A debate over some simple probabilities

For the discussion of math. Duh.

Moderators: gmalivuk, Moderators General, Prelates

Who is right about this simple probability problem?

You are right, sir.
9
24%
You are mistaken. Your friends are absolutely correct.
24
65%
You're both wrong! I'll tell you why!
0
No votes
You're both right. Let me explain...
4
11%
 
Total votes: 37

User avatar
hencethus
Posts: 24
Joined: Thu May 31, 2007 4:25 am UTC
Location: Port Orange, FL
Contact:

A debate over some simple probabilities

Postby hencethus » Fri Jun 01, 2007 11:12 pm UTC

I happened to visit the same restaurant twice in the same week, accompanied by different people each time. Both times it was someone else (besides me) that chose where we sat. And both times we happened to sit in the same place.

On the second visit I mentioned that we were sitting in the same place that I'd sat the last time I'd been there, and someone else asked (probably rhetorically), "What are the odds of that?"* So I said that assuming there were 50 places to sit, and assuming that our seating selection was truly random with an equal chance of sitting at any of the 50 places (which of course isn't true), then the odds were 1 in 2500.

Immediately everyone protested. My companions insisted that the odds were in fact 1 in 50 since the initial seating choice was "a given" (whatever that means). One of my friends even used an analogy, saying that by my logic, if he flipped a coin I would have a 1 in 4 chance of guessing correctly on which side it would land.

I tried to explain that while the odds of choosing correctly were 1 in 2, the odds of him flipping heads and me randomly guessing heads specifically was only 1 in 4 (and that the same was true for tails).

Another one of my companions again insisted that I was wrong, and suggested that I go back to school and take a statistics course.

So who's right here, guys?

*Addendum: We're talking about the odds of sitting in that specific place twice, which I thought was made clear by my explanation of the coin analogy, but apparently I'm biased. :P
Last edited by hencethus on Fri Jun 01, 2007 11:48 pm UTC, edited 1 time in total.

User avatar
Mighty Jalapeno
Inne Juste 7 Dayes I Wille Make You A Hero!
Posts: 11265
Joined: Mon May 07, 2007 9:16 pm UTC
Location: Prince George In A Can
Contact:

Postby Mighty Jalapeno » Fri Jun 01, 2007 11:15 pm UTC

You're right, and they're probably from Texas.

User avatar
Woxor
Posts: 506
Joined: Mon May 07, 2007 11:28 pm UTC

Postby Woxor » Fri Jun 01, 2007 11:27 pm UTC

It depends on what the exact question is, but I'm afraid I think you're wrong.

If the question were, "What are the odds of sitting in THIS place twice in a row?" you'd be right. But as you've phrased it, they asked the odds of sitting in "the same place as last time" the second time you come. The odds of the latter happening are indeed 1 in 50, since there are 2500 possible ways to be seated twice, and 50 possible ways to be seated at the same place twice.

But what do I know, I'm from Texas. Maybe math is different down there.

User avatar
Vaniver
Posts: 9422
Joined: Fri Oct 13, 2006 2:12 am UTC

Postby Vaniver » Fri Jun 01, 2007 11:32 pm UTC

You're wrong.

The probability is sitting in the same spot.

The specific place the first spot is in doesn't matter- just if the second is the same. Just like it doesn't matter whether you guessed heads or tails; just whether or not you were right. 1/4 (guessed heads and it was heads) + 1/4 (guessed tails and it was tails) = 1/2 (guessed what it was); you can do a similar addition for each table, or just accept the principle that the first is a "given", and you're just comparing if they are the same.

[edit]I grew up in Texas as well. Odd.
I mostly post over at LessWrong now.

Avatar from My Little Pony: Friendship is Magic, owned by Hasbro.

User avatar
hencethus
Posts: 24
Joined: Thu May 31, 2007 4:25 am UTC
Location: Port Orange, FL
Contact:

Postby hencethus » Fri Jun 01, 2007 11:33 pm UTC

Woxor wrote:If the question were, "What are the odds of sitting in THIS place twice in a row?" you'd be right.


That is indeed the question. Otherwise I wouldn't have bothered with the coin flipping explanation.

