Make Money at Roulette!

[Hyperborea]

I ran across this in the news and considering that we had been talking about "risk" and comparing it to putting it "all on red" this seemed apropos. Though it does seem that it does require brains and probably some work to implement.

http://www.thisislondon.co.uk/news/a...0Standard&ct=5

Quote:

Gamblers who "won" £1.3million on the Ritz hotel's roulette wheel with the help of hi-tech computer gadgetry have escaped prosecution. The three - a Hungarian woman and two Serbian men - have been told that because their scanners only provided information and did not manipulate the equipment, there were no grounds to prosecute. Casino managers became suspicious after the trio won £1.3 million over two days in March. On the first night they won £100,000. They returned the next night to win a further £1.2 million, which they quickly cashed. After declining offers of free caviar and champagne, they departed soon afterwards. Suspicious security experts examined CCTV footage and officers from Scotland Yard's gaming squad later arrested the three at a nearby hotel and impounded the cash and mobile phones. The three - a 32-year-old Hungarian woman described as "chic and beautiful" and two Serbs aged 33 and 38 - were questioned and released on bail. Now they have been told that they are free to leave the country. Police sources said it was notoriously difficult to get enough evidence to prosecute people over alleged casino fraud and in this case there was no proof they had "technically" interfered with the equipment.

[Nords]

The news has been out for years on roulette wheels' individual idiosyncracies. Usually casinos encourage gamblers to keep track of the spins because a gambler with a system is a happy big-spending gambler. At the same time they're changing wheels & croupiers frequently to disrupt the wearing-in process. It's the casino's fault if they were too lazy/disorganized to change the wheel after a bad night.

1.3 million-- that's what, a week's casino earnings? Heck, they must be paying at least that much in bribes to Scotland Yard.

Ever notice how successful gamblers are always "chic & beautiful"? None of them ever resemble Phyllis Diller or Don Rickles!

[cute fuzzy bunny]

But did they get to keep the 1.3M??

[azanon]

On Roulette, just find a table that have low table bid minimums (like $5 dollars), and just start out betting 5. *If you lose, bet 10. *If you lose, bet 20. *If you lose, bet 40, and so on and so forth. * Once you eventually win, (say you won at 40) the $40 pays back the lost 5, 10, and 20 and gives you $5 dollars profit. * Go back to $5 dollars, wash rinse and repeat. Stand there and make $5 dollars at a time all night long.

Gotta have a lot of money in your pocket though case you have to go 7+ stages deep. *Also, just hope you're not so unlucky that you lose 7+ times in a row :-). *Other than that, you can make $5 bucks over and over :-).

The color you pick is irrelevant, and you can switch if you want. *Its a statistical falacy to think that because it hit, say, red 5 times previous, its more likely to hit black next. *The reason is each outcome is "mutually exclusive".

[sgeeeee]

Quote:

On Roulette, just find a table that have low table bid minimums (like $5 dollars), and just start out betting 5. *If you lose, bet 10. *If you lose, bet 20. *If you lose, bet 40, and so on and so forth. * Once you eventually win, (say you won at 40) the $40 pays back the lost 5, 10, and 20 and gives you $5 dollars profit. * Go back to $5 dollars, wash rinse and repeat. *Stand there and make $5 dollars at a time all night long. Gotta have a lot of money in your pocket though case you have to go 7+ stages deep. *Also, just hope you're not so unlucky that you lose 7+ times in a row :-). *Other than that, you can make $5 bucks over and over :-). The color you pick is irrelevant, and you can switch if you want. *Its a statistical falacy to think that because it hit, say, red 5 times previous, its more likely to hit black next. *The reason is each outcome is "mutually exclusive".

Odds are still against you in the long run because the house gets 00.

[BigMoneyJim]

My aunt had the same wonderful idea, but as Salaryguru said it neglects the 0 and 00, both of which are losing bets for black and red. By the way, American tables have both 0 and 00 but I hear European tables have just 0, so we get worse odds on this side of the pond.

