Logicians, Lateral Thinkers, Mathematicians, Bored Dopers...

You take a group of 100 people in to a casino, each with $100. Your task is to have a probability of over 50% that a majority of those 100 people will walk out of the casino with more money than when they went in. RULES:

  • The people you take in to the casino are allowed to obtain their money only from gambling.

  • The people are only allowed to play games found in standard casinos and abide by house rules.

  • No side betting with other gamblers, all money must be gambled stricty on games offered by casinos.

  • No Card Counting.

  • The people can lay any number of bets they want, on any game they want.

  • It is not required that everyone bet.

  1. pair up the gamblers.
  2. take a quarter, and flip heads or tails for each pair.
  3. whoever loses gives the $100 to the other person.
  4. the last pair wins the quarter as well as the $100.
  5. everyone leaves immediately.

This beats your conditions by 25 cents.

What, you say that’s not fair? Of course not. This game has negative “house odds.” The games in a casion all have positive house odds. There is no game you can play in a casino with anything close to even house odds.

This violates rules 2 & 3.

"* The people are only allowed to play games found in standard casinos and abide by house rules.

  • No side betting with other gamblers, all money must be gambled stricty on games offered by casinos."
    Wrong answer.

… and fails the question I asked anyway.

“Your task is to have a probability of over 50% that a majority of those 100 people will walk out of the casino with more money than when they went in”.

In your scenario, you would have 50 winners and 50 losers. That isn’t a majority of winners.

They could all use the Monte Carlo method (I do have the right method, right?) on roulette…

Everyone places a dollar on red.
Anyone who wins, leaves. Anyone who loses, doubles his previous bet and places it on red.

-tries to work it out-

Ok, I’ve got 96 people winning a dollar (on a 38 number wheel), and 4 losing $63…

The numbers are probably wrong, but I present my solution to you anyway. Does it work?

Your monte carlo method only works on one condition, that you have an infinite reserve of cash to double bets with. And you forgot one thing: once in a while you’ll hit 0 on the wheel.

I’ll clarify: There is no method that can fulfil the stated conditions. All games have a vigorish to the house. Even if you just played red/black on roulette, once in a while 0 would come up and the house wins all bets. The games are rigged, even if you play perfectly, you can beat other players but you can’t beat the house.

Hmmm. Isn’t it possible to play poker at Casinos? And if so, could these gamblers play purely amongst themselves? And if so, does that constitute a ‘side bet’ or not?

What you refered to as the “Monte Carlo” method is more properly called a “Martingale”. “Monte Carlo method” is a term usually used to describe a problem solving technique using random numbers and statistical sampling.

The problem with a Martingale lies in the fact that you have limited resources, as already observed, and that you are taking increasing risk for a fixed return. By the time you have doubled 8 times, you are risking $256 to win $1 overall.

We went through a whole series of these things with the now-banned MadHatter a while back. I’ll reiterate a reccomendation I made back then - “Lady Luck”, by Warren Weaver, is a very readable and informative elementary probability book for the layman, even though it was written many many years ago.

Oh, and I didn’t mean to suggest that MadHatter was banned because of those questions - he made an ass of himself on other fronts.

Even though Martingale is a loser in the long run, I think ryoushi is right, since the requirement is only that more than 50 people come out ahead by any amount, not that there be an overall gain. A player only needs to win once to come out ahead, but must lose several times in a row to lose all. With a $5 min bet, a player would have to lose four times in a row before being unable to continue the system. On a 38 number wheel, chance of losing four times in a row are (.53)^4 = .0789, so chance of winning at least once in four = 1 - .0789 or .9211. So if 100 people played successively (and each stopped immediately after winning or losing four in a row), 92 would win $5 and 8 would lose $75. House still makes $140 (I didn’t try to figure out the total amount bet, but would assume this is the expected house percentage of that total amount). House take can be reduced further by having everyone stop as soon as any player becomes the 51st winner.

Correct (though I didn’t stop to check your numbers), and you are actually just illustrating what is wrong with the Martingale as a “system”, though it might satisfy the OP’s rules, which do not require an overall gain. An 8/100 chance of losing $75 against a 92/100 chance of winning 5 are not odds most people would like.

I agree that ryoushi is correct. Due to the house advantage, total expected losses should exceed winnings, so the winners win a little and the losers lose a lot, but there is no reason why there can’t more winners than losers.

Each person could pick 35 different numbers at the roulette table (35 different numbers for one spin- do they allow that?, or 35 different spins). Most would make a small profit, and a few would lose 35 times the bet.

If rules permit betting on 35 different numbers for one spin, all 100 players could each play the same 35 numbers (if 100 people at one table is too much, they could pool their money for that one spin). There would then be 35/38= 92% chance of all 100 people making a profit. That might please the gamblers, and the casino should like it most of all.