User avatar
Woxor
Posts: 506
Joined: Mon May 07, 2007 11:28 pm UTC

Postby Woxor » Fri Jun 01, 2007 11:35 pm UTC

hencethus wrote:That is indeed the question. Otherwise I wouldn't have bothered with the coin flipping explanation.

Then you should probably change this:
hencethus wrote:On the second visit I mentioned that we were sitting in the same place that I'd sat the last time I'd been there, and someone else asked (probably rhetorically), "What are the odds of that?"

... because that's not the same question.

User avatar
Vaniver
Posts: 9422
Joined: Fri Oct 13, 2006 2:12 am UTC

Postby Vaniver » Fri Jun 01, 2007 11:38 pm UTC

As well, the question "will an event be repeated twice" is generally more frequently asked than "will this event be repeated twice," unless there is something to differentiate "this event" from other events. If it's just the coincidence of sitting at the table twice, the first is what people will expect. If it's the coincidence of being seated by the kitchen, or bathroom, or wherever, twice, they might expect the second.
I mostly post over at LessWrong now.

Avatar from My Little Pony: Friendship is Magic, owned by Hasbro.

User avatar
hencethus
Posts: 24
Joined: Thu May 31, 2007 4:25 am UTC
Location: Port Orange, FL
Contact:

Postby hencethus » Fri Jun 01, 2007 11:43 pm UTC

Woxor wrote:... that's not the same question.

Isn't it? I think it's phrased rather ambiguously. Anyhow, although I didn't include it in my post, I did explain to my friends that I was talking about the odds of sitting in that specific spot both times. They still insisted that the original seating selection was a given regardless.

I'll make an addendum to the original post, though.

User avatar
Woxor
Posts: 506
Joined: Mon May 07, 2007 11:28 pm UTC

Postby Woxor » Fri Jun 01, 2007 11:45 pm UTC

Yeah, I mean it would be pretty silly if someone asked, "What are the odds that I will get heads when I flip a coin?" and you replied with "About 1 in a trillion because there are a billion places to stand and 500 things you could do, so the chance of you standing in that exact spot and flipping a coin to begin with are about 1 in 500 billion." They weren't asking what the probability of the universe being in a given state is, they were asking what the chances that a coincidence like that would occur.

EDIT:
hencethus wrote:I did explain to my friends that I was talking about the odds of sitting in that specific spot both times. They still insisted that the original seating selection was a given regardless.

Okay, if they still disagreed with you, then you bested them mathematically, but I agree with Vaniver's post. vvvvvvvvv
Last edited by Woxor on Fri Jun 01, 2007 11:48 pm UTC, edited 3 times in total.

User avatar
Vaniver
Posts: 9422
Joined: Fri Oct 13, 2006 2:12 am UTC

Postby Vaniver » Fri Jun 01, 2007 11:46 pm UTC

I did explain to my friends that I was talking about the odds of sitting in that specific spot both times. They still insisted that the original seating selection was a given regardless.
Well, that still seems ambiguous. I mean, unless it's special for a reason other than you sitting there, asking about the odds of sitting at table 16 twice in a row is a silly question. Since the odds are high you would have commented on a repeated seating regardless of where you sat, your friends probably interpreted the question the way we did, regardless of you rephrasing it to a less meaningful question.
I mostly post over at LessWrong now.

Avatar from My Little Pony: Friendship is Magic, owned by Hasbro.

User avatar
hencethus
Posts: 24
Joined: Thu May 31, 2007 4:25 am UTC
Location: Port Orange, FL
Contact:

Postby hencethus » Fri Jun 01, 2007 11:52 pm UTC

Vaniver wrote:Well, that still seems ambiguous. I mean, unless it's special for a reason other than you sitting there, asking about the odds of sitting at table 16 twice in a row is a silly question. Since the odds are high you would have commented on a repeated seating regardless of where you sat, your friends probably interpreted the question the way we did, regardless of you rephrasing it to a less meaningful question.

Agreed. Eventually I conceded that we were both right depending on how the problem was approached, but they made no such concession. The initial seating was "a given." Period.

User avatar
Vaniver
Posts: 9422
Joined: Fri Oct 13, 2006 2:12 am UTC

Postby Vaniver » Fri Jun 01, 2007 11:57 pm UTC

The initial seating was "a given." Period.
Well, was there anything special about the table, other than that you sat there?