[Hyperborea]

Quote:

The news has been out for years on roulette wheels' individual idiosyncracies. *Usually casinos encourage gamblers to keep track of the spins because a gambler with a system is a happy big-spending gambler. *At the same time they're changing wheels & croupiers frequently to disrupt the wearing-in process. *It's the casino's fault if they were too lazy/disorganized to change the wheel after a bad night.

It doesn't require a worn-in wheel or one that is skewed though that may help. *The process uses a measuring system (in this case a laser) and a computer to calculate the probable resting place of the ball. *I read a book that described another similar system that was developed in the early 80's. *It was a computer system in a shoe that required two to operate. *One would use the computer's toe inputs to enter information about the ball and wheel. *The other operator would get signals from their shoe computer about which octant of the wheel was the most probable. *By skewing the bets to the right octant you would improve your odds tremendously. *The guys who developed that system were grad students at the time and IIRC many of them went on to do some of the seminal work in chaotic systems (an area of math/physics).

In doing a quick Google search on this I came across this rather interesting paper by Thorp (of Beat the Dealer fame) who claims to have done just this same thing in the 50's with Claude Shannon (the father of information theory).
http://c2000.cc.gatech.edu/classes/c...ings/thorp.pdf

I remember reading that article shortly before I went to LV last year. As I recall it said their device had the appearance of a cell phone. I don't think the scam would work in LV - on every spin you would have to appear to be using your phone, then rush in a bet after the wheel is spun. Seems to me that would get you "noticed" pretty fast out there, and anyway they usually tell you to step away from table games when using a phone.

But I think this group kept the winnings since they didn't violate any law. Just like BJ card counting is not illegal in LV - casinos can "bar" you but they can't take the money away after you win it...

[Hyperborea]

Quote:

I don't think the scam would work in LV - on every spin you would have to appear to be using your phone, then rush in a bet after the wheel is spun. Seems to me that would get you "noticed" pretty fast out there, and anyway they usually tell you to step away from table games when using a phone.

I don't think that's exactly how it worked. One person was standing down by the wheel "watching" the game and not playing. The cell phone was in their hand but not being used for calling. It was pointed at the wheel and an invisible laser beam was scanning the wheel/ball. It either did a calculation on that data or passed the scanning info to somebody else's hidden phone which then did the calculation. Now, it gets kind of fuzzy but the player's phone needs to signal them as to which octant of the wheel to bet on - it probably used vibration in some way (patterns). Now there were three of them so perhaps there were two bettors alternating somehow as to who played the predicted octant - to help conceal the scheme though obviously not well enough or we wouldn't have heard the story.

[azanon]

Quote:

Odds are still against you in the long run because the house gets 00.

Yeah, a friend of mine and I used a basic program to simulate it, and basically the way it works out is more times than not, you could come out of the casino with a bunch of 5 dollars wins, but eventually you lose really big. * I was aware of the 2 greens that the house gets.

BUT if you simply looked at the probabality you'd come out of the casino any one night with more money than you came in with, odds are strongly in your favor if you only used that technique, esp if you're prepared to go at least 7 spins deep. *

I forgot how many total numbers there are but it'd be roughly (.48 )7th (superscript) chance of losing big on every 7 spin series. *(not .5 to the 7th cause house gets 2 greens)
(which comes out to roughly 0.5% chance that would happen).

[gratefuled]

The only even bet in most casinos is the free odds line on a craps table. Every other bet in the house is skewed against the player. Even the sports book.

My inner geek loves probability theory. Las Vegas is like a gigantic case study and every time I'm there I wander around figuring odds in my head. You can have the best conversations with dealers about the house edge; some of them are geeks too. "How does this game work? Why is it worth the floor space?" Etc.

Ed

[BigMoneyJim]

Quote:

BUT if you simply looked at the probabality you'd come out of the casino any one night with more money than you came in with, odds are strongly in your favor if you only used that technique, esp if you're prepared to go at least 7 spins deep.

What do you mean by "odds are strongly in your favor"?

Quote:

I forgot how many total numbers there are but it'd be roughly (.48 )7th (superscript) chance of losing big on every 7 spin series. (not .5 to the 7th cause house gets 2 greens) (which comes out to roughly 0.5% chance that would happen).

Define "losing big".

