E(B) = 10 * .5 + 10,000 * .5 = 5,005
E(A) = 1,000
E(B) >E(A)
E(B) = 10 * .5 + 10,000 * .5 = 5,005
E(A) = 1,000
E(B) >E(A)
That's a failing of the underlying theory.
The fuck are you on about? The distribution you're trying to describe is nonsensical.
So what, you chose a big number,... bravo?
You are right with your repeating "no uniform distribution over IR", But I have to repeat myself too: You don't know the distribution. And more importantly, by knowing box A contains 1000 you gain no information about the distribution. So what matters is whether you have taken the box with the high value or the one with the low value - 50:50.
>So what matters is whether you have taken the box with the high value or the one with the low value - 50:50.
Not knowing a distribution does not make it 50/50, it makes it unknown.
>And more importantly, by knowing box A contains 1000 you gain no information about the distribution.
Of course sampling a random variable gives you information about its distribution. Even a single sample of a uniform distribution gives you an informed guess at its upper bound.
en.wikipedia.org
the other box could have either $10000 or $100 but it's 50/50
10000-1000>|100-1000|