Explain this statistics/probability thing for me

[Ririon]

have anybody heard about some "cool series" to be lottery winner?, I mean does someone remember to get 5-10-15-20-25-30 in their national lottery contest?... it should have been some news if it has happened.. or something like 1-1-1-1-1-1... personally I can't.. but I would like to see if it has happened

Ospalh's example of 2-3-4-5-6-26 in the German lottery is the closest I've heard about. One interesting thing is that you can make money from this knowledge.

Since choosing common combinations (series, patterns, numbers under 12, numbers under 31 etc.) gives you a lower payout than average, avoiding these combinations would give you a higher payout than average. It is still not a reliable way of making money, though.

What you do is sell a lottery system that incorporates this. Lottery systems is a reasonably big business based almost entirely on ignorant costumers. "With our extra turbo system you only need 4 numbers, our system will cover for the rest". So selling a system that actually gives you higher winnings would even be ethical as long as you consider lotteries ethical.

Ririon

 

[a_unique_person]

Melendwyr and Art Vandelay: There's a weak law of large numbers and a strong law of large numbers. Both are true.

a somewhat related humorous link:
http://geomblog.blogspot.com/2004/06/strong-law-of-truly-misquoted.html

Frivolous Theorem of Arithmetic

http://mathworld.wolfram.com/contact/Almost all natural numbers are very, very, very large.

 

[a_unique_person]

I was interested to note that there are in fact two ways of looking at the list published by the lottery company of how many times each number has come up. Some pick the ones that have come up the most, to my surprise, since they are the ones that come up the most. I think the reasoning is just as 'valid' as the alternative, the numbers drawn the most are the least likely to come up next.

 

[gnome]

I was interested to note that there are in fact two ways of looking at the list published by the lottery company of how many times each number has come up. Some pick the ones that have come up the most, to my surprise, since they are the ones that come up the most. I think the reasoning is just as 'valid' as the alternative, the numbers drawn the most are the least likely to come up next.

Not really. If repetition is due to a flaw in the equipment that makes it slightly non-random, you could try to improve your odds a bit by playing the numbers it seems to favor. I belive this is mechanically possible, though I'm sure they tested it with a larger sample than players have available to them, so the point is probably moot.

 

[tsg]

Not really. If repetition is due to a flaw in the equipment that makes it slightly non-random, you could try to improve your odds a bit by playing the numbers it seems to favor. I belive this is mechanically possible, though I'm sure they tested it with a larger sample than players have available to them, so the point is probably moot.

I know some lotteries use multiple sets of balls chosen at random to eliminate that kind of thing.

 

[Jon.]

have anybody heard about some "cool series" to be lottery winner?, I mean does someone remember to get 5-10-15-20-25-30 in their national lottery contest?... it should have been some news if it has happened.. or something like 1-1-1-1-1-1... personally I can't.. but I would like to see if it has happened

I seem to remember that on Sept. 11 2001, the New York Lottery numbers were 9-1-1.

Ah yes, here it is. Of course, the odds are that about 1 of every thousand draws will draw the date (where it's possible, of course - you couldn't draw Dec. 24th, for instance, in a 3-digit lottery). Too bad it happened that day, setting off the conspiracy theorists and all.

 

[Jon.]

I would think that the only way to reduce the odds is to pick 7 numbers and premutate them , so if any 6 were correct you would win .Cost you 6 times as much of course .

In Canada, the lottery offers this kind of buy. You can pick up to 10 (I think) numbers and the computer will print you a ticket that plays all possible 6-number combinations. (We have a 6/49 system.) Of course, you actually have a better chance of winning something by playing more tickets with different numbers (3 or more correct wins) but with the combination, you would stand to win bigger if you did win (if you got all 6 on one combination, you would get several 5/6s, 4/6s, etc.).

Actually the correct term for this kind of thing is a combination , permutation implies order as well .
Best way to increase your chance of a win is to play once a week instead to twice a week and double your stake to £2 .Even better is to save the £2 each week and buy premium bonds instead .Trouble with that is that it takes a year before you can buy any and the odds are poor since only the interest is in the prize pool , but you do keep your money .
They don't call the lottery the ' Idiots Tax ' for nothing .

The only rational approach to the lottery is to calculate the odds of winning, multiply by the cost of a ticket, and only play if the jackpot is that high or higher. This, of course, ignores the possibility of having to share your jackpot, but it also ignores the smaller winning conditions, so I reckon those kind of balance out.

Oh, the other rational time to buy a ticket is when you've got a pool at work. Then not only do you have a chance of winning, but you also eliminate the possibility that you're the only one coming in to work next week!