You're starting the thought process from before the first time you went to the restaurant, when you should be starting it the second time you enter the restaurant.
I mostly post over at LessWrong now.

Avatar from My Little Pony: Friendship is Magic, owned by Hasbro.

User avatar
hencethus
Posts: 24
Joined: Thu May 31, 2007 4:25 am UTC
Location: Port Orange, FL
Contact:

Postby hencethus » Sat Jun 02, 2007 12:03 am UTC

Vaniver wrote:Well, was there anything special about the table, other than that you sat there?

It was right by the trash bins, which I why I remembered that it was the same table in the first place. I hate sitting by the trash.

Vaniver wrote:You're starting the thought process from before the first time you went to the restaurant, when you should be starting it the second time you enter the restaurant.

I don't think there's a time from which I should start my thought process. Just as there's no reason that I shouldn't ask the question, "If we both flip a coin now, what are the odds that both coins will land on heads?" The answer is 1 in 4. It would be silly to say, "No, you shouldn't ask that question. The answer is 1 in 2 because the coins could also both land on tails."

User avatar
Woxor
Posts: 506
Joined: Mon May 07, 2007 11:28 pm UTC

Postby Woxor » Sat Jun 02, 2007 1:19 am UTC

hencethus wrote:I don't think there's a time from which I should start my thought process. Just as there's no reason that I shouldn't ask the question, "If we both flip a coin now, what are the odds that both coins will land on heads?" The answer is 1 in 4. It would be silly to say, "No, you shouldn't ask that question. The answer is 1 in 2 because the coins could also both land on tails."

I think it makes sense to say what the natural question to consider is, though. Ex: my 1-in-a-trillion coin-flipping example above.

User avatar
3.14159265...
Irrational (?)
Posts: 2413
Joined: Thu Jan 18, 2007 12:05 am UTC
Location: Ajax, Canada

Postby 3.14159265... » Sat Jun 02, 2007 1:21 am UTC

"What are the odds of that?"


Well it was surprising to you that you sat at the same table twice not that you sat at table (49 or w.e.) twice.

So they are right.

No concessions!
"The best times in life are the ones when you can genuinely add a "Bwa" to your "ha""- Chris Hastings

User avatar
hencethus
Posts: 24
Joined: Thu May 31, 2007 4:25 am UTC
Location: Port Orange, FL
Contact:

Postby hencethus » Sat Jun 02, 2007 1:34 am UTC

Woxor wrote:I think it makes sense to say what the natural question to consider is, though. Ex: my 1-in-a-trillion coin-flipping example above.

Natural to whom? Anyway, although I understand the point you're making, I don't think your reductio ad absurdum is especially relevant here. Obviously the question was quite natural to me.

3.14159265... wrote:Well it was surprising to you that you sat at the same table twice not that you sat at table (49 or w.e.) twice.

It was surprising that "we were sitting in the same place that I'd sat the last time" (exactly how it was phrased in the original post), which was a very specific place (table 49 or whatever).

The_Spectre
Posts: 33
Joined: Sun Dec 03, 2006 2:18 pm UTC

Postby The_Spectre » Sat Jun 02, 2007 1:35 am UTC

I'll tell you a shocking story. This morning, I parked my bike next to a car that had the following license plate number:

JF-24-UE

Amazing, isn't it? I mean, the odds of that exact license plate number are really really tiny!

User avatar
hencethus
Posts: 24
Joined: Thu May 31, 2007 4:25 am UTC
Location: Port Orange, FL
Contact:

Postby hencethus » Sat Jun 02, 2007 1:47 am UTC

The_Spectre wrote:I'll tell you a shocking story. This morning, I parked my bike next to a car that had the following license plate number:

JF-24-UE

Amazing, isn't it? I mean, the odds of that exact license plate number are really really tiny!

Even more shocking: I was playing a game of hearts, and I was dealt 13 random cards: J(c) 3(d) K(d) A(d) 2(s) 8(s) J(s) Q(s) K(s) A(s) 3(h) 10(h) Q(h).

The odds of being dealt that exact hand are something like 1 in 600 billion. Amazing! I can hardly believe it happened!

