Do you purchase lottery tickets

[cowboy]

At any point in time I own several hundred. As far as I can tell, all they do is eat a lot, crap a lot, and make it difficult to avoid swearing in front of my children.

[cougjunkie]

Seems like everyone here knows someone that has hit a jackpot. So maybe your odds are better than you think.

My grandma sends me scratch-offs for my birthday every year. The most I have ever won is a free scratch off.

My dad made an Idaho run once when powerball was over 150 million. He is not a gambler at all so that surprised me, he went to the first gas station as you cross over from Utah to Idaho and the line was out the door he said. He said he had to drive another 40 miles in to the state before he found a gas station where the line was not at least 40-50 people long.

[Indy Coug]

Seems like everyone here knows someone that has hit a jackpot. So maybe your odds are better than you think.

Makes me wonder if that is due to the same mathematical reasons behind the surprising probabilities of shared birthdays.

[ame="http://en.wikipedia.org/wiki/Birthday_problem"]Birthday problem - Wikipedia, the free encyclopedia@@AMEPARAM@@/wiki/File:Birthday_paradox.png" class="image"><img alt="" src="http://upload.wikimedia.org/wikipedia/commons/thumb/5/56/Birthday_paradox.png/450px-Birthday_paradox.png"@@AMEPARAM@@commons/thumb/5/56/Birthday_paradox.png/450px-Birthday_paradox.png[/ame]

[cougjunkie]

Makes me wonder if that is due to the same reasons behind the surprising probabilities of shared birthdays.

Birthday problem - Wikipedia, the free encyclopedia

Should this go in your statistical anomaly thread? That is pretty crazy.

[wuapinmon]

Makes me wonder if that is due to the same mathematical reasons behind the surprising probabilities of shared birthdays.

Birthday problem - Wikipedia, the free encyclopedia

I've noticed on facebook that my "friends;" birthdays seem to run in clusters. There are sometimes when out of 500+ people, it's nobody's birthday for a few days.

[pelagius]

I've noticed on facebook that my "friends;" birthdays seem to run in clusters. There are sometimes when out of 500+ people, it's nobody's birthday for a few days.

You would expect clusters even if birthdays follow a distribution where each calender day is equally likely (1/365). How many clusters you would expect on average for each cluster length is a difficult problem to solve (much harder than the birthday problem) but it is easy to simulate (few lines of code). Here is a simulation of the number of clusters in a calendar year assuming a uniform distribution using 10,000 trials:

Friends = 500;

Summary Statistics
            Cluster=1   Cluster=2    Cluster=3   Cluster=4   Cluster=5
Mean        92.63590    23.35940     5.90650     1.47960     0.37150 
St. Dev      6.13179     4.74433     2.82944     1.50632     0.78061 
Min         70.00000     8.00000     0.00000     0.00000     0.00000 
Max        114.00000    46.00000    22.00000    10.00000     7.00000

With 500 friends, 3 day clusters of no birthdays isn't that unusual even if everything is random. 4-5 day clusters are rare. If you have 750 friends, clusters of 3 days without a birthday get rare assuming an uniform distribution (which, of course, doesn't reflect reality but its probably not a bad first order approximation):

Friends=750

            Cluster=1    Cluster=2   Cluster=3   Cluster=4   Cluster=5
Mean        46.58870     5.89380     0.74460     0.08790     0.01060 
St. Dev      5.36211     2.51526     0.96159     0.33434     0.12037 
Min         28.00000     0.00000     0.00000     0.00000     0.00000 
Max         68.00000    19.00000     8.00000     4.00000     3.00000

Most likely the first order effect you are observing is just plain ordinary randomness.

[Clark Addison]

I've noticed on facebook that my "friends;" birthdays seem to run in clusters. There are sometimes when out of 500+ people, it's nobody's birthday for a few days.

This is very similar to something I have noticed, which is that on facebook my "friends" always "forget" my birthday.

Jerks.

