the month. They won't go broke until there is a run of losses that is mathematically likely to occur only once in a million bets. That will happen once in about twenty years.

To make this profit of $260 a month, you have to play three or four hours a day, five or six days a week — in total, about twenty hours a week. If you have $20,000 capital, you are due to go broke within twenty years. It might happen the first year and it might happen the twentieth, but the mathematical expectancy is that it will happen toward the end of the twelfth year. In this case, you will have taken out $3,120 each year, or better than 15% on your $20,000 capital, and you will have grossed $37,440 while living all the while at a rate of $3,120 a year. It may seem like a meagre living, but (as umpire Hank O'Day so classically remarked) "You can't beat the hours."


LAB0UCHÈRE, or CANCELLATION. Oswald Jacoby, in his book How to Figure the Odds, says this is a system used by the gambler "who has a childlike faith in figures." The object of the system is to make one win compensate for two losses. You write down a series of figures like this:








You bet the sum of the top and bottom figures, in this case 4. If you lose, you write the amount of the lost bet at the bottom of the column:










—and again you bet the sum of the top and bottom figures, which now is five. When you win, you cross off the top and bottom figures and bet the sum of the figures that are now at the top and bottom.

This system is so plausible that a few years ago Esquire itself published an article about it, showing a "typical" run at roulette in which one player always played red and one player always played black and they both won. The fallacy is there, nevertheless. If the zero happens to fall when both players are making low bets, the house will lose. If the zero happens to fall when both players are making high bets, the players will lose.

In the long run the house will make its usual percentage, but this is still a good system for players who want some paper work to occupy them while they are going through the dull rounds of making one bet after the other. What with crossing off old figures and writing in new ones, they won't be bored. The paper work doesn't increase their chances of winning, but neither does it reduce them. The player of "Labby" with $20,000 capital and a four-dollar starting bet, playing twenty hours a week, doesn't figure to go broke for nine years, in which time he will have won almost $360 a month average and will have grossed $39,700 — better than the martingales player if you are willing to starve three years for the sake of nine years of comparative luxury.


ALEMBERT, or PROGRESSION. This is the most respectable of all the systems because it is named for and may even have been devised by Jean d'Alembert, great French mathematician of the 18th century (he was born in 1717 and died in 1783).

Also it is probably the best of the systems. If I were going to become a subsistence gambler I would use the Alembert, despite the practical experience of the 90% of professiona1s who use martingales.

The Alembert works like this: You start with a bet of ten units. Every time you win a bet you decrease the bet by one unit. Every time you lose a bet you increase the bet by one unit. So if you bet 10 and win, your next be is 9; If you bet 10, and lose your next bet is 11.

If your wins exactly balance your losses, you win 1/2 unit per bet; or, to put it another way, you win one unit per winning bet. It doesn't work this way every time, but it averages up this way until along comes a succession of losing series that breaks you. In the best available games you win thirty-six bets to every thirty-seven you lose, and you can stand that extra loss for a long time. It would take about fifty years to break you if you never spent your winnings, and at least ten years on the average if you spend your winnings of nearly $5,000 a year and save your $20,000 capital intact. By that time you would have taken out about $50,000. The usual warning must be appended: Of all the gamblers who try the system, about 2% will go broke the first year and at least 20% will go broke within five years and at least 1% will still be winning when they're a hundred years old; like the gambling house itself, you have to take your chances. The majority will subsist for ten years or more.

Everything that has been said about these systems assumes that you will pick the best available game and play it properly, so that must be the next section.




Whatever his system, the subsistence gambler makes only even-money bets and plays only in a game in which the house has the lowest advantage.

Ultimately the gambler is beaten by a long series of losses. The mathematical advantage of the house makes such strings come more frequently. For example, if there were no house advantage, eleven consecutive losses could be expected once in 14,096 bets — about once in a month's play. At the best available games, where the house has an edge of 1.4%, this expectancy is increased to once in less than 4,000 bets. In most gambling-house games, where the house edge is 5% to 7-1/2% or more, the string of eleven straight will occur about once in 3,000 bets, or three weeks' play.