User avatar
Darth Eru
Posts: 77
Joined: Fri May 25, 2007 10:52 pm UTC

Postby Darth Eru » Sat Jun 02, 2007 2:52 am UTC

I agree with the people saying that it was understandable your friends misunderstood the question. The first time you sat at that particular table, it didn't really matter which table you sat at. It only becomes significant once you sit at the same table later on, so the significant question is what the odds are of sitting at any table twice, not a particular table twice.

Consider this: The chances of sitting at a particular table twice are the same as those of sitting at any two particular tables, whereas the chances of sitting at the a table twice are very small compared to the chance of sitting at two different tables.
Signature:
This is a block of text that can be added to posts you make. There is a 255 character limit

There are 2 types of people in the world: those who know hexadecimal, and those who don't.

Dvorak > QWERTY

User avatar
bitwiseshiftleft
Posts: 295
Joined: Tue Jan 09, 2007 9:07 am UTC
Location: Stanford
Contact:

Postby bitwiseshiftleft » Sat Jun 02, 2007 2:59 am UTC

A mathematician is flying from America to Europe, and when his bag is searched at security, the screener finds a bomb in it! He angrily asks the mathematician for an explanation. "Well," the mathematician explains, "I wasn't going to set it off, and what are the odds there being two bombs on the same flight?"

User avatar
Woxor
Posts: 506
Joined: Mon May 07, 2007 11:28 pm UTC

Postby Woxor » Sat Jun 02, 2007 3:05 am UTC

hencethus wrote:It was surprising that "we were sitting in the same place that I'd sat the last time" (exactly how it was phrased in the original post), which was a very specific place (table 49 or whatever).

This clarification only makes it worse for your case. The chances of "sitting in the same place that you sat last time" are 1 in 50.

hencethus wrote:Natural to whom? Anyway, although I understand the point you're making, I don't think your reductio ad absurdum is especially relevant here. Obviously the question was quite natural to me.

The reason it was more natural to ask the 1 in 50 question is that the past event had already happened, and the second seating provoked the question. Like someone else said above, it's the fact that you sat in the same place that was of interest, not that you sat at table "49" in particular twice. You're right about your probability, but I think it would have been more reasonable to interpret the question in the 1 in 50 sense.

And my 1 in a trillion example makes a valid point: it's important to consider what part of the problem is relevant. The penny flipping is what we care about, not where you are or how you do it. Similarly, the coincidence of sitting at the same table is what we care about, not which table it was.

User avatar
3.14159265...
Irrational (?)
Posts: 2413
Joined: Thu Jan 18, 2007 12:05 am UTC
Location: Ajax, Canada

Postby 3.14159265... » Sat Jun 02, 2007 5:35 am UTC

It was surprising that "we were sitting in the same place that I'd sat the last time" (exactly how it was phrased in the original post), which was a very specific place (table 49 or whatever).


Would you have been just as surprised if you sat at table 48 or 47 twice, if yes they are right.

If that table is special in that it was a bad one, or always served by the hot waitress or something else then you are.

I am sure its the first case, so you are wrong, concede 100%! now!
"The best times in life are the ones when you can genuinely add a "Bwa" to your "ha""- Chris Hastings

User avatar
Woxor
Posts: 506
Joined: Mon May 07, 2007 11:28 pm UTC

Postby Woxor » Sat Jun 02, 2007 5:38 am UTC

3.14159265... wrote:If that table is special in that it was a bad one, or always served by the hot waitress or something else then you are.

He said it was by the trash, so he's got a leg to stand on. I just think his friends had two legs. :D

User avatar
Zohar
COMMANDER PORN
Posts: 8533
Joined: Fri Apr 27, 2007 8:45 pm UTC
Location: Denver

Postby Zohar » Sat Jun 02, 2007 8:53 am UTC

I think you got a good enough answer already (if you asked what are the chances to pick this specific table twice, it's 1/2500, if it's to sit twice at the same table it's 1/50).

I just thought to mention that everyone assumes a uniform distribution on the tables (not that it matters for the sake of the argument). Obviously they aren't. Some people have favorite tables, some don't want to sit in the center of the room, the good areas might get crowded and noisy so you'd choose a "worse" table etc. I guess that would be interesting to see. The different distributions according to different people, kind of like the choose the right urinal test.
Mighty Jalapeno: "See, Zohar agrees, and he's nice to people."
SecondTalon: "Still better looking than Jesus."