The roulette wheel is pretty simple. All the odds are variations of 1:38 paying out 36:1. The house has about a 5.26% advantage on all bets all the time. If you focus your bets or use the doubling strategy on red/black your wins will be bigger and your losses will be bigger. If you play steady bets on red/black or 1/3 of the board your wins and losses are smaller.

Aside from the odds being in the house's favor, the bankroll is in their favor, too. You may try to play until you're on an upswing, but if you lose too much you have to stop from lack of funds. I'm curious if this increases the real take of the house; intuitively it seems like it might, but I can't apply the math to it. Anyway, the doubling strategy increases the chance of winning more but it also increases the odds of going broke fast. The house edge is still just over 5% no matter how you play.

When I play the casino I look for the $5 craps table and start with pass line and single or double odds. I don't really believe I have the ability to predict rolls, but it's still fun to place $10 pass or higher odds if I get a feeling. Sometimes I'll play a $5 field or a $1 hardway on feeling even though I know those are worse payout odds. (Although the field isn't worse than roulette, especially if 12 is tripled instead of doubled.) If I lose I bet more conservatively. If I lose rapidly I leave and play quarter or even nickel slots, and I pick machines with lower top payouts because at the same overall odds a lower top payout means I'm likely to win more frequently and play longer on the same bankroll; I never never never play the megajackpot slots, and if possible I avoid the machines that penalize the top jackpot when less than max coins are played. I don't expect to win, I just want to maximize my entertainment time for my money spent. Winnings and comps are a happy bonus. And free drinks. Yum. That's what I miss about Vegas when I go to the Indiana boat casinos: free pina coladas while playing.

By the way: my favorite slot machine ever is "100 or Nothing". You put a dollar in; put 3 7's on the line and you get $100. No other wins are possible. I don't play it continuously, though; I'll drop a dollar in as I pass or sit down and put $20 through it. I'm ahead on it lifetime so far.

[azanon]

Quote:

What do you mean by "odds are strongly in your favor"?

Other than the odds speaking for themselves, best way i know to explain it is to just give an example; *Lets say i used the technique I described above and decided I'd make exactly 28 spins.

The odds i'd lose 7 deep, and thus lose money (a lot of money) for the night are (.995) 4 (superscript)
Math scribble:
1-.05 (the chance of losing one series of 7 spins)= .995 *(chance i wouldnt lose in a series of 7 spins.

.995 x .995 x .995 x .995 = ~ 98% chance i'd walk out of the casino with a winning night using that technique.

The exact amount of money won though would just depend on how lucky i guess right each time.

It might sound too good to be true but it really isnt. *My wins wouldnt be near as large as the possible loss. *Over an infinite amount of time, the house wins per the 2 greens on the dial that the house gets.

(edit) hmm, think my math is wrong. *i'd have to use y to the x power so the real odds would be .995 to the 21 power (last 21 + 7=28 ) . *That still comes out to a 90% chance of not losing a series of 7.

/scratches head. ive been out of school too long.

[BigMoneyJim]

Quote:

.995 x .995 x .995 x .995 = ~ 98% chance i'd walk out of the casino with a winning night using that technique.

The 28-spin number you describe is closer to 95.6% (roughly 21 in 22), and the max upside is $140 with a max downside of $2540 (assuming you go back to $5 after losing 7 spins). The real odds are lower and more complex than this because the 7-spin series restarts whenever there's a win, and you'll probably quit playing after losing a 7th spin in a row or at least concede the $635 you put in so far and start at $5 again.

Also, if you lose any 7 spins in a row it wipes out 127 wins!

20/38 chance of losing a spin ~= 52.63%. Chance of losing any 7 spins in a row ~= 1.12%, about 1 in 89 trials, but a new trial starts after every winning spin--did your simulation account for this? The numbers get harder after that, but I bet your long term house advantage is still 5.26%--they get $5.26 for every $100 you play.

Quote:

It might sound too good to be true but it really isnt. My wins wouldnt be near as large as the possible loss.

"Odds strongly in your favor" is too good to be true. It's like saying crossing a busy street without looking is favorable because 95.6% of the time drivers will be alert and stop for you. Technically it's true but the upside is small with a large downside that will happen given any meaningful trial time.