[wuapinmon]

You would expect clusters even if birthdays follow a distribution where each calender day is equally likely (1/365). How many clusters you would expect on average for each cluster length is a difficult problem to solve (much harder than the birthday problem) but it is easy to simulate (few lines of code). Here is a simulation of the number of clusters in a calendar year assuming a uniform distribution using 10,000 trials:

Friends = 500;

Summary Statistics
            Cluster=1   Cluster=2    Cluster=3   Cluster=4   Cluster=5
Mean        92.63590    23.35940     5.90650     1.47960     0.37150 
St. Dev      6.13179     4.74433     2.82944     1.50632     0.78061 
Min         70.00000     8.00000     0.00000     0.00000     0.00000 
Max        114.00000    46.00000    22.00000    10.00000     7.00000

With 500 friends, 3 day clusters of no birthdays isn't that unusual even if everything is random. 4-5 day clusters are rare. If you have 750 friends, clusters of 3 days without a birthday get rare assuming an uniform distribution (which, of course, doesn't reflect reality but its probably not a bad first order approximation):

Friends=750

            Cluster=1    Cluster=2   Cluster=3   Cluster=4   Cluster=5
Mean        46.58870     5.89380     0.74460     0.08790     0.01060 
St. Dev      5.36211     2.51526     0.96159     0.33434     0.12037 
Min         28.00000     0.00000     0.00000     0.00000     0.00000 
Max         68.00000    19.00000     8.00000     4.00000     3.00000

Most likely the first order effect you are observing is just plain ordinary randomness.

Yet another reason why you're very cool. Thanks for explaining this. My mind doesn't work this way, and I'm glad that there are people like you who can do this stuff, seemingly effortlessly.

[Indy Coug]

Yet another reason why you're very cool. Thanks for explaining this. My mind doesn't work this way, and I'm glad that there are people like you who can do this stuff, seemingly effortlessly.

Pelagius is truly the Master.

[pelagius]

Yet another reason why you're very cool. Thanks for explaining this. My mind doesn't work this way, and I'm glad that there are people like you who can do this stuff, seemingly effortlessly.

The mind is very good at picking out patterns ... no doubt there are lots of really good evolutionary reasons for this ... the downside is the mind isn't intituitively great with with random, particularly what random draws can looks like.

[I.J. Reilly]

Makes me wonder if that is due to the same mathematical reasons behind the surprising probabilities of shared birthdays.

Birthday problem - Wikipedia, the free encyclopedia

Reading this article has led me to invent a new party game.

Whenever I go to a party with 23+ people, I will take bets with guests on whether there are any two people in the room that share the same birthday for the first hour of the party. After that time, I will ask all to write down their birthday and then turn it in to an independent third party. Statistically, I should come out a winner.

If I show up at a place with 50+ I'll even advertise a great payout, like 10-1 if they win.

Now, I just need to get invited to enough parties where the people don't overlap so I can't be discovered.

[marsupial]

I bought a few on my 18th birthday because I could. I haven't bought any since.

[chrisrenrut]

A guy I that worked for me went on an Idaho run a few years back, and collected money from anyone that wanted him to buy them tickets. 4 or 5 people gave him money.

When he came back, one of the tickets he gave to one lady scratched off for $10,000.00. She was so excited, jumping up and down, screaming, making a big scene in the department. The guy just stood there smiling. It seems he had stopped at a novelty shop along the way and bought her a fake lottery ticket that was a guaranteed winnner. When he finally told her she was devastated. It was a terrible prank, that still doesn't get any laughs to this day.

[RC Vikings]

Reading this article has led me to invent a new party game.

Whenever I go to a party with 23+ people, I will take bets with guests on whether there are any two people in the room that share the same birthday for the first hour of the party. After that time, I will ask all to write down their birthday and then turn it in to an independent third party. Statistically, I should come out a winner.

If I show up at a place with 50+ I'll even advertise a great payout, like 10-1 if they win.

Now, I just need to get invited to enough parties where the people don't overlap so I can't be discovered.

You are one party animal.

[I.J. Reilly]

You are one party animal.

Hey, thanks! I'm always trying to think up stuff that won't get me invited to the next party happening