Not how I say my name

User avatar
Yakk
Poster with most posts but no title.
Posts: 11128
Joined: Sat Jan 27, 2007 7:27 pm UTC
Location: E pur si muove

Postby Yakk » Sat Jun 02, 2007 11:26 pm UTC

So you walk into the restaurant, and your friend chooses a table.

You sit down and say "man, we came into the restaurant, and we sat at the table my friend chose! What are the odds of that? You know, assuming we sit where my friend chooses."

The transformation of the situation, where you change the table on both visit 1 and visit 2 to the same different table, generates a situation where saying "what are the odds we sit at the same table twice?" is equally valid.

When one talks about how unlikely something is, you should talk about how unlikely it is up to such automorphisms, and up to "even more unlikely things" (which is a hard part).

"What are the odds I would have to sit next to the trash two visits in a row?"
vs
"What are odds we sit in the same spot two days in a row?"

The first one is 1/2500 (assuming only one set of seats next to the trash)
The second one is 1/50.

User avatar
__Kit
Posts: 1576
Joined: Tue May 08, 2007 5:12 am UTC
Location: 16/M/NZ
Contact:

Postby __Kit » Sat Jun 02, 2007 11:31 pm UTC

I agree with your friends sorry, I have often pondered this. If I roll a dice it doesn't matter WHAT it is (1-6, lets say 3) th chance of me getting the same number as the first time is 1 in 6
=]

User avatar
Blatm
Posts: 638
Joined: Mon Jun 04, 2007 1:43 am UTC

Postby Blatm » Mon Jun 04, 2007 2:39 am UTC

To repeat what's been said about a dozen times, it all depends on what you mean. Your friends are right on a practical level, but you're right on a technical level, and being technicaly correct is the best kind of correct. I voted in your favour, but only because the poll isn't going to matter. If it were, I would probably vote for you both being right. Actualy, I'm very surprised that only one person has picked option 4 up to now.

Have fun with your pointless bickering. :)

User avatar
Woxor
Posts: 506
Joined: Mon May 07, 2007 11:28 pm UTC

Postby Woxor » Mon Jun 04, 2007 2:56 am UTC

Blatm wrote:Your friends are right on a practical level, but you're right on a technical level, and being technicaly correct is the best kind of correct.

"Dammit hencethus! They've got us surrounded! I thought you said the odds of them bombing Pearl Harbor twice were one in 2500, over!"

"Well, there are 50 U.S. states, so 50 squared is 2500, over!"

"But we already knew that Pearl Harbor was bombed once! Oh God, I'm going down! Mayday!"

"You didn't specify what the exact question-" *BOOM*

User avatar
Gelsamel
Lame and emo
Posts: 8237
Joined: Thu Oct 05, 2006 10:49 am UTC
Location: Melbourne, Victoria, Australia

Postby Gelsamel » Mon Jun 04, 2007 7:15 am UTC

Jesus fucking christ, if you explained that you're talking about on both occasions and they still insisted it was 1 in 50 then they're retarded.

The chance of getting 2 heads in a row is 1/4 do a fucking probability tree.

Code: Select all

        Heads 1/4
 Heads<
/       Tails 1/4
\       Heads 1/4
 Tails<
        Tails 1/4






However the probability of getting the same face of the coin as last time is 1/2.

Code: Select all

        Heads 1/2
 Heads<
        Tails 1/2




If you made it clear you were talking about the first situation and they dispute that then they're idiots. And There is no use in trying to reason with them.
"Give up here?"
- > No
"Do you accept defeat?"
- > No
"Do you think games are silly little things?"
- > No
"Is it all pointless?"
- > No
"Do you admit there is no meaning to this world?"
- > No

User avatar
scowdich
The Hedgehog
Posts: 771
Joined: Tue May 22, 2007 4:55 am UTC
Location: University of Illinois (Urbana-Champaign)
Contact:

Postby scowdich » Mon Jun 04, 2007 7:43 am UTC

Woxor wrote:"Dammit hencethus! They've got us surrounded! I thought you said the odds of them bombing Pearl Harbor twice were one in 2500, over!"

"Well, there are 50 U.S. states, so 50 squared is 2500, over!"

"But we already knew that Pearl Harbor was bombed once! Oh God, I'm going down! Mayday!"