I think the strategy is psychologically dangerous because it begins with the premise that losing is increasingly unlikely over a larger number of spins, so it would be too tempting after the 7th losing spin to say "8 losers in a row must be nigh impossible" when in fact it happens more than half the time immediately after 7 losing spins in a row.

The strategy minimizes the number of losses by increasing dramatically the size of the loss. I prefer smaller jumps down myself.

A correction to my earlier post where I said the doubling strategy and focused betting increases the size of winnings and losses: They trade the frequency of the wins for the size of the wins.

[Hyperborea]

This is one of those old betting methods that is based on the "fact" that you can't lose an infinite number of times. *Let's imagine that the game is actually a 50/50 game and that you will play an infinite number of rounds.

So, when you finally win on round n and have bet 2^(n-1) units you will win 2^(n-1) units. *You have lost 2^0 + 2^1 + ... + 2^(n-2) units along the way. *If you sum that up that is 2^(n-1) - 1 (the sum of a geometric series). *Now subtract that cost from your winnings and you have won the princely sum of one unit.

Now, let's apply it to a more real life example and what do we find. *Well, you can't bet an inifinite number of times so even if roulette was 50/50 you would top out at the casino betting limits or your self imposed 7 rounds. *So, one out of every 128 sets of runs (2^7) you would top out and lose. *That means that on average (i.e. over a lot of runs - a whole lot) you will make 127/128ths of a unit on every run.

We however know that roulette isn't even odds but instead has odds of 1/38th of winning and a payout of 36 (including the original bet). *You have a (37/38)^7 *100% = 83% chance of making it to the end of the 7 round without a win. *Yes, what we have here is not just a simple geometric series but the more general power series.

The terms of the power series will be:

(bet payoff factor) * (odds of winning in this round) * (bet)
= (bet payoff factor) * (odds of winning * odds of making it to this round) * (bet)
= (36) * [(1/38) * (37/38)^(n - 1)] * 2^(n - 1)

= (36/38) * (37/38)^(n - 1) * 2^(n - 1)

So, what we want is the sum of that from n = 1 to n = 7. *That will give us the expected value of your winnings on 7 rounds of roulette.

After 7 rounds
Expected Winnings = 105.2
Total Bet = 127
Expected loss every 7 rounds ~= 21.8

I computed this numerically. *If someone wants to do it analytically then go for it. *I wouldn't put too much money out on this "great" system.

[azanon]

Quote:

20/38 chance of losing a spin ~= 52.63%. Chance of losing any 7 spins in a row ~= 1.12%, about 1 in 89 trials, but a new trial starts after every winning spin--did your simulation account for this?

That's what was causing me problems with the math. *I realized i couldnt just simply take 4 groups of 7 because a new trial starts every time you win, so the math gets a little complicated for me. *If i sat down and spent some time on it, i'd eventually figure it out, but too lazy to do that.

Bottom line - its no get rich quick scheme. *If you ask me, casinos are illogical from a money standpoint so i only endulge for the entertainment value alone..... or for the buffets. * But if you can count cards (the original poster) by all means, have at it! *Sounds a lot like work to me though =p.

Azanon

[Hyperborea]

Ok, I've thought about this some more (the kind of things that run through my head while waiting for a bus or taking a shower) and I realize two mistakes in my previous work: I worked it out as though betting on a single number; and I didn't include the probability odds on the costs. *To make this easier to understand (at least for me) I'm also going to rearrange the layout somewhat.

Ok, I'm going to calculate expected return at each round of a run, ER. *These round number we will denote by n. *We will need to start by computing the return, R, achievable at round n. *Note that this is just the return for a specific round and not a summation of all rounds in a run.

R(n) = (winnings) - (costs)
= 2 * BET(n) - (SUM(1 to n)[BET(i)])
= BET(n) - SUM(1 to n-1)[BET(i)]

We know that (assuming that you start with a bet of 1 - could generalize by using a variable but it makes things messier),
BET(n) = 2^(n-1)

SUM(1 to n-1)[BET(i)] = 2^(n-1) - 1

So,

R(n) = 2^(n-1) - 2^(n-1) -1
= 1

Now we need to figure out the expected return, ER(n). *Again this will be the expected return of a single round and not the expected return of the whole sequence of rounds in a run.

