settle a statistics argument?

Discuss the language of the universe.

Moderators: Darkchilde, Calilasseia

settle a statistics argument?

 
 

settle a statistics argument?

#1  Postby CookieJon » May 14, 2011 10:39 am

So my friend is telling me that in games like lotto, you reduce your chances of winning if you limit your own number choices by only ever choosing, say, your family's birthdays. But surely since the winning numbers are random, you got the same chances of winning no matter what.

Who is right?
.

:cookies:
User avatar
CookieJon
THREAD STARTER
 
Name: Reverend Lionel
Posts: 6255
Male

Country: England
England (eng)

Re: settle a statistics argument?

#2  Postby z8000783 » May 14, 2011 10:49 am

CookieJon wrote:So my friend is telling me that in games like lotto, you reduce your chances of winning if you limit your own number choices by only ever choosing, say, your family's birthdays.

Does he explain why that is the case?

John
I don’t simply believe in miracles - I rely on them
z8000783
 
Name: WTF
Posts: 8323
Age: 58
Male

Greece (gr)

Re: settle a statistics argument?

#3  Postby chairman bill » May 14, 2011 10:59 am

If you enter only birthday dates, you limit your choice of number to between 1 & 31. That would limit your chances of winning where the choice of numbers rises to 49, for obvious reasons. You are relying on numbers from a more limited range than for other people in the lottery.
Image
The truth may be out there, but the lies are inside your head. Terry Pratchett
User avatar
chairman bill
 
Posts: 13052
Male

Country: UK
United Kingdom (uk)

Re: settle a statistics argument?

#4  Postby z8000783 » May 14, 2011 11:02 am

Makes no difference I'm afraid. If you only entered the numbers 1 to 6 every week you would still have the same chance of winning as everybody else.

John
I don’t simply believe in miracles - I rely on them
z8000783
 
Name: WTF
Posts: 8323
Age: 58
Male

Greece (gr)

Re: settle a statistics argument?

#5  Postby CookieJon » May 14, 2011 12:21 pm

chairman bill wrote:If you enter only birthday dates, you limit your choice of number to between 1 & 31. That would limit your chances of winning where the choice of numbers rises to 49, for obvious reasons. You are relying on numbers from a more limited range than for other people in the lottery.


That's what he said.
.

:cookies:
User avatar
CookieJon
THREAD STARTER
 
Name: Reverend Lionel
Posts: 6255
Male

Country: England
England (eng)

Re: settle a statistics argument?

#6  Postby CookieJon » May 14, 2011 12:23 pm

z8000783 wrote:Makes no difference I'm afraid. If you only entered the numbers 1 to 6 every week you would still have the same chance of winning as everybody else.

John


That's what I said.

He's further added that both are right and this is some sort of known paradox, but really?? I think he's thinking of something else.
.

:cookies:
User avatar
CookieJon
THREAD STARTER
 
Name: Reverend Lionel
Posts: 6255
Male

Country: England
England (eng)

Re: settle a statistics argument?

#7  Postby z8000783 » May 14, 2011 12:30 pm

How can both be right, either the probabilities change or they don't? In this case they don't.

There are a number of myths around probabilities, he's probably thinking of something else. For example, if you are playing roulette and 10 reds turn up in a row is it better to put you money on Black?

John
I don’t simply believe in miracles - I rely on them
z8000783
 
Name: WTF
Posts: 8323
Age: 58
Male

Greece (gr)

Re: settle a statistics argument?

#8  Postby virphen » May 14, 2011 12:33 pm

CookieJon wrote:So my friend is telling me that in games like lotto, you reduce your chances of winning if you limit your own number choices by only ever choosing, say, your family's birthdays. But surely since the winning numbers are random, you got the same chances of winning no matter what.

Who is right?


If you are talking about a single entry, you are, without doubt.

If you're talking about about a number of "lines" on a single ticket, your friend could be right, in a very obscure way. To take the simplest case, say you need to choose a minimum of 6 combinations of numbers when buying a ticket (I think that is the case here, although I've never actually bought one myself).

