well known proba problem

All non-Go discussions should go here.
Post Reply
User avatar
perceval
Lives in gote
Posts: 312
Joined: Thu Aug 05, 2010 3:35 am
Rank: 7K KGS
GD Posts: 0
KGS: tictac
Has thanked: 52 times
Been thanked: 41 times

well known proba problem

Post by perceval »

At work last friday i brought up the following proba problem and the open space productivity almost stopped due to the argument, i thought it might be fun to share here.

Its a fairly well know problem so you might already know of it, but almost everybody that hear it for the first time gets it wrong.

Suppose you have 3 Boxes:
- One with 2 Gold coins
- One with 2 Silver coins
- One with 1 Gold and 1 Silver Coin

Suppose you randomly pick a box, then take one of the 2 coins out from the box.
This Coin turns out to be a Gold coin. What is the probability that the second coin in the box is also Gold ? :mrgreen:
In theory, there is no difference between theory and practice. In practice, there is.
Kirby
Honinbo
Posts: 9553
Joined: Wed Feb 24, 2010 6:04 pm
GD Posts: 0
KGS: Kirby
Tygem: 커비라고해
Has thanked: 1583 times
Been thanked: 1707 times

Re: well known proba problem

Post by Kirby »

This reminds me of the monty hall problem. Intuition says that it's a 50% chance, but intuition didn't serve me well with the monty hall problem, either. ;-)

My rationale:
Two possible events: first box or third box.
be immersed
User avatar
HermanHiddema
Gosei
Posts: 2011
Joined: Tue Apr 20, 2010 10:08 am
Rank: Dutch 4D
GD Posts: 645
Universal go server handle: herminator
Location: Groningen, NL
Has thanked: 202 times
Been thanked: 1086 times

Re: well known proba problem

Post by HermanHiddema »

2/3?

There are six possible grabs, each with equal probability:

  1. Box GG grab G1
  2. Box GG grab G2
  3. Box GS grab G
  4. Box GS grab S
  5. Box SS grab S1
  6. Box SS grab S2

The last three options are eliminated by the fact that we grabbed gold, therefore one of the first three must be the one that happened. If it is either 1 or 2, then the other coin will also be gold. If it is 3, then the other will be silver.

Therefore, we have 2/3 chance of another gold.
User avatar
EdLee
Honinbo
Posts: 8859
Joined: Sat Apr 24, 2010 6:49 pm
GD Posts: 312
Location: Santa Barbara, CA
Has thanked: 349 times
Been thanked: 2070 times

Post by EdLee »

perceval, I can't believe you didn't convert this for here :)

Suppose you have 3 Go bowls:
- :white: :white: One with 2 Shell stones
- :black: :black: One with 2 Slate stones
- :white: :black: One with 1 Shell and 1 Slate stone

Suppose you randomly pick a bowl, then take one of the 2 stones out from the bowl.
This stone turns out to be :white:. What is the probability that the second stone in the bowl is also :white: ?
Kirby
Honinbo
Posts: 9553
Joined: Wed Feb 24, 2010 6:04 pm
GD Posts: 0
KGS: Kirby
Tygem: 커비라고해
Has thanked: 1583 times
Been thanked: 1707 times

Re: well known proba problem

Post by Kirby »

I thought about it more, and I think Herman is correct (I read his answer).

Another way to explain it is that there are 3 gold coins and 3 silver coins in the entire problem. Since you didn't draw a silver coin, the two silver coins in the silver-only scenario can't be from the pool of possibilities.

Therefore, there are 3 gold coins and 1 silver coin in the pool you selected from. You picked 1 gold coin, so there's 2 gold coins and 1 silver coin left. So you have 2/3 chance.
be immersed
User avatar
perceval
Lives in gote
Posts: 312
Joined: Thu Aug 05, 2010 3:35 am
Rank: 7K KGS
GD Posts: 0
KGS: tictac
Has thanked: 52 times
Been thanked: 41 times

Re:

Post by perceval »

EdLee wrote:perceval, I can't believe you didn't convert this for here :)

Suppose you have 3 Go bowls:
- :white: :white: One with 2 Shell stones
- :black: :black: One with 2 Slate stones
- :white: :black: One with 1 Shell and 1 Slate stone