The Black Jackal said:

Assuming a double zero roulette wheel, three spins will give you better than 85% chances.

Spin 1: 50 bet $1 on W, 50 bet $1 on R. 36/38 of the time 50 are +1 and 50 are -1.

Spin 2: Only the -1 guys bet, 25 on W and 25 on R. With probability (36/38)[sup]2[/sup] = 89.75%, after this spin there will be 50 +1, 25 0, and 25 -2.

Spin 3: This time pair up 24 of the 0 guys, 12 W, 12 R. With probability (36/38)[sup]3[/sup] = 85.03%, after this spin there will be 62 +1, 1 0, 12 -1, and 25 -2. Cash in.

I doubt this technique is optimal and of course I don’t deal with the 2/38 chances of zero coming up where you would still have chances of getting everyone back even and trying again.

Note the laws of probability have not been repealed. You’re merely rearranging the pattern of losses. Play long enough, you lose it all.

Assuming a roulette table with a $5 minimum bet per square and a table where up to 4 players may gamble, there is a 100% chance that at least 82 people can leave the casino with more money than they came in with.

Mono – I’m assuming you’re trying to be funny?

Martingale? I didn’t know it was called that :o

In my admittedly limited experience in Las Vegas casinos (and this was 13 years ago):

Yes, no, and it doesn’t matter, since #2 is no. The house ALWAYS had a dealer who played, and usually some shills among the players, too. I do NOT recommend trying to work as a shill, as most casinos will make you use your own money to gamble with, and if you are winning too much at one table, you’ll be given a signal to get up and move to another table.

Building on ryoushi’s method, assuming 18 red, 18 black, 0 and 00 spaces. Also, I assume winning on red or black wins the player the amount of the bet, and winning on 0 or 00 wins at least 5 times the bet. This also assumes a person can bet a space and a color on one wheel spin. (Never played roulette, so I don’t know the rules.)

Pair off the people, and for each pair, one bets $10 on black and $5 on 0, the other bets $10 on red, and $5 on 00. For each pair, one comes out ahead at least $5, and leaves. You’ve now got 50 people with $85. Pair them off, and for each pair, one bets $50 on black and $10 on 0, the other bets $50 on red, and $10 on 00.

At this point, 75 people are up, so there’s a 100 percent chance a majority are ahead.

I’m no logician, lateral thinker, or mathematician, so I guess that makes me bored. Not sure if this makes sense or would even work, but for what it’s worth, here goes.

Let’s play some roulette. We could try this at the crap table with Pass and Don’t Pass, but roulette offers a much simpler result.

We’re going to pair off our gamblers. For every spin of the wheel, one player in each pair will bet one $5 chip on Red, the other will bet one $5 chip on Black. (Red and Black are even money wagers.) After each spin is done, we should have one winner and one loser (though not in all cases–see below), and it is time for the next pair of gamblers to go to the table. Each pair only plays one spin of the wheel, and no gambler plays again until all others have played at least once.

In the first 38 spins, assuming the numbers work the way they should over time, after each of our 76 (so far) gamblers has played one spin of the wheel, we should have 36 winners, and 40 losers, since neither Red nor Black wins on Zero or Double Zero. Our 36 winners each have $105 now, and we send them home with more money than they came in with.

We still need 15 winners, however, if we are to fulfil the terms of the OP. (36 winners + 15 winners = 51 winners, a majority of the original 100.)

Left are 40 losers, each of whom have have $95; and 24 gamblers, each of whom have not yet played and thus have their original $100 stakes.

We’ll send the 24 who haven’t played yet to the table. They pair off and each wagers one $5 chip on Red or Black for the next 12 spins. For the following 20 spins, we pair off our original 40 losers, each of whom wagers $10 on Red or Black, as they did before. (I know this is basically following the Martingale system, but because we’re using pairs of disciplined players in rotation, and only need to achieve 51 winning players, we won’t need to continue the Martingale to its normally ruinous end.) To complete our series of 38 spins, we need 12 players to play 6 spins–we’ll use the ones who have only played once so far if we can, so they only have to wager $10.

Again, assuming the numbers work in the second set of 38 spins, we have 36 winners. No winner should have more than $105, but that still counts as a profit as far as the OP conditions go. With our original 36 winners, and the new group of 36, we have 72 winners. This is more than enough to fulfil the conditions of the OP (a majority of players being 51), and provides a little leeway in the cases where Zero and Double Zero come up a little more often than they should.

Hm. Using this method, I don’t know what the probability is that a majority would walk out winners, but it’s a thought anyway.