1. Cars, Goats, and Sample Spaces
这道题的精髓在于:(Relevant) information improves your chance of winning.
1. 不换:You are essentially not using the extra information to improve your chances. Thus your likelihood to win is the same as before Monty opens the other door.
2. 换:在这里很多人摔跤,因为觉得现在只剩下两扇门了,随机选一个的胜率就肯定是1/2。但是这两扇门是不一样的。对比一下下面这两种情况: - 有A、B两扇门,其中一扇门背后有大奖,you do not know anything else about A or B and if you make random choice, your chance to win is indeed 1/2. - 有A、B两扇门,其中一扇门背后有大奖。你选了A,但是现在你知道B是知情人Monty没有选的。这样的话,A和B两扇门的“信息量”就不同了。
Better Explained上的解释说的比我好:
在这种情况下把Car = A、Car = B两个事件model成independent random events就不行了。 利用Bayes theorem的解法:
不换的胜算:1/3
换的胜算:假设有A、B、C三门,你选了A,大奖在B,Monty打开了C门。换的话你就得选B。 P(car = B | Monty opens C) = P(Monty opens C | Car = B) * P(car = B) / P(Monty opens C) - P(Monty Opens C | Car = B) = 1 - P(car = B) = 1/3 - P(Monty opens C) = 1/2 - 换的胜算是2/3
我们也可以用Bayes theorem来写一下stick with A的胜率: P(car = A | Monty opens C) = P(Monty opens C | Car = A) * P(car = A) / P(Monty opens C) - P(Monty opens C | Car = A) = 1/2 - P(Car=A) = 1/3 - P(Monty opens C) = 1/2 - 不换的胜算是1/3。
SyampuuiYojeon对本书的所有笔记 · · · · · ·
-
1. Cars, Goats, and Sample Spaces
-
3. Polya urn model
https://en.wikipedia.org/wiki/P%C3%B3lya_urn_model
说明 · · · · · ·
表示其中内容是对原文的摘抄