Mojo
Mostly harmless
- Do you agree with any of my premises?
I agree with the premise that you exist. Unfortunately you are trying to use your own existence as an event subsequent to your initial assessment of the probability of H.
- Do you agree with any of my premises?
Your existence is predicted by the model! The universe is working exactly the way we think it does. It's not unusual that somebody occasionally wins the lottery.Dave,
- Maybe this can clear things up. We aren't talking about the givens (note the s) of OOFLam, we're talking about the likelihood of something when OOFlam is given.
LL,
- Do you agree with any of my premises?
Dave,
- I think that the answer is, "That's how Bayesian statistics works."
Your existence is predicted by the model! The universe is working exactly the way we think it does. It's not unusual that somebody occasionally wins the lottery.
I think stating "your existence is predicted by the model" is wrong. Neither his "scientific" model nor his unscientific model predicted specifically his existence. Nobody made a prediction that Jabba would exist, and then stood back and watched whether the prediction came true. Rather Jabba is hypothesizing after the results are known. He observes his own existence, and then manufactures a hypothesis that fits the observation.
Neither [Jabba's] "scientific" model nor his unscientific model predicted specifically his existence. Nobody made a prediction that Jabba would exist, and then stood back and watched whether the prediction came true. Rather Jabba is hypothesizing after the results are known. He observes his own existence, and then manufactures a hypothesis that fits the observation.
Fine. His existence is consistent with the model. The mere fact of his existence is no more a reason to doubt the scientific universe than the existence of Real Housewives Of Atlanta is to doubt the workings of the television industry.
Right, so when you calculate the likelihood of E occurring if OOFLam is true you do it based on the givens of OOFLam.
Dave,
- Maybe this can clear things up. We aren't talking about the givens (note the s) of OOFLam, we're talking about the likelihood of something when OOFlam is given.
Dave,I was using those to mean the same thing.
If we're talking about the likelihood of something when OOFlam is given, we're basing that likelihood on OOFLam being correct.
But note that we are not saying that OOFLam is correct; we are asking, "What is the likelihood of E happening IF OOFLam is correct?" Likelihood is based upon a hypothetical.
Dave,
- Exactly. We are basing that likelihood on OOFLam being correct.
- We're saying, "IF OOFLam is correct, then..."
- But note that we are not saying that OOFLam is correct; we are asking, "What is the likelihood of E happening IF OOFLam is correct?" Likelihood is based upon a hypothetical.
...would be meaningless.
Dave,So, when talking about P(E|OOFLam), why does it matter if I think it's possible OOFLam is wrong?
It would have personal meaning for someone who is emotionally entrenched in a supernatural belief and wanted to buttress that with the notion that there was some objective or scientific support for it.
Dave,
- Cause, if my estimates for the other variables in the formula are roughly correct, P(H|E) is virtually zero.
Dave,
- Cause, if my estimates for the other variables in the formula are roughly correct, P(H|E) is virtually zero.
Dave,The variables on the other side of the formula have no bearing on P(H|E). P(H|E) is only based on E. P(H|~E) is only based on ~E. The value of P(H|~E) does not change the value of P(H|E), and vice versa.
That seems like you're saying that Bayesian statistics doesn't work...