ER(n) = (probability of happening) * (return)
= (probability of getting here) * ( probability of winning) * (return)
= (probability of not having won in all previous rounds) * (probability of winning) * (return)
= (20/38)^(n-1) * (18/38) * R(n)
= (20/38)^(n-1) * (18/38) * 1
= (20/38)^(n-1) * (18/38)

Finally, we need to sum the expected returns from all the rounds played in a particular run to get the Total Expected Return, TER. *In the case of this example we sum from 1 to 7.

TER = SUM(1 to 7)[ER(n)]
= SUM(1 to 7)[(20/38)^(n-1) * (18/38)]
= (18/38) * SUM(1 to 7)[(20/38)^(n-1)]
= (18/38) * (2.0875)
= 0.9888

So you will win slightly less than 1 betting unit every set of runs on average, very long sets of runs. *This is slightly less than you would get from a 50/50 game and is the result that is expected. *As the number of rounds in a run increases you begin to converge on 1.

[lets-retire]

All this math confuses the snot out of me. Ancedotally I can say every time I played roulette I went in with $50 and walked out with over $300. I got to the point I was using it as a second income, playing 3-4 times per week. This was with a European wheel so it didn't have the 00 on it. The method was simple wait until red/black, even/odd, or high/low hit 3 times then bet the other rarely did it go past 4 more deals or hit 0. I always started with the minimum bet and doubled it every deal I lost until the streak stopped, or I hit maximum bet.

After the night I would always figure the stats. This was about 10 years ago and I'm doing it from memory so if they're off a little, oh well. I would bet apporximately 10% of the deals and would win apporximately 95% of the bets I placed. The streak would change within 3 deals 98% of the time.

[BigMoneyJim]

On a European wheel the house advantage is about 2.7%. Long term they should take $2.70 for every $100 played. (Substitute eruo symbol as necessary.)

There's nothing wrong per se about the doubling strategy. Statistically in the long term it will have the same results as straight betting, and during play you'll have frequent small wins with occasional large losses. I guess I personally object because although I like frequent small wins I hate large sudden losses that can send you home early.

My objection is the promotion of it as a somehow superior winning strategy. When the downside hits, it hits big. If that's more exciting play for someone, fine, but it doesn't increase the payout odds.

As far as waiting for 3 consecutive hits then betting against it, if that feels more fun then do it, but--assuming a balanced wheel and fair game--past results have no correlation with future results.


Now for all the griping I do about this strategy I do plenty of unnumeric betting behavior myself. I play on hunches and bet on poor odds bets here and there because it's more fun for me. But I don't tell anyone else that doing that increases their odds of winning.

Incidentally, I've heard that if you count cards playing blackjack but mess up once or maybe twice that it's worse than playing without counting cards. On a wheel-gaguing device I expect the same disadvantage would happen if you didn't have the physics and readings just right.

EDIT: I've heard that baccarat has the lowest house advantage of all games (card counting blackjack aside), and unlike blackjack it requires no skill to play. i.e. you can't lose more money by playing poorly. I haven't played it myself, though. Craps pass line bets (and don't pass, come and don't come, and especially with odds bets) and blackjack are next lowest house advantage, depending upon the player's blackjack skill. "0" roulette is next. After that is probably 00 roulette and craps field bets, but I'm not sure. As far as electronic games I understand video poker has the best odds if the player is good. Worst odds in the place are the big wheel, then keno and some of the modified poker table games.

[sgeeeee]

I never gambled much myself. Once I understood the math behind the games I didn't feel the need to run the long term experiment and prove the mathematics correct.

But I know many people who visit Vegas to gamble on a regular basis. All of them routinely come out ahead. In fact, it is rare for one of them to return from a trip and to have done worse than "break even". They all "understand" the games they play and don't "get crazy" or make "stupid mistakes". They are always careful when they gamble. And at the end of the weekend or week, they walk away winners. It's amazing to me that Las Vegas doesn't go broke.