Old-timers will assure you that the best game of all was faro, which was America's national gambling game until craps came along. Faro was unique among house games in one way: There were some bets in which the house had no advantage at all.

But you can't find a faro game today except in some of the biggest Nevada houses, where it is maintained more as a historical curiosity than as a paying proposition. Even when faro did flourish, the house soon chased a gambler who waited for the "cases" — the even bets — and besides, these occurred so seldom that the gambler would spend three-quarters of his time waiting and only one-quarter playing. Nearly all faro players were nearly always broke. Forget about faro.

The classic game, of course, is roulette; and in the casinos at Monte Carlo and some other European watering places it is also the best game. In the United States and every other part of the Western Hemisphere it is not a good game at all.

The best available game for most Americans is craps. But there are crap games and crap games, and you have to skirt certain pitfalls in picking your game and your bet.

Since these are the ONLY games in which the subsistence player has a decent chance, each deserves a fuller description.


ROULETTE. All the glamour traditionally associated with gambling surrounds roulette. The paraphernalia is colorful, with the wheels of polished mahogany and burnished metal, the green felt splashed with red and black lacquer, the chips of bright translucent plastics inlaid with German silver, as eye-catching as costume jewelry. The men in evening dress, the women bejeweled and décolleté, the suave, uniformed croupiers, all in a setting of resort gaiety, seem glamorous too — at a distance. Close up, their visages are typical of the gambler's mien, as grim and vicious, as distorted by the usual frustration of loss or the occasional elation of victory, as Holbein would ask if he were seeking anew models for the danse macabre.

In the United States roulette is a minor game, something to amuse the ladies and incidentally collect their quarter and dollar bets while the bolder men risk real money on the dice. Along the Riviera it's the other way around. Except for the Greek syndicate and a few card-conscious plungers who prefer baccarat, and a few others who like the novelty of craps (recently introduced there), most of the big money heads to the roulette tables.

To attract this big money, the casinos make the greatest concessions known on either side of the Atlantic. There are thirty-six numbers on the wheel (half red, half black) and one zero, where American wheels have two. When the zero shows, the house takes only half the bettor's money, instead of all.* Overall, the house advantage is 1.38%

There are six possible even-money bets, usually called by their French names: rouge (red) or noir (black); pair (an even number such as 14) or impair (an odd number such as 15); manque (a low number, 1 to 18) or passe (a high number, 19 to 36). Mathematically it does not matter which of these you play or how you shift your bets from one to the other. The chances of winning and losing streaks remain the same.

Compared to other gambling games, the pace is slow at roulette. The casino tries to establish a rhythm by which there will be a fixed number of coups or turns of the wheel per hour. This number ranges from 50 to 60. Playing three or four hours a day, the subsistence gambler gets in about two hundred bets.

On the even-money bets, the casino invariably establishes a low minimum and a high limit — say $1 minimum and anywhere from $2,000 to $3,000 limit. For the casino knows that players of some systems want to double up, and the casino wants to attract the system-players. That is one of the many paradoxes of gambling — however sure you are that you are going to win, above all things the house wants to see you try.


CRAPS. For about forty years this has been the principal gambling game of the United States. In American gambling houses it replaced faro for several reasons, but the principal reason was the speed of the crap game.

A gambling-house crap game, or "bank craps," is played on a big green-felt-covered table that resembles a billiard table except that it is bigger even than the 5-by-10 championship tables. The felt is marked with a layout on which players place their bets. A "stick man" handles the dice and runs the game; three to six men pay and collect bets. There are no seats as at roulette. Everybody stands. It is tough on the feet.

To the uninitiated the speed of a house crap game is likely to be bewildering. Whereas the rule of thumb in faro was, and in roulette is, one turn per minute, in craps a bet is decided every twenty or thirty seconds and a rate of 150 bets per hour is not unusual. The chatter of the stick man is like bursts from a machine gun; the house men are trained down so fine, raking in and paying out the chips, that they jerk in their sleep (I quote from one of them, and with him it was much too serious a subject to permit of facetiousness); and before the novice knows whether he won or lost the whole game is likely to be three rolls ahead of him.

