Roland, you seem to be missing the point here. He’s making two of them. One, asking about explosives is in some way privileging the hypothesis. Second, the probability of A and B must be at most the probability of A and B together. Your response has nothing to do with either of these issues.
One I have answered at least twice before, I’m getting tired of reiterating it again and again.
Second my point was to screen off arguments by getting close to the issue. So the conjunction with other events is irrelevant, but even if it weren’t, you are assuming that the events are independent, what if they aren’t?
So the conjunction with other events is irrelevant, but even if it were, you are
assuming that the events are independent, what if they aren’t?
No. Even if probabilities are not independent, the probability of both happening is still less than the probability of either one happening. This should be intuitively obvious, but since it apparently isn’t, here’s the math:
P(A v B)= P(A) + P(B) - P(A^B). So P(A^B) = P(A) + P(B) - P(A v B). Since P(A v B) >= P(A) and P(A v B) >= P(B) we have that P(A^B) ⇐ min(P(A),P(B)).
A more salient point here might be that the conjunctive issue doesn’t matter much because the probability that planes did crash into the Twin Towers is so high that it is almost 1 (unless anyone here thinks that the planes were actually holograms disguising a missile http://www.youtube.com/watch?v=q5NNh6WnVZo )), and thus the conjunction argument doesn’t by itself substantially reduce the probability of explosives being planted in WTC7. This sort of conjunction argument only works when both aspects of the conjunction have a probability substantially lower than 1.
You are not just confused about basic probability. You also seem to be confused about what Eliezer is talking about when he discusses closing off arguments. Please read his example again with the Wright Brothers. He’s talking about closing off classes of arguments as of much lower relevance to estimating an expectation. Thus, physics arguments are better in a general sense than arguments from authority. And physical demonstrations in many conditions are better in a general sense than arguments from the laws of physics. This is not at all the same as closing off questions about what connected events occurred (such as questions of motivation or the nature of the crashing planes or the like). It may help to reread that sequence.
Yes his point is more general than physics but that’s not the point! He’s not closing off explanations of related issues in reality. He’s closing off methods of weak evidence. He’s closing off classes of evidence, not evaluations of expected probabilities. Maybe expanding his example will make this more clear: If his example were equivalent to yours, instead of closing off Kelvin’s opinion, you’d be closing off whether or not the Wright Brothers had built an air tunnel. See the difference?
Roland, you seem to be missing the point here. He’s making two of them. One, asking about explosives is in some way privileging the hypothesis. Second, the probability of A and B must be at most the probability of A and B together. Your response has nothing to do with either of these issues.
One I have answered at least twice before, I’m getting tired of reiterating it again and again. Second my point was to screen off arguments by getting close to the issue. So the conjunction with other events is irrelevant, but even if it weren’t, you are assuming that the events are independent, what if they aren’t?
No. Even if probabilities are not independent, the probability of both happening is still less than the probability of either one happening. This should be intuitively obvious, but since it apparently isn’t, here’s the math:
P(A v B)= P(A) + P(B) - P(A^B). So P(A^B) = P(A) + P(B) - P(A v B). Since P(A v B) >= P(A) and P(A v B) >= P(B) we have that P(A^B) ⇐ min(P(A),P(B)).
A more salient point here might be that the conjunctive issue doesn’t matter much because the probability that planes did crash into the Twin Towers is so high that it is almost 1 (unless anyone here thinks that the planes were actually holograms disguising a missile http://www.youtube.com/watch?v=q5NNh6WnVZo )), and thus the conjunction argument doesn’t by itself substantially reduce the probability of explosives being planted in WTC7. This sort of conjunction argument only works when both aspects of the conjunction have a probability substantially lower than 1.
You are not just confused about basic probability. You also seem to be confused about what Eliezer is talking about when he discusses closing off arguments. Please read his example again with the Wright Brothers. He’s talking about closing off classes of arguments as of much lower relevance to estimating an expectation. Thus, physics arguments are better in a general sense than arguments from authority. And physical demonstrations in many conditions are better in a general sense than arguments from the laws of physics. This is not at all the same as closing off questions about what connected events occurred (such as questions of motivation or the nature of the crashing planes or the like). It may help to reread that sequence.
(Edits for grammar and punctuation)
I read the Eliezer’s post before commenting. Yes, he has some specific examples using physics, etc… but his point is general.
Yes his point is more general than physics but that’s not the point! He’s not closing off explanations of related issues in reality. He’s closing off methods of weak evidence. He’s closing off classes of evidence, not evaluations of expected probabilities. Maybe expanding his example will make this more clear: If his example were equivalent to yours, instead of closing off Kelvin’s opinion, you’d be closing off whether or not the Wright Brothers had built an air tunnel. See the difference?