"You didn't specify what the exact question-" *BOOM*


For the record, there were only 48 U.S. states when Pearl Harbor was bombed (the first time), resulting in a "technical probability" of 1/2400, not 1/2500. Or perhaps since Hawaii wasn't a state then, using a base probability of 1/48 won't work...hmm.

User avatar
evilbeanfiend
Posts: 2650
Joined: Tue Mar 13, 2007 7:05 am UTC
Location: the old world

Postby evilbeanfiend » Mon Jun 04, 2007 10:40 am UTC

Gelsamel wrote:Jesus fucking christ, if you explained that you're talking about on both occasions and they still insisted it was 1 in 50 then they're retarded.

The chance of getting 2 heads in a row is 1/4 do a fucking probability tree.

Code: Select all

        Heads 1/4
 Heads<
/       Tails 1/4
\       Heads 1/4
 Tails<
        Tails 1/4





i think their point is that the first toss has already happened so we know exactly how it came out .'. assigning a probability other than 1 to this event is meaningless.
in ur beanz makin u eveel

User avatar
Gelsamel
Lame and emo
Posts: 8237
Joined: Thu Oct 05, 2006 10:49 am UTC
Location: Melbourne, Victoria, Australia

Postby Gelsamel » Mon Jun 04, 2007 11:13 am UTC

evilbeanfiend wrote:i think their point is that the first toss has already happened so we know exactly how it came out .'. assigning a probability other than 1 to this event is meaningless.


No, "We know how the first decision came out therefore the answer is 1/2" is the 2nd tree I came up with. Note how Heads starts the chain, because it already happened.

The OP said he explained that he was evaluating the hypothetical "What are the chances?" and they still disagreed.

And the hypothetical isn't meaningless, otherwise we could answer questions like "Pr(2 heads in a row) = 1/2 because after the first I'll know the side"

Or why not just flip 2 coins and if it's Not the same Say Pr=0, or if it is say Pr=1?

Or http://xkcd.com/c221.html

To the question "What is the chance of sitting in the same seat twice in a row" IS actually 1/50. But the question "What is the chance of sitting in this particular seat twice is a row" is 1/2500.

ie.


"What is the chance of getting the same side twice in a row."

Code: Select all

        Heads 1/4 <------
 Heads<
/       Tails 1/4  AND
\       Heads 1/4
 Tails<
        Tails 1/4 <------

= 1/2



But "What is the chance of getting this particular side twice in a row"

Code: Select all

        Heads 1/4 <------
 Heads<
/       Tails 1/4    OR
\       Heads 1/4
 Tails<
        Tails 1/4  <-------

=1/4



The OP was answering the 2nd.

Sure there was obviously a heap of ambiguity which explains why they disagreed initially (really they should've realised this ambiguity and asked the OP to clarify).

But to continue to disagree after the OP clarified is ridiculous. Unless you disagree with the method of mathematics altogether in which case you better have a damn good reason.
"Give up here?"
- > No
"Do you accept defeat?"
- > No
"Do you think games are silly little things?"
- > No
"Is it all pointless?"
- > No
"Do you admit there is no meaning to this world?"
- > No

User avatar
evilbeanfiend
Posts: 2650
Joined: Tue Mar 13, 2007 7:05 am UTC
Location: the old world

Postby evilbeanfiend » Mon Jun 04, 2007 12:26 pm UTC

yes but i think they are trying to point out that the hypothetical question answered by the op is always hypothetical
at the time the question was asked because 1 outcome is known.

i.e we have these possible questions

q1. given 2d50 what is the chance of getting 50 on each before we roll
q2. given 2d50 what is the chance of getting the same on each before we roll
q3. given 2d50, and the first die has already been thrown and is known to be 50, what is the chance of getting 50 on both.

i think they are interpreting it as q3, and given that the other two are purely theoretical questions as they don't considered the fact that one roll has already been observed .'. i think their position that they are right and the op is wrong is defensible even though i think the op being right is also a defensible position. i.e. it doesn't just depend on what the question was but when it was asked
in ur beanz makin u eveel

User avatar
Gelsamel
Lame and emo
Posts: 8237
Joined: Thu Oct 05, 2006 10:49 am UTC
Location: Melbourne, Victoria, Australia

Postby Gelsamel » Mon Jun 04, 2007 1:12 pm UTC

Right, they were interpreting it differently, yet when the OP explained what he was answering they still disagreed.

