• Welcome to Valhalla Legends Archive.
 

"Wrong Number"

Started by iago, April 18, 2004, 08:31 PM

Previous topic - Next topic

iago

Somebody called my house looking for Ron (which, as you should all know, is my name).  After the embarassed pauses and "hello?"'s, it went like this:
Him: "Is this Ron?"
Me: "Yes"
"Who answered the phone?"
"My Dad, Jim"
"I think I have the wrong number"
"Well, this IS Ron"
"Ron Warren?"
"No"
"Oh"
"I can see why we were confused"
"haha, yeah"
"Well, have a good night.  Good luck :)"

And yes, I smiled.

That was a weird wrong number :)
This'll make an interesting test for broken AV:
QuoteX5O!P%@AP[4\PZX54(P^)7CC)7}$EICAR-STANDARD-ANTIVIRUS-TEST-FILE!$H+H*


Grok

A variation of the birthday paradox.

Adron


Grok

#3
Yes the other discussion we had about *some* two people (out of 23) having a 50% probability of the same birthday, where intuition normally tells them it takes half of 365 to have a 50% probability for a matching birthday.   In reality, it is true that you need 183 people to have a 50% probability of someone having *your* birthday.

In this case, some random person is calling for "Ron", and calls a "Ron".  At first it looks improbable.  What are the chances someone calling for "Ron" calls iago?  Pretty good, since we were not looking for that to happen specifically.  Had someone called for "Robert" and gotten me, or called "Brian" and gotten Skywing, instead of calling iago for "Ron", same thing.  But now that we're looking for a specific occurrence, the odds of someone looking for a specific name calling a specific person we pre-name, of the same name, is pretty slim.

Did I make any sense?

Akamas

Maybe it was me? Muahhaahha!
Quote from: Arta[vL] on August 14, 2006, 04:57 PM
Well, I want some too. Greedy Yoni should stop hogging it.

iago

Now if somebody had called here and asked for iago, I'd be worried.

The only problem is, there are 365(ish) possible birthdays.  But there are manymany more names than that.  I guess if you pick a random person, what are the odds that his name is Ron?
This'll make an interesting test for broken AV:
QuoteX5O!P%@AP[4\PZX54(P^)7CC)7}$EICAR-STANDARD-ANTIVIRUS-TEST-FILE!$H+H*


Grok

Quote from: iago on April 18, 2004, 10:26 PM
Now if somebody had called here and asked for iago, I'd be worried.

The only problem is, there are 365(ish) possible birthdays.  But there are manymany more names than that.  I guess if you pick a random person, what are the odds that his name is Ron?

See, you're looking at it from your own perspective.  That dramatically alters the statistics problem.  If you look at it as "the probability that when calling a wrong number, someone of the correct name will answer" your odds are far better than "the probability that someone having the wrong number will call iago, asking for iago."

iago

#7
Granted, that's part of it.  But the birthday thing is still a much smaller test area, 365, compared to names, which are thousands or more.  It's really hard to picture what you're talking about, but I do see what you mean and how this relates back to the birthday paradox.  I'll have to think more about this tomorrow :)

In case somebody missed our previous discussion, see
http://en.wikipedia.org/wiki/Birthday_paradox
and
http://mathworld.wolfram.com/BirthdayProblem.html
and
http://science.howstuffworks.com/question261.htm

<edit> I was just thinking, I should change the statement, "If I pick a random person, what are the odds that his name is Ron" to "If I pick a random person, what are the odds that he has a name?"
This'll make an interesting test for broken AV:
QuoteX5O!P%@AP[4\PZX54(P^)7CC)7}$EICAR-STANDARD-ANTIVIRUS-TEST-FILE!$H+H*


j0k3r

It differs ALOT for each person though, someone named michael is a lot more likely to get a call from someone asking for michael than someone named skylar to get a call asking for someone named skylar.
QuoteAnyone attempting to generate random numbers by deterministic means is, of course, living in a state of sin
John Vo

iago

Quote from: j0k3r on April 19, 2004, 05:44 AM
It differs ALOT for each person though, someone named michael is a lot more likely to get a call from someone asking for michael than someone named skylar to get a call asking for someone named skylar.

In co-op (that CS job-placement thingy I'm in), we have 9 people named michael and 2 females (out of 50 people).  Therefore, 18% of the population of the world is named Michael, and 4% of the population is female, and we can assume, with reasonable accuracy, that those don't overlap.
This'll make an interesting test for broken AV:
QuoteX5O!P%@AP[4\PZX54(P^)7CC)7}$EICAR-STANDARD-ANTIVIRUS-TEST-FILE!$H+H*


Adron

Quote from: Grok on April 18, 2004, 09:31 PM
Yes the other discussion we had about *some* two people (out of 23) having a 50% probability of the same birthday, where intuition normally tells them it takes half of 365 to have a 50% probability for a matching birthday.   In reality, it is true that you need 183 people to have a 50% probability of someone having *your* birthday.

In this case, some random person is calling for "Ron", and calls a "Ron".  At first it looks improbable.  What are the chances someone calling for "Ron" calls iago?  Pretty good, since we were not looking for that to happen specifically.  Had someone called for "Robert" and gotten me, or called "Brian" and gotten Skywing, instead of calling iago for "Ron", same thing.  But now that we're looking for a specific occurrence, the odds of someone looking for a specific name calling a specific person we pre-name, of the same name, is pretty slim.

Did I make any sense?

You make some sense, but I don't think what you're saying really applies. For birthdays, anyone having the same birthday as anyone else is the key concept.

For the phone case, I believe you'd have to include someone calling you and asking for Ron, Brian or Adron in addition to someone calling you and asking for Robert. Otherwise the probability won't be as high as for birthdays.

iago

Actually, that makes sense.  I knew something sounded weird there, but I couldn't quite get it.

If we take the odds of either me, Grok, or Skywing getting a call for either Ron, Rob, or Brian, then it would match a birthday problem, but since we're only counting somebody calling Grok looking for Rob, calling me looking for Ron, etc., it would be considerably less.

Thanks, Adron :)
This'll make an interesting test for broken AV:
QuoteX5O!P%@AP[4\PZX54(P^)7CC)7}$EICAR-STANDARD-ANTIVIRUS-TEST-FILE!$H+H*


Yoni

The Birthday Paradox is O(sqrt(n)). If there are a million different names, the probability that would happen is in the order of 1000 to 1.

Grok

Quote from: Yoni on April 19, 2004, 05:03 PM
The Birthday Paradox is O(sqrt(n)). If there are a million different names, the probability that would happen is in the order of 1000 to 1.

I wasn't implying the odds of birthdate paradox and wrong-number paradox were equivalent; just that they were the same type of statistical problem.