Talk:Absence of evidence
From RationalWiki
Aaaah Vienna! ("clearer explanations required not invoking" any "Bayesian style maths") Also the sections of the diagram could be labelled Pippa (talk) 15:31, 25 August 2011 (UTC)
- Difficult task as it's only convincing in a Bayesian framework with the probabilities involved. Otherwise you're still stuck with the "but it could still exist but we haven't seen it yet" mentality. If you contrast a missing transitional fossil with Russell's teapot it might help because that will introduce the idea of "prior" probability that you need for this thing to work. Namely, the prior probability, that is the probability before you look for this evidence, of a transitional fossil existing is higher than a celestial teapot. So finding no evidence of a teapot in orbit is stronger evidence against a teapot than no evidence of a transitional fossil is evidence against that. ADK...I'll discombobulate your balloon! 16:17, 25 August 2011 (UTC)
- Really? If look around me now I do not observe an elephant; that is, there is an absence of evidence for an elephant in the room. Seems to me that the absence of evidence in this case, that I don't see an elephant, is even on an intuitive level abundant evidence that there is, in fact, no elephant in the room. This seems relatively easy to grasp and seems to illustrate at least one of the fundamental points pretty well. A Bayesian framework is needed to model precisely how this works in a formal model, but not to grasp the basic idea. Am I missing something? (the elephant example does not capture the analogy to transitional fossils or Russell's teapot completely, but it does get the first step)G.D. (talk) 20:51, 28 July 2012 (UTC)
- Intuitively, that works as an example. Although it does limit the scope a little bit to situations where the only way you'd be wrong is for the nature of reality and your senses to be lying to you - which is always going to be what keeps you that tiny minuscule remote minute infinitesimal fraction away from total certainty about anything.
moral 23:30, 29 July 2012 (UTC)
- Intuitively, that works as an example. Although it does limit the scope a little bit to situations where the only way you'd be wrong is for the nature of reality and your senses to be lying to you - which is always going to be what keeps you that tiny minuscule remote minute infinitesimal fraction away from total certainty about anything.
- Really? If look around me now I do not observe an elephant; that is, there is an absence of evidence for an elephant in the room. Seems to me that the absence of evidence in this case, that I don't see an elephant, is even on an intuitive level abundant evidence that there is, in fact, no elephant in the room. This seems relatively easy to grasp and seems to illustrate at least one of the fundamental points pretty well. A Bayesian framework is needed to model precisely how this works in a formal model, but not to grasp the basic idea. Am I missing something? (the elephant example does not capture the analogy to transitional fossils or Russell's teapot completely, but it does get the first step)G.D. (talk) 20:51, 28 July 2012 (UTC)
[edit] Worst page on the site
My God. The very first line is a bold and stupid denial of a reasonable logical principle. The fact that there is no evidence for a claim is decidedly not evidence it is false.
For instance, I have no evidence that the 90 trillionth digit of pi is a 7. Nor do I have evidence it is not a 7. According to this moronic thesis, I have reason to believe both claims are false. Phiwum (talk) 15:33, 12 December 2013 (UTC)
- I don't see how that's wrong. --81.175.225.92 (talk) 16:46, 13 December 2013 (UTC)
- It should just be tumbleweeds. Ba-dum-pshh. Occasionaluse (talk) 18:44, 13 December 2013 (UTC)
- The stupid thing here is you. Observe:
- •
- H := "The 90 trillionth decimal digit of pi is a 7."
- P(H) := 0.1
- E1 := "Phiwum finds a proof of H within one year, without really looking."
- P(E1) := 10^(-10)
- E2 := "Phiwum finds a proof of ~H within one year, without really looking."
- P(E2) := 10^(-9)
- P(H|E1) := 0.99
- •
- => P(H|~E1) = 0.099999999910^_ (period 10) < 0.1 = P(H)
- So there is a shift here, but it's very minor, because you did not really expect to encounter the evidence. The evidence consists of finding the proof; not in the proof existing. There is about a 0.1 chance that the proof exists, of course.
- •
- Note that even now, we still have 2 degrees of freedom:
- P(H|E2) is not constrained. And while I would insist that P(E1∩E2) ≤ 1.1*10^(-9) (because E1∩E2 is either a paradox or a mistake; one of the proofs does not lead to the other), it is also otherwise free.
- Indeed, we can still require "P(H|~(E1∪E2)) = P(H)" without contradiction, for example.
- •
- And indeed, P(H|~(E1∪E2)) is what we are really looking for, here. Or did you actually manage to find one of the above proofs within the last year? Did you try hard enough to have some hope of succeeding?
- Dinya (talk) 01:44, 17 February 2015 (UTC)
[edit] Elephant
There seems to be a desire to remove the following without debate
- Take another even more obvious example. Someone maintains that they keep an elephant in their garage. A perfectly normal visible large grey elephant. You go to look and you cannot see any elephant. You enter the garage, which is quite small, and look around it. There is no elephant smell or any elephant droppings. In fact there is a total lack of evidence of any sort which would suggest there is an elephant in the garage. You would quite reasonably assume that the absence of elephantine evidence indicated the absence of the elephant.
Please explain before deleting.--Weirdstuff (talk) 20:51, 13 December 2013 (UTC)
- It's painful to read and adds nothing to the article that isn't clearly stated already. PowderSmokeAndLeather: Say something once, why say it again?.
20:54, 13 December 2013 (UTC)