 

[Art Vandelay]

Melendwyr and Art Vandelay: There's a weak law of large numbers and a strong law of large numbers. Both are true.

a somewhat related humorous link:
http://geomblog.blogspot.com/2004/06/strong-law-of-truly-misquoted.html

The strong law of large of large numbers presented in that link is not true. In fact, the weak law in that link is not true, either. It's the sample mean, not the sample sum, that has an increasing probability of being close to the true mean. In fact, if you take larger and larger samples, the probability of the sample sum being within a set distance of the expected sum actually goes to zero. For instance, if one million people were to each roll a die, and the value of each of those rolls were added up, the "expected value" would be 3,500,000, even though the probability of getting exactly 3,500,000 is extremely small.

 

[Ririon]

I seem to remember that on Sept. 11 2001, the New York Lottery numbers were 9-1-1.

Sept. 11 2002.

Anybody know when "running the numbers" started to be something the state did? Some movies and books (Malcolm X's autobiography comes to mind) mentions a similar three-digit lottery run by gangsters.

 

[CurtC]

The only rational approach to the lottery is to calculate the odds of winning, multiply by the cost of a ticket, and only play if the jackpot is that high or higher.

Sort of, but there's another factor. You could say that if the probability of winning is 1/16million, and the payout is more than $16M (after taxes), then the lottery is a good deal. What that doesn't take into account is that money has different incremental "value" to me depending on how much I already have. Right now, a million dollars to me has a very high value. However, if I already had 15 million, the extra million would not be nearly as significant. That basically means that the dollar I spend on my ticket is valued a lot higher to me than any one of those several million dollars that I might win, so that effect should skew the calculation, but I'm not quite sure how exactly.

 

[Jon.]

Sept. 11 2002.

Doh! I should learn to read one of these days.

Anybody know when "running the numbers" started to be something the state did? Some movies and books (Malcolm X's autobiography comes to mind) mentions a similar three-digit lottery run by gangsters.

I don't know about the US, but in Canada, I think it was in the late '70s to help pay off some of the debt from the 1976 Montreal Olympics. Of course, in the UK it was just a few years ago.

 

[Jon.]

Sort of, but there's another factor. You could say that if the probability of winning is 1/16million, and the payout is more than $16M (after taxes), then the lottery is a good deal. What that doesn't take into account is that money has different incremental "value" to me depending on how much I already have. Right now, a million dollars to me has a very high value. However, if I already had 15 million, the extra million would not be nearly as significant. That basically means that the dollar I spend on my ticket is valued a lot higher to me than any one of those several million dollars that I might win, so that effect should skew the calculation, but I'm not quite sure how exactly.

Fair enough, but that becomes unquantifiable pretty fast because generally there are too many options for how to spend the dollar. All I'm looking at is the point at which playing the lottery is no longer a tax on people who are bad at math.

 

[Rodney]

I was interested to note that there are in fact two ways of looking at the list published by the lottery company of how many times each number has come up. Some pick the ones that have come up the most, to my surprise, since they are the ones that come up the most. I think the reasoning is just as 'valid' as the alternative, the numbers drawn the most are the least likely to come up next.

Which lottery organizations publish lists of how many times each number has come up? Do you have any links?

 

[ynot]

A lottery is a very good investment for people that don't understand math

 

[tsg]

Which lottery organizations publish lists of how many times each number has come up? Do you have any links?

New Jersey's Pick6 has Winning Number History here.

 

[Rodney]

New Jersey's Pick6 has Winning Number History here.

Thanks very much.

 

[hodgy]

The only rational approach to the lottery is not to do it. If you must gamble then put your money on the horses - its more fun and the odds are much better.

The lottery is a self-imposed tax.

 

[tsg]

The only rational approach to the lottery is not to do it. If you must gamble then put your money on the horses - its more fun and the odds are much better.

Unless you're betting among your friends, it's still a negative EV.

 

[mgidm86]

I saw a website several years ago that listed the most commonly selected numbers (by the player). Don't know where they got the info. The theory is to pick the least popular numbers so that the chances of sharing the prize is minimized.

 

[homer]

When we studied Probability and Statistics some years ago we refered to it as sadistics . Very difficult and not a little boring .I prefer Pure Math.
It's amazing that something (probability that is ) that has a range of between 0 and 1 can be so complicated . 0 means no chance at all , 1 means it absolutely will . Probability of winning the UK lottery something like .00000007 (not very good ).
So when MR Spock tells the capt: that the probability is 10,000,000 to 1 he is giving odds not a measure of probability .