There are "square" crap games and "gyp" crap games. You can easily differentiate between them, by the layout on the green felt of the table. The square game will have a space marked "DON'T PASS — BAR 1-1." (". . . BAR 6-6" is also a square game.) The gyp game will say, in the same space , "DON'T PASS — BAR 1-2." That digital difference makes all the difference.

These esoteric markings mean that if you bet the thrower of the dice will lose, and he does lose by throwing 1-1, or 6-6, or 1-2, respectively, you don't win your bet. You just get your money back. When the house holds out 1-1 or 6-6, its advantage is 1.4%. The 1-1 (or 6-6) is due to come up once in thirty-six rolls. When the house holds out 1-2, its advantage is 4.4%. The 1-2 is due to come up once in eighteen rolls. To the subsistence gambler the difference is enormous. He can never do better than surrender 1.4% to the house. He cannot contend with anything as big as 4.4%.

Actually it is unfair to call the 1-2 games gyp games. They aren't gypping anybody. They are simply demanding a bigger profit. Usually they have a low-betting clientele and need more profit to pay their overhead. Big money is wise money and wise money looks for the best odds. A gambling house will always reduce its edge to attract big money. Remember that if a house can get 1.4% on an average play of $100 it will make $1.40 every twenty or thirty seconds. If it takes 4.4% on an average play of $10 it will get only 44 cents.

The subsistence gambler should look for the so-called square game, but he doesn't have to. He can play in any gambling house, bet right, and surrender only 1.42%, for that is the house advantage when you bet on the dice to win. The difference of two one-hundredths of one percent is not going to cost him a great deal, even over a period of ten or twelve years.

Most gambling houses have a low spread between the minimum and maximum bets on a crap game. If the minimum is $1 the maximum will usually be about $200, as compared with $2,000 or more in roulette. This seriously restricts players of some systems but does not materially affect players of the systems described earlier in this article.




My calculations have been based on a capital of $20,000 plus the desire to live comfortably for ten or twelve years at a rate of $75 to $100 a week. This is far more than you could get out of your twenty G's if you simply invested them at the best available interest rate and invaded the capital for the rest of your living expense. It is far less than you would have if you could find any kind of decent job and add your earnings to the interest on your invested capital.

Let there be no mistake about it: If you are physically employable and temperamentally able to work, you are better off working. Subsistence gambling appeals mostly to elderly persons who cannot find any jobs (or who cannot find palatable jobs) and to unfortunate youngsters whose psychopathic personalities make them unwilling to work.

Subsistence gambling is a career for the patient and the resigned. It cannot appease the hustler, the chiseler, the get-rich-quick. The harder you try to break the bank, the sooner the bank will break you.

These are the rules for the subsistence gambler:

1. Set an amount you want to win each day. It should not be more than one-tenth of one percent of your capital.

2. A day ends when you have won that much. If luck is with you and you win it fast, keep on playing and add any additional winnings to your capital. It doesn't belong to you, and furthermore you are going to lose it back when luck is against you.

3. Select a system and stick to it. (See analysis of systems, above.) Deviation will not affect your mathematical chances, but it will affect you psychologically. Being human, you cannot help feeling regret over "might-have-beens." If you are irrevocably wedded to one system in advance, this will not happen.

4. Don't worry when your capital goes down. Five or six times, over the course of your planned gambling career, you will be almost broke and still come back — perhaps to restore your entire original capital or more. Once you won't come back, and then you will be an ex-gambler.

5. Don't splurge on winnings or retrench when you lose. Either policy lures you away from the purely objective approach. Most users of martingales have gone broke making maximum bets in an effort to get even before starting the system again.

Capital of $20,000 is beyond the reach of most gamblers. I have known many who had $5,000, or even less, and who are still going strong after almost thirty years. But my friend Jean de Beausacq, who lives in Monaco and has been playing roulette up and down the shore for nearly thirty years, would view them with horror. "They," he would exclaim, "are gamblers!" No more opprobrious term exists among the gentry who rely on paradoxical percentages for their meager livings.