If you had (for the purposes of an example only) 7 numbers you were willing to choose, then obviously you're going to have a big degree of overlap on each entry - and the chances of the ticket being a winner (which I take as winning any prize at all, e.g. for having 4/6 numbers) are reduced. However if it did win, it would likely win on more than 1 line and thus win more money.

The chances of your friend referring to something like this though, are about the same as the chances of they winning the big prize.
/pedantry


z8000783 wrote:Makes no difference I'm afraid. If you only entered the numbers 1 to 6 every week you would still have the same chance of winning as everybody else.

John


True - but if the numbers 1-6 did somehow win, you'd take home less money, as apparently 1-6 is the single most popular entry in these games, so the prize would be shared with many more people.
Beatrice wrote:Did you guys see the golden shower after the game?
User avatar
virphen
 
Name: Lacerta ad Astra
Posts: 5156
Male

Country: Illuminatia
Jolly Roger (arr)

Re: settle a statistics argument?

#9  Postby virphen » May 14, 2011 12:35 pm

z8000783 wrote:For example, if you are playing roulette and 10 reds turn up in a row is it better to put you money on Black?


No, you should put it on red, as that would be fairly good evidence that the wheel is not playing true.
Beatrice wrote:Did you guys see the golden shower after the game?
User avatar
virphen
 
Name: Lacerta ad Astra
Posts: 5156
Male

Country: Illuminatia
Jolly Roger (arr)

Re: settle a statistics argument?

#10  Postby z8000783 » May 14, 2011 12:43 pm

virphen wrote:
z8000783 wrote:For example, if you are playing roulette and 10 reds turn up in a row is it better to put you money on Black?


No, you should put it on red, as that would be fairly good evidence that the wheel is not playing true.

mmmmm...possibly.

John
I don’t simply believe in miracles - I rely on them
z8000783
 
Name: WTF
Posts: 8323
Age: 58
Male

Greece (gr)

Re: settle a statistics argument?

#11  Postby susu.exp » May 14, 2011 6:22 pm

CookieJon wrote:So my friend is telling me that in games like lotto, you reduce your chances of winning if you limit your own number choices by only ever choosing, say, your family's birthdays. But surely since the winning numbers are random, you got the same chances of winning no matter what.

Who is right?


Let´s look at this:
Every combination is equally likely to be drawn, so this is uniformly random (random is a more general term - a uniform distribution with all possible outcomes equally probable is a special case). To elaborate further: If you look at the random variable "number of numbers you picked that are also drawn", it has the same distribution for any combination.

What your friend may have though of is something different:
When several people have the winning combination things like jackpots are split. This means that if you pick a common number combination, you are more likely not to be the only winner.
This in turn means that if you look at the random variable "winnings in money", it does not have the same distribution for any combination and in particular differes in it´s expectation value. Your expected winnings are lower, the more common your chosen numbers are.

Now treating playing the lotto for several rounds as a random walk, with increments of winnings in money - cost of tickets, picking less common numbers will lead to the expectation of the increments being closer to 0 (they of course are negative). This in turn means that the probability that you are ahead in money after a couple of plays does increase when you pick the latter.
Here´s a diagram based on the winnings of the lowest category (and hence with the broadest data) in the german lottery of the last year:
Image
These are scaled winnings for the numbes, with 1 being the generally expected win. Error bars give the root of the sample variance.
susu
susu.exp
 
Posts: 1047


Re: settle a statistics argument?

#12  Postby Berthold » May 15, 2011 1:01 pm

z8000783 wrote:Makes no difference I'm afraid. If you only entered the numbers 1 to 6 every week you would still have the same chance of winning as everybody else.

John

And you think you are the only one who has this idea? Oh no, there are hundreds!

Also diagonals, etc., are tipped by great numbers of players.
Berthold
 
Posts: 436
Age: 61
Male

Austria (at)

Re: settle a statistics argument?

#13  Postby z8000783 » May 15, 2011 1:10 pm

Berthold wrote:
z8000783 wrote:Makes no difference I'm afraid. If you only entered the numbers 1 to 6 every week you would still have the same chance of winning as everybody else.