And I wouldn't give them the benefit of the doubt saying that they were 'arguing the OP's question is always hypothetical' all the OP has said is that they disagreed with his calculation of HIS question (Question 1 in your list).

And the OP's question isn't always hypothetical: "If we choose one of the 50 seats totally at random then within the next two times we go to the restaurant what is the probability that we end up at the same seat both times?"

The fact of the matter is Question 3 is NOT useful, it tells us NOTHING.

Why? It sets any events that has happened = 1. And of course that's true, but does it tell us anything? No. "What is the chance that we got the exact same seat as last time?" "100%, since we did get it again"

"Wow I won the lotto!"
"Wow, what're the odds of that?!"
"100% chance, since I did win it!"
Of course once again in hindsight this is perfectly correct, however it tells you shit all.

The only meaningful questions are 1 and 2 in your list because they actually tell you something.

So unless someone is looking for a totally meaningless answer when they ask "What is the chance of that?" Or "What are the odds?", then the question is either 1 or 2, and which one of those two it is depends on the situation.
"Give up here?"
- > No
"Do you accept defeat?"
- > No
"Do you think games are silly little things?"
- > No
"Is it all pointless?"
- > No
"Do you admit there is no meaning to this world?"
- > No

User avatar
evilbeanfiend
Posts: 2650
Joined: Tue Mar 13, 2007 7:05 am UTC
Location: the old world

Postby evilbeanfiend » Mon Jun 04, 2007 2:00 pm UTC

Gelsamel wrote:The fact of the matter is Question 3 is NOT useful, it tells us NOTHING.


it does tell us something, it tells us the probability of one roll and makes it explicit that we consider previous rolls to have no effect on future rolls, this is a point that is often missed by the general public (although pretty obvious to a bunch of mathematicians) it is important however as you clearly get a different (and very wrong result) if you modify your expectation of future rolls depending on the value of previous rolls (this all depends on the events being genuinely random of course, if they are highly correlated then modifying your expectation depending on past results is exactly the right thing to do)
in ur beanz makin u eveel

User avatar
Andrew
Posts: 619
Joined: Tue Jan 02, 2007 9:59 pm UTC
Location: Manchester, UK
Contact:

Postby Andrew » Mon Jun 04, 2007 2:08 pm UTC

I went out for a meal one day, and I sat in a different place to last time. The odds of me sitting in those two places in that order must be 1 in 2500. Big whoop. Yeah, so a 1 in 2500 chance came up. But there were 2500 mutually exclusive one-in-2500 chances. One of them had to come up. It's highly unlikely, but it's not special. There were 50 possible combinations of chairs that would have caused you to comment in this way, and you had a 1 in 50 chance of hitting one.

It's all well and good saying "they're answering the wrong question" but this is a conversation, not an arithmetic exam. It's entirely reasonable for them to tell you you're asking the wrong question.

Scarblac
Posts: 43
Joined: Wed Feb 28, 2007 11:39 am UTC

Postby Scarblac » Mon Jun 04, 2007 2:13 pm UTC

I like the dice analogy.

If you roll a 5, and then you roll a 5 again, then say "Gee, the same number as last time! What are the odds of that!" then the answer is 1 in 6, period.

If you then try to argue that you meant that specific number, not just any same number, so the answer should be 1 in 36 - then that just makes you look lame, since that obviously wasn't what you meant initially.

User avatar
Gelsamel
Lame and emo
Posts: 8237
Joined: Thu Oct 05, 2006 10:49 am UTC
Location: Melbourne, Victoria, Australia

Postby Gelsamel » Mon Jun 04, 2007 2:17 pm UTC

evilbeanfiend wrote:
Gelsamel wrote:The fact of the matter is Question 3 is NOT useful, it tells us NOTHING.


it does tell us something, it tells us the probability of one roll and makes it explicit that we consider previous rolls to have no effect on future rolls, this is a point that is often missed by the general public (although pretty obvious to a bunch of mathematicians) it is important however as you clearly get a different (and very wrong result) if you modify your expectation of future rolls depending on the value of previous rolls (this all depends on the events being genuinely random of course, if they are highly correlated then modifying your expectation depending on past results is exactly the right thing to do)


.......We're not "modify[ing] your expectation of future rolls depending on the value of previous rolls".