There is no dearth of literature on "how to beat the bank." There are pamphlets on systems, money management, the record of the wheel (what numbers have come up, hour after hour, for years and decades), and everything else imaginable about gambling.

Most of these pamphlets prove that you can live on your gambling winnings.

And who do you think puts them out?

Right. The gambling houses themselves.

Once I was an intellectual expert on the subject of legalized gambling. I was a profound scholar on the subject. I had steeped myself in statistics and surveys. From this scholarly data I knew that legalized gambling, wherever it had been tried during the cynical 19th century, had resulted in moral and pecuniary impoverishment of the people.

It wasn't till years later that I learned that every pamphlet of statistics had been prepared and promulgated by the entrepreneurs of the Monte Carlo casino. Object: To discourage competition from other small European principalities and grand duchies.

Those same Monte Carlo gambling impresarios conduct a running census of the minor professionals who live on their gambling winnings. These professionals make five to ten dollars a day. Their numbers are inexorably diminishing. There were about 200 of them in 1910, about 120 in 1930, about 50 in 1950. Some of them quit because they went broke and had to, some of them died (because their average age was at least sixty), but most of them simply found jobs that paid better. Professional gambling falls as employment opportunities rise, and we are now in a period of full employment.

There are some, of course, who gamble professionally because they prefer gambling to any other means of livelihood. These we have always with us. They are the gentry one sees along the Riviera from Monte Carlo through Nice to Cannes, at Deauville and at Hialeah, at Reno and Las Vegas and Gardena et passim. They have existed from the earliest times and in the most primitive societies. The American Indians, long before the coming of Columbus, knew their itinerant dice players, the Yenadizzibug, who went from wigwam to wigwam looking for a bet. The modern archetype is a little old man with a goatee, a refined manner, a suit well-pressed under the mattress but shabby at the principal points of contact, and a collar stiffly starched but fraying under the edges. The Yenadizzibug and the goateed old man are brothers under the skin.

Behind all the archaeological data and the unreliable propaganda one fact stands out:


The gambling houses know that some minor professionals make steady money gambling, and they don't care.

Every gambling house employs shills anyway. Shills in a gambling house are like respectable-looking old ladies in sightseeing buses. No one wants to be first to do anything so the shills pretend to be gambling until cash customers arrive. Then the shills fold their tents and steal (perhaps a better word would be sneak?) away.

The subsistence gambler doesn't take any more out of the gambling house than the average shill. You don't have to pay social security on him and he doesn't take up any shelfroom. Let him live.

But there's still a limit, just as there is on the even-money chances. A couple of dozen at Harold's Club, a couple of hundred at Monte Carlo, and who cares? But too big a rush and the house will put its foot down. It's not hard to be barred from a gambling house, and then where do you go?

So do we all go around in circles and find ourselves finally at our starting point. I've told you how to win money gambling. I tried it once myself, and I didn't like it. I've known many others who tried it and didn't like it either. It can be done, but I don't recommend it.

*The casino does not actually take half your money. Your bet is "put in prison" and decided by the next turn of the wheel. Then, if you win you get all your money back, if you lose you lose it all. This is equivalent to taking half your money; but it seems to be a difficult concept to grasp — among others, John Scarne, an authority on gambling, couldn't figure it out — so I had better explain it. When you place a $10 bet on an even-money chance such as red, theoretically you put up $10 and the house puts up $10, making a $20 pool, winner take all. When your money is put in prison, the total pool becomes $10, winner take all, and you have put up not only your $5 share but also the house's $5 share. Therefore you have only a $5 share in the new pool and have lost about half of your original bet. It would be exactly half except that on the determination of the new bet the house still has its mathematical advantage. If the house actually gave you back half of your bet, its mathematical advantage would be 1.37 %. Using the "prison" system, it increases this advantage to 1.387%. The calculations on which this article is based do not take into account this infinitesimal difference; they assume that you get back exactly half of your bet. (back)