SWC75
Bored Historian
- Joined
- Aug 26, 2011
- Messages
- 33,979
- Like
- 65,541
This started on the Eddie Lampkin thread and had nothing to do with it so I thought I'd put it in the portal to here. It's about a game show and math so it's kind of an "other sport".
It started here:
MS Orange: Exactly. I'm well aware that Lampkin has some limitations in his game, but I wasn't at all excited about most of the options being mentioned behind door #2.
My jokey response:
Ottoingrotto: This is a good post.
Me: Two theories:
- You take away one option, you are left with two options. It must be 50-50, right?
- You have a 1/3 chance it's behind the first door and that doesn't change just because it's not behind the third door.
I will go to my grave thinking that the first theory makes the most sense. What if you started out with two doors, had the contestant chose one of them, then wheeled out a third door, opened it and there was a goat there. Would that have made it a good idea for the contestant to switch? Isn't that the same situation?
I decided to test it with a deck of cards. I separated the black and red cards. The black cards are cars, the red goats. I combined one black card with two red cards. I shuffled the three cards as best I could. I took the first three and put them face down. I 'chose' #1 and flipped the third card over. if it was black, I put those three cards aside. If it was red, I then flipped the second card and recorded when it was black and when it was red. I had, as I recall 10 of 13 sets where the third card was red and the second card was black 7 times. I reshuffled them and did it a gain a couple more times and the 2/3 percentage held. I then realized that I was wrong but still don't understand why.
Skeeter223: And somehow, "Let's Make a Deal" became one of the most intriguing parts of the discussion.
Forloveoforang(e?): It took me 20 hours to figure out that it is absolutely twice as likely that its door number 2!
Door number one is a 1/3 chance. Doors 2 and 3 combined are a 2/3 chance. Monte does the contestant a huge favor by saying that Door number 3 is not it. So that means that number 2 has the full 2/3 chance.
Showing a non-car door does not change the 1/3 chance for door number 1. Again the bottom line is that doors 2 and 3 combined are a 2/3 chance, and Monte let the contestant know that the 2/3 chance is in door number 2 by opening door number 3.
That's quite a tough one to get your head around!
ROCuse: The first time you pick out of three doors the chance of success is 33%. The next time you pick out of two doors the chance of success is 50%. The scenario is are independent of each other, so given the chance to switch you should always switch to the “roll” with the best odds for success.
Ragman: This is incorrect. Look up the Money Hall problem explained. The correct answer is 67% chance of meeting correct if you switch, not 50.
Fat Tire 13: Hurts my brain too - and if you want to go down the Wiki rabbit hole….
Forloveoforang: Yes, depending on how the host is playing the game...There is a 2/3 chance that Monte did not pick door number 2 because the car was behind it, right? That's why the 2 doors are not equal 50-50 chances.
Me: Yeah, that's the second option. But why does it rule over the first option?
AZOrange: Take the issue to an absurd level. Let’s say you had 100 suitcases in them 1 with $1mm in it. You pick 1 and the host reveals 98 losers on the other side. Does it make any sense that you have a 50/50 chance now when you started with 1%?
Me: Yes. That doesn't make it true, (as I explained), but you are left with two suitcases, one of which has $1mm in it. I understand both arguments. I just don't understand why the second one rules the situation.
Forloveoforang: Like I said, I understand it is hard to understand. But I think the key is knowing that the combined 2 and 3 doors hold a 2 out of 3 chance. When its known that 3 is not it, it swings all of that 2 out of 3 chance to the 2nd door. Its like ah ha, the host had to pick over the 2nd door ( the scenario being that he knows which door has the car). The 2 remaining doors are not created equally, anymore. The second door absolutely holds the 66 and 2/3 % chance that is missing. (There is still a 33 and 1/3 % chance that it is Door number 1). Its hard to get your head around that fact.
JustPheru: I’ll find Jimmy Hoffa before I can understand this.
Capt Tuttle: Took me a while to get it, too.
The scenarios aren’t independent. They exist on a continuum. The 50/50 would only exist if Monty had the curtain closed so they could move the prizes around (or not) between the 2 choices. Because that doesn’t happen, door #1 will always have a1/3 probability.
Another way to think of it is after choosing door #1, there is a 2/3 probability the car is behind one of the other two. That doesn’t change with the elimination of door 3.
The determinating condition is that there were 3 choices for the original placement of the car. That doesn’t change.
Cuse181: I think that is sort of the simple way to go about it. You are basically being given a second door. All the doors had 33% chance, you picked 2, he gave you 3 the idea (I think) is you’re going into your next decision with a 66% of being right because you KNOW 3 is a goat and 2 was your choice. You implicitly received information that 2 is not 1 of the 2 non-car doors, thus giving you the improved chance of being right.
AZOrange: You odds don’t change just because an answer is revealed. When this ”problem” came out many disputed it until they did a large number of testing of the hypothesis.
