It’s interesting that 20 out of 27 people say that there was no change in their opinion. Given conservation of expected evidence, this seems a bit odd. To a rough approximation, if a paper published on the subject could have provided evidence for it, but when investigated it failed to provide evidence, then we should update in favor of it being less likely. And if it succeeds in providing evidence, we should update in favor of it being more likely.
Likewise, if someone didn’t read the paper, he could still update one way or the other based on the existence of the paper.
Given conservation of expected evidence, this seems a bit odd.
Wrong event.
If someone tells me that they constructed a perpetual motion machine out of chains, cranks and pulleys, I do not expect to update my estimate of whether such a device is likely. I do however, expect to update my estimate of the probability of whether taking this person seriously is worthwhile.
If a serious scientist publishes an earnest paper claiming a revolutionary novel effect (like the superluminal neutrino paper last year), I would update my probability of this effect being real, until further information is available.
Rossi matches the pattern of a con artist, and none of the linked paper’s authors appear to be experts in debunking clever schemes. After reading the paper I have lower opinion of the paper authors, so yes, I have updated.
However, if someone makes the perpetual motion claim, unless you update your probability that they are worth taking seriously to 0%, you should also update your probability of the perpetual motion machine.
I don’t need to estimate “worth taking seriously” to 0, just “too low to bother”. (E.g., the update to my “perpetual motion machin probability” would be lower than the margin of error of my estimates.)
It’s interesting that 20 out of 27 people say that there was no change in their opinion. Given conservation of expected evidence, this seems a bit odd. To a rough approximation, if a paper published on the subject could have provided evidence for it, but when investigated it failed to provide evidence, then we should update in favor of it being less likely. And if it succeeds in providing evidence, we should update in favor of it being more likely.
Likewise, if someone didn’t read the paper, he could still update one way or the other based on the existence of the paper.
I was one of them, IIRC.
Yes. To a rough approximation. Excluding things like, ‘I can’t adjust my beliefs by such a small amount’.
Wrong event.
If someone tells me that they constructed a perpetual motion machine out of chains, cranks and pulleys, I do not expect to update my estimate of whether such a device is likely. I do however, expect to update my estimate of the probability of whether taking this person seriously is worthwhile.
If a serious scientist publishes an earnest paper claiming a revolutionary novel effect (like the superluminal neutrino paper last year), I would update my probability of this effect being real, until further information is available.
Rossi matches the pattern of a con artist, and none of the linked paper’s authors appear to be experts in debunking clever schemes. After reading the paper I have lower opinion of the paper authors, so yes, I have updated.
I agree with you about Rossi.
However, if someone makes the perpetual motion claim, unless you update your probability that they are worth taking seriously to 0%, you should also update your probability of the perpetual motion machine.
I don’t need to estimate “worth taking seriously” to 0, just “too low to bother”. (E.g., the update to my “perpetual motion machin probability” would be lower than the margin of error of my estimates.)
If you do, then you are prone to a version of the Pascal mugging attack: given enough false claims, you start taking them seriously.