Bayesian Theory sounds like we can start a logic statistical demonstration based in nothing but thin air and then build from it, hoping that we will reach solid soil by hopping from priors to posteriors, guided by logic rules.
Good video! Funny point about the "very reliable test" that i learned somewhere: If the test is 95% accurate, and we give the test to 20 people, then if one of them tests positive, then it's equally likely that it is a false positive as it is that that person is positive positive. ;)
At the end you mentioned Glymour's argument about the "problem of old evidence" as an objection to Bayesian probability.
Yet, the "problem of old evidence" is bogus, and no one outside of philosophy take it seriously. Glymour in his paper used sloppy notation and conditioned on the wrong events, and thus reached a nonsensical result.
I suggest you read the presentation "Bayesians Can Learn From Old Data", at omega.albany.edu:8008/ME07dir/WilliamJefferys/slides.pdf
@MrFairObserver That's a great link, thanks for that. I'm not an expert on interpretations of probability theory so I can't judge the adequacy of Jaynes' approach to these issues, but I do agree that he's under-read among philosophers who write on probability theory.
@gjsterp Any interpretation of probability needs to make sense of statements like "It's true that the probability of a fair coin landing heads is 0.5". They'll differ on what makes such a claim true. Sounds like you're thinking of "relative frequency" interpretations of probability? Check out the "video response" on frequency interpretations of probability.
@menonfire12 Not yet, but these are planned. First up I'm finishing a course on the basic rules for reasoning with probabilities. Then I'm doing a course on fallacies of probabilistic reasoning. I do have plans for a course on elementary decision theory and game theory, because this helps to set up a whole other set of fallacies and cognitive biases that we can talk about. If you go to my site you can sign up for an email newsletter and I can let you know when new courses become available.
@menonfire12 I would surmise that the term "Dutch Book" is in keeping with the the tradition of affixing "Dutch" as a modifier to connote ersatz inauthenticity (eg, Dutch oven, Dutch courage). This has its origins in the seventeenth century rivalry between the Dutch and the English (rather like Italians referring to syphilis as the "French disease", and vice versa).
How is "Dutch Book" worded in Holland ? Can you imagined calling "nigger idea" to an idea that makes no sense? It would create a lot of fuss and even a few murders, and rightly so. (Let´s hope that nobody pays attention to expressions like "black list" or "dark side of the Force".)
Netherlanders should protest and ask for compensation against the Book Printers that are publishing Bayesian books. And Germans should also do, because they call themselves Dutchs.
This has been flagged as spam show
Bayesian Theory sounds like we can start a logic statistical demonstration based in nothing but thin air and then build from it, hoping that we will reach solid soil by hopping from priors to posteriors, guided by logic rules.
powerdriller10 1 week ago
Comment removed
powerdriller10 1 week ago
Good video! Funny point about the "very reliable test" that i learned somewhere: If the test is 95% accurate, and we give the test to 20 people, then if one of them tests positive, then it's equally likely that it is a false positive as it is that that person is positive positive. ;)
CPLains 4 weeks ago
At the end you mentioned Glymour's argument about the "problem of old evidence" as an objection to Bayesian probability.
Yet, the "problem of old evidence" is bogus, and no one outside of philosophy take it seriously. Glymour in his paper used sloppy notation and conditioned on the wrong events, and thus reached a nonsensical result.
I suggest you read the presentation "Bayesians Can Learn From Old Data", at omega.albany.edu:8008/ME07dir/WilliamJefferys/slides.pdf
MrFairObserver 3 months ago
@MrFairObserver That's a great link, thanks for that. I'm not an expert on interpretations of probability theory so I can't judge the adequacy of Jaynes' approach to these issues, but I do agree that he's under-read among philosophers who write on probability theory.
PhilosophyFreak 3 months ago
Thanks for the great lecture. Very well produced, very good content.
Spadie1 6 months ago
I don't see how "probability" can be said to be 'true'.
It is only a chance that it will be true.
But the 'truth' of the probability is only known after the fact.
After the coin is tossed.
The 'truth' will change after each toss.
gjsterp 8 months ago
@gjsterp Any interpretation of probability needs to make sense of statements like "It's true that the probability of a fair coin landing heads is 0.5". They'll differ on what makes such a claim true. Sounds like you're thinking of "relative frequency" interpretations of probability? Check out the "video response" on frequency interpretations of probability.
PhilosophyFreak 8 months ago
Do you have complete courses on decision theory and on game theory on the critical thinking academy ??
menonfire12 8 months ago
@menonfire12 Not yet, but these are planned. First up I'm finishing a course on the basic rules for reasoning with probabilities. Then I'm doing a course on fallacies of probabilistic reasoning. I do have plans for a course on elementary decision theory and game theory, because this helps to set up a whole other set of fallacies and cognitive biases that we can talk about. If you go to my site you can sign up for an email newsletter and I can let you know when new courses become available.
PhilosophyFreak 8 months ago
@PhilosophyFreak I will, thanks.
menonfire12 8 months ago
@menonfire12 I would surmise that the term "Dutch Book" is in keeping with the the tradition of affixing "Dutch" as a modifier to connote ersatz inauthenticity (eg, Dutch oven, Dutch courage). This has its origins in the seventeenth century rivalry between the Dutch and the English (rather like Italians referring to syphilis as the "French disease", and vice versa).
ChadSmith1452 8 months ago
@ChadSmith1452
How is "Dutch Book" worded in Holland ? Can you imagined calling "nigger idea" to an idea that makes no sense? It would create a lot of fuss and even a few murders, and rightly so. (Let´s hope that nobody pays attention to expressions like "black list" or "dark side of the Force".)
Netherlanders should protest and ask for compensation against the Book Printers that are publishing Bayesian books. And Germans should also do, because they call themselves Dutchs.
powerdriller10 1 week ago
@PhilosophyFreak Oops. I meant to reply to you, not menonfire12.
ChadSmith1452 8 months ago