Suppose you randomly pick a bowl, then take one of the 2 stones out from the bowl.
This stone turns out to be :white:. What is the probability that the second stone in the bowl is also :white: ?

shame on me :oops:
In theory, there is no difference between theory and practice. In practice, there is.
tj86430
Gosei
Posts: 1348
Joined: Wed Apr 28, 2010 12:42 am
Rank: FGA 7k GoR 1297
GD Posts: 0
Location: Finland
Has thanked: 49 times
Been thanked: 129 times

Re: well known proba problem

Post by tj86430 »

Herman has it, 2/3 is correct


I was quite recently involved in a lengthy (10 pages or so) discussion concerning this problem on another forum, and later there was another, even lengthier discussion concerning a similar problem, but where the wording was a bit more vague, on the same forum.

The other problem was like this (I have attempted to remove the vagueness of the wording): There are 150 cards. 40 of them are white on both sides, 50 of them are black on one side and white on the other side and 60 of them are black on both sides. You take a card, and see that one side is white, but you don't see the other side. What is the probability that the other side is white too?
Offending ad removed
User avatar
Bonobo
Oza
Posts: 2223
Joined: Fri Dec 23, 2011 6:39 pm
Rank: OGS 9k
GD Posts: 0
OGS: trohde
Universal go server handle: trohde
Location: Germany
Has thanked: 8262 times
Been thanked: 924 times
Contact:

Re: well known proba problem

Post by Bonobo »

50/50
“The only difference between me and a madman is that I’m not mad.” — Salvador Dali ★ Play a slooooow correspondence game with me on OGS? :)
User avatar
perceval
Lives in gote
Posts: 312
Joined: Thu Aug 05, 2010 3:35 am
Rank: 7K KGS
GD Posts: 0
KGS: tictac
Has thanked: 52 times
Been thanked: 41 times

Re: well known proba problem

Post by perceval »

tj86430 wrote:
Herman has it, 2/3 is correct


I was quite recently involved in a lengthy (10 pages or so) discussion concerning this problem on another forum, and later there was another, even lengthier discussion concerning a similar problem, but where the wording was a bit more vague, on the same forum.

The other problem was like this (I have attempted to remove the vagueness of the wording): There are 150 cards. 40 of them are white on both sides, 50 of them are black on one side and white on the other side and 60 of them are black on both sides. You take a card, and see that one side is white, but you don't see the other side. What is the probability that the other side is white too?

i'll have a go at this one:

Using the same kind of resaonning that the problem above: lets forget about the cards black on both side, we have not drawn one of them.
What we know is that we have drawn a white side. It is reasonnable to say that the proba to get any White side is the same. There are 40*2+50 = 130 white sides, amongst with 80 belong to a card that is white on both sides , so i would say 80/130 = 8/13 chances of the other side being white
Last edited by perceval on Tue Feb 12, 2013 5:25 am, edited 1 time in total.
In theory, there is no difference between theory and practice. In practice, there is.
tj86430
Gosei
Posts: 1348
Joined: Wed Apr 28, 2010 12:42 am
Rank: FGA 7k GoR 1297
GD Posts: 0
Location: Finland
Has thanked: 49 times
Been thanked: 129 times

Re:

Post by tj86430 »

Offending ad removed
tj86430
Gosei
Posts: 1348
Joined: Wed Apr 28, 2010 12:42 am
Rank: FGA 7k GoR 1297
GD Posts: 0
Location: Finland
Has thanked: 49 times
Been thanked: 129 times

Re: well known proba problem

Post by tj86430 »

perceval wrote:
tj86430 wrote:I was quite recently involved in a lengthy (10 pages or so) discussion concerning this problem on another forum, and later there was another, even lengthier discussion concerning a similar problem, but where the wording was a bit more vague, on the same forum.