In summary, I understand the 1/3 vs. 2/3 argument and the alternative "now it's 50-50" scenario and have tested it and found that 1/3 vs. 2/3 is the one that works. I just don't understand why it does.
It started here:
MS Orange: Exactly. I'm well aware that Lampkin has some limitations in his game, but I wasn't at all excited about most of the options being mentioned behind door #2.
My jokey response:
Monty Hall problem - Wikipedia
en.wikipedia.org
Ottoingrotto: This is a good post.
Me: Two theories:
- You take away one option, you are left with two options. It must be 50-50, right?
- You have a 1/3 chance it's behind the first door and that doesn't change just because it's not behind the third door.
I will go to my grave thinking that the first theory makes the most sense. What if you started out with two doors, had the contestant chose one of them, then wheeled out a third door, opened it and there was a goat there. Would that have made it a good idea for the contestant to switch? Isn't that the same situation?
I decided to test it with a deck of cards. I separated the black and red cards. The black cards are cars, the red goats. I combined one black card with two red cards. I shuffled the three cards as best I could. I took the first three and put them face down. I 'chose' #1 and flipped the third card over. if it was black, I put those three cards aside. If it was red, I then flipped the second card and recorded when it was black and when it was red. I had, as I recall 10 of 13 sets where the third card was red and the second card was black 7 times. I reshuffled them and did it a gain a couple more times and the 2/3 percentage held. I then realized that I was wrong but still don't understand why.
Skeeter223: And somehow, "Let's Make a Deal" became one of the most intriguing parts of the discussion.
Forloveoforang(e?): It took me 20 hours to figure out that it is absolutely twice as likely that its door number 2!
Door number one is a 1/3 chance. Doors 2 and 3 combined are a 2/3 chance. Monte does the contestant a huge favor by saying that Door number 3 is not it. So that means that number 2 has the full 2/3 chance.
Showing a non-car door does not change the 1/3 chance for door number 1. Again the bottom line is that doors 2 and 3 combined are a 2/3 chance, and Monte let the contestant know that the 2/3 chance is in door number 2 by opening door number 3.
That's quite a tough one to get your head around!
ROCuse: The first time you pick out of three doors the chance of success is 33%. The next time you pick out of two doors the chance of success is 50%. The scenario is are independent of each other, so given the chance to switch you should always switch to the “roll” with the best odds for success.
Ragman: This is incorrect. Look up the Money Hall problem explained. The correct answer is 67% chance of meeting correct if you switch, not 50.
Fat Tire 13: Hurts my brain too - and if you want to go down the Wiki rabbit hole….
Forloveoforang: Yes, depending on how the host is playing the game...There is a 2/3 chance that Monte did not pick door number 2 because the car was behind it, right? That's why the 2 doors are not equal 50-50 chances.
Me: Yeah, that's the second option. But why does it rule over the first option?
AZOrange: Take the issue to an absurd level. Let’s say you had 100 suitcases in them 1 with $1mm in it. You pick 1 and the host reveals 98 losers on the other side. Does it make any sense that you have a 50/50 chance now when you started with 1%?
Me: Yes. That doesn't make it true, (as I explained), but you are left with two suitcases, one of which has $1mm in it. I understand both arguments. I just don't understand why the second one rules the situation.
Forloveoforang: Like I said, I understand it is hard to understand. But I think the key is knowing that the combined 2 and 3 doors hold a 2 out of 3 chance. When its known that 3 is not it, it swings all of that 2 out of 3 chance to the 2nd door. Its like ah ha, the host had to pick over the 2nd door ( the scenario being that he knows which door has the car). The 2 remaining doors are not created equally, anymore. The second door absolutely holds the 66 and 2/3 % chance that is missing. (There is still a 33 and 1/3 % chance that it is Door number 1). Its hard to get your head around that fact.
JustPheru: I’ll find Jimmy Hoffa before I can understand this.
Capt Tuttle: Took me a while to get it, too.
The scenarios aren’t independent. They exist on a continuum. The 50/50 would only exist if Monty had the curtain closed so they could move the prizes around (or not) between the 2 choices. Because that doesn’t happen, door #1 will always have a1/3 probability.
Another way to think of it is after choosing door #1, there is a 2/3 probability the car is behind one of the other two. That doesn’t change with the elimination of door 3.
The determinating condition is that there were 3 choices for the original placement of the car. That doesn’t change.
Cuse181: I think that is sort of the simple way to go about it. You are basically being given a second door. All the doors had 33% chance, you picked 2, he gave you 3 the idea (I think) is you’re going into your next decision with a 66% of being right because you KNOW 3 is a goat and 2 was your choice. You implicitly received information that 2 is not 1 of the 2 non-car doors, thus giving you the improved chance of being right.
AZOrange: You odds don’t change just because an answer is revealed. When this ”problem” came out many disputed it until they did a large number of testing of the hypothesis.
In summary, I understand the 1/3 vs. 2/3 argument and the alternative "now it's 50-50" scenario and have tested it and found that 1/3 vs. 2/3 is the one that works. I just don't understand why it does.