John

And you think you are the only one who has this idea? Oh no, there are hundreds!

I wasn't proposing it a suggested selection.

John
I don’t simply believe in miracles - I rely on them
z8000783
 
Name: WTF
Posts: 8323
Age: 58
Male

Greece (gr)

Re: settle a statistics argument?

#14  Postby CarlPierce » May 18, 2011 9:16 am

Chance of winning is exactly the same whatever you pick....

But if you wish to reduce the number of people you share the jackpot with pick 6 numbers over 31 in a random fashion.
CarlPierce
RS Donator
 
Posts: 1380
Age: 47
Male

United Kingdom (uk)

Re: settle a statistics argument?

#15  Postby Berthold » May 18, 2011 4:51 pm

CarlPierce wrote:But if you wish to reduce the number of people you share the jackpot with pick 6 numbers over 31 in a random fashion.

I've heard that too many players are doing this by now. It was a good tactic before it was popularised.
Berthold
 
Posts: 436
Age: 61
Male

Austria (at)

Re: settle a statistics argument?

#16  Postby Someone » May 25, 2011 2:00 am

Your expected payout is affected by the likelihood of having to split prizes by choosing popular sets of numbers, but it's also affected when you have multiple entries by the overlaps between them (for all but the jackpot with an exact match). The bottom line, though, is that lotteries are a poor gambling choice, the payouts being pretty far off from a fair game. The main selling point for lotteries is the ratio between the jackpot and the cost of entry, with the utility of a single dollar or pound being, for most people who play, insignificant compared to the (minute) possibility of becoming instantly wealthy.
Here infrequently. I am a messenger and human. Haven't actually been given much info on who is responsible for sending the message. I will get back to you on that if I find out before you.
Someone
 
Name: Why-Me?/Just-Kidding-I
Posts: 1253
Age: 47

Country: USA
Myanmar (mm)

Re: settle a statistics argument?

#17  Postby Sityl » May 25, 2011 2:32 am

You'd have the same chance of winning if you picked numbers 1,2,3,4,5,6 as if you chose any random set of numbers, so I don't see how limiting your number pool will have any effect.
Stephen Colbert wrote:Now, like all great theologies, Bill [O'Reilly]'s can be boiled down to one sentence - 'There must be a god, because I don't know how things work.'


Image
User avatar
Sityl
 
Posts: 4454
Age: 30
Male

United States (us)

Re: settle a statistics argument?

#18  Postby Someone » May 25, 2011 4:41 pm

Somebody else already pointed out that you have less risk of having to split a jackpot by choosing a set of numbers not likely to also be chosen by somebody else.
Here infrequently. I am a messenger and human. Haven't actually been given much info on who is responsible for sending the message. I will get back to you on that if I find out before you.
Someone
 
Name: Why-Me?/Just-Kidding-I
Posts: 1253
Age: 47

Country: USA
Myanmar (mm)

Re: settle a statistics argument?

#19  Postby Sityl » May 25, 2011 4:44 pm

Someone wrote:Somebody else already pointed out that you have less risk of having to split a jackpot by choosing a set of numbers not likely to also be chosen by somebody else.


That's fine, but the OP's conversation wasn't about reward/risk ratio, it was about chances of wining.
Stephen Colbert wrote:Now, like all great theologies, Bill [O'Reilly]'s can be boiled down to one sentence - 'There must be a god, because I don't know how things work.'


Image
User avatar
Sityl
 
Posts: 4454
Age: 30
Male

United States (us)

Re: settle a statistics argument?

 
 

Re: settle a statistics argument?

#20  Postby Tyrannical » May 25, 2011 9:29 pm

The last thing you want is multiple winning lottery numbers that split the pot. Even if certain lucky numbers did come up more often, large amounts of peole will be playing those numbers thus reducing your share of the pot.
Good fences make good neighbors
User avatar
Tyrannical
Banned Troll
 
Posts: 6708
Male

United States (us)


Return to Mathematics

Who is online

Users viewing this topic: No registered users and 1 guest