The other problem was like this (I have attempted to remove the vagueness of the wording): There are 150 cards. 40 of them are white on both sides, 50 of them are black on one side and white on the other side and 60 of them are black on both sides. You take a card, and see that one side is white, but you don't see the other side. What is the probability that the other side is white too?

i'll have a go at this one:

Using the same kind of resaonning that the problem above: lets forget about the black cards on both side, we have not drawn one of them.
What we know is that we have drawn a white side. It reasonnalbe to say that the proba to get any White side is the same. There are 40*2+50 = 130 white sides, amongst with 80 belong to a card that is white on both sides , so i would say 80/130 = 8/13 chances of the other side being white

Correct
Offending ad removed
User avatar
Bonobo
Oza
Posts: 2223
Joined: Fri Dec 23, 2011 6:39 pm
Rank: OGS 9k
GD Posts: 0
OGS: trohde
Universal go server handle: trohde
Location: Germany
Has thanked: 8262 times
Been thanked: 924 times
Contact:

Re: well known proba problem

Post by Bonobo »

I don’t get it.

Choices at first:

- BB
- BW
- WW

I choose a box and take out a B stone, so the WW option has just been eliminated, hasn’t it? How should then a 2/3 chance remain? There is no /3 anymore, so it should be 2/2 -> 1:1 chance, no?

OK, I’m slightly dyscalculic, meaning I have to take of my shoes when counting to 20 ;-)

Please just disregard if I’m simply too stupid.
“The only difference between me and a madman is that I’m not mad.” — Salvador Dali ★ Play a slooooow correspondence game with me on OGS? :)
User avatar
perceval
Lives in gote
Posts: 312
Joined: Thu Aug 05, 2010 3:35 am
Rank: 7K KGS
GD Posts: 0
KGS: tictac
Has thanked: 52 times
Been thanked: 41 times

Re: well known proba problem

Post by perceval »

Bonobo wrote:I don’t get it.

Choices at first:

- BB
- BW
- WW

I choose a box and take out a B stone, so the WW option has just been eliminated, hasn’t it? How should then a 2/3 chance remain? There is no /3 anymore, so it should be 2/2 -> 1:1 chance, no?

OK, I’m slightly dyscalculic, meaning I have to take of my shoes when counting to 20 ;-)

Please just disregard if I’m simply too stupid.


Definitly not stupid, everybody starts with this argument, that is the beauty of the problem.
Here is some food for thought:

Lets change the game, now each 3 boxes contains 1000 coins:
- one contains 1000 G
- one contains 1000 S
- one contains 999 S and one G

You draw a coin from a box and notice its a gold coin, what is the probability that the box contains 999 other Gold coins ?

If you happen to draw a silver coin, what is the probability that the box contains 999 silver coins ?
In theory, there is no difference between theory and practice. In practice, there is.
tj86430
Gosei
Posts: 1348
Joined: Wed Apr 28, 2010 12:42 am
Rank: FGA 7k GoR 1297
GD Posts: 0
Location: Finland
Has thanked: 49 times
Been thanked: 129 times

Re: well known proba problem

Post by tj86430 »

Bonobo wrote:I don’t get it.

Choices at first:

- BB
- BW
- WW

I choose a box and take out a B stone, so the WW option has just been eliminated, hasn’t it? How should then a 2/3 chance remain? There is no /3 anymore, so it should be 2/2 -> 1:1 chance, no?

OK, I’m slightly dyscalculic, meaning I have to take of my shoes when counting to 20 ;-)

Please just disregard if I’m simply too stupid.

I don't know if I can explain it any more clearly than it has already been explained, but:

1. originally there are three black stones, right? (lets call them A, B and C, and let's also say that A & B are in the same bowl)
2. If 1. is correct, then you can take any one of those three with equal probability, right?
3. If 2. is correct, then there is 1/3 chance that you have taken the stone we decided to call A; 1/3 chance that you have taken B and 1/3 chance that you have taken C, right?
4. If 3. is correct, then there is 2/3 chance that there is another black in the same bowl (1/3 if you had taken A + 1/3 if you had taken B) and 1/3 chance that there is a white in the same bowl (in case you had taken C)
Offending ad removed
tj86430
Gosei
Posts: 1348
Joined: Wed Apr 28, 2010 12:42 am
Rank: FGA 7k GoR 1297
GD Posts: 0
Location: Finland
Has thanked: 49 times
Been thanked: 129 times

Re: well known proba problem

Post by tj86430 »

perceval wrote:everybody starts with this argument

I beg to disagree.
Offending ad removed
Post Reply