It's not as if we're saying the chance of picking 1 of 50 seats randomly is 1/2500.

We're saying the chance that seat 50 is chosen 2 times in a row totally randomly out of 50 seats is 1/2500, there is NOTHING wrong with that.

Random Choice 1 in question 1 and 2 has NO EFFECT on Random Choice 2, no effect in this situation AT ALL. However the WHOLE probability tree of TWO choices gives you a 1/2500 chance, 1/50 each time.

What I meant is by setting all happened events = 1 you LOSE information. And I was also saying since they asked the question after they HAD sat down at the same table wouldn't that happened event = 1?

Setting events = 1 is useless. It is better to totally IGNORE them, in which case 1d50 suffices. Of course then you're not answering the question, which is why only question 1 or 2 can answer "What are the odds?" meaningfully.
"Give up here?"
- > No
"Do you accept defeat?"
- > No
"Do you think games are silly little things?"
- > No
"Is it all pointless?"
- > No
"Do you admit there is no meaning to this world?"
- > No

User avatar
Woxor
Posts: 506
Joined: Mon May 07, 2007 11:28 pm UTC

Postby Woxor » Mon Jun 04, 2007 2:37 pm UTC

Gelsamel wrote:.......We're not "modify[ing] your expectation of future rolls depending on the value of previous rolls".


It's not as if we're saying the chance of picking 1 of 50 seats randomly is 1/2500.

We're saying the chance that seat 50 is chosen 2 times in a row totally randomly out of 50 seats is 1/2500, there is NOTHING wrong with that.

Random Choice 1 in question 1 and 2 has NO EFFECT on Random Choice 2, no effect in this situation AT ALL. However the WHOLE probability tree of TWO choices gives you a 1/2500 chance, 1/50 each time.

No one's disputing the mathematics. We're disputing the relevance.

Gelsamel wrote:What I meant is by setting all happened events = 1 you LOSE information. And I was also saying since they asked the question after they HAD sat down at the same table wouldn't that happened event = 1?

No, you gain/retain the information that it happened. You can figure out what the prior probability was, and no, there's nothing wrong with doing that, you just shouldn't pretend it's relevant when it's not.

Gelsamel wrote:Setting events = 1 is useless. It is better to totally IGNORE them, in which case 1d50 suffices. Of course then you're not answering the question, which is why only question 1 or 2 can answer "What are the odds?" meaningfully.

This is just nonsense. Ignoring past events isn't better in any way; interpreting them correctly is. So long as your epistemology is certain, events that have already happened have probability 1, but more importantly, they did happen (since events can have probability 1 and still not happen). This allows us to derive probabilities of other things that do depend on them with greater accuracy.

As many people in this thread have said a dozen different ways, the fact that it was in the given place twice was not what made the event notable and is thus irrelevant to the natural question it raised. The OP is just trying to cover for being bad at interpreting their question by hiding behind an irrelevant question that he WAS right about. I can't understand why you agree with him, though.

Questions 1 and 2 were only relevant before the event transpired, and Question 1 was still only relevant if the place in question was special in some way. These weren't the conditions when the question was asked, so they are irrelevant.

User avatar
evilbeanfiend
Posts: 2650
Joined: Tue Mar 13, 2007 7:05 am UTC
Location: the old world

Postby evilbeanfiend » Mon Jun 04, 2007 2:47 pm UTC

Gelsamel wrote:.......We're not "modify[ing] your expectation of future rolls depending on the value of previous rolls".


no, i know that, that's cos you are answering a different question. they are perfectly allowed to think that your question is irrelevant, in that way their position is defensible (note i'm not saying correct).

glossing over ignoring previous rolls is all very well but so many people make mistakes about that (e.g. "number 43 has come up every 3 weeks in a row in the lottery, so it must be less likely it will come up now!" except it isn't) so i'm not surprised the op's friends seem to be concerned that this sort of mistake is being made (even though it isn't) hence their instance that the first table was 'a given'.
in ur beanz makin u eveel


Return to “Mathematics”

Who is online

Users browsing this forum: No registered users and 10 guests