How to get the most out of realizing you are wrong by using Bayes’ Theorem to update your beliefs


[Read the post]


Nope, still don’t listen to online video or sound recordings, clickbait or not.

Until I’ve had the chance to look through the transcript that is…


Yep. This.

@jlw, what else could be done with this content? The topic sounds pretty cool, but I frankly cannot scan an audio clip for highlights, and I don’t have the time to listen with anything like the level of intensity I can read with. I like the idea, but I don’t like that it’s exclusively audio. I really never will click over to listen without some level of detail.


It is a podcast. The creators intend you listen to it. It is their art. If you would like transcripts I suggest you let them know you’d like them. I’d hope they read the comments here, but again we would never require our guests do that.


Thanks, great idea!


Regarding article’s illustration: Everyone knows it’s turtles all the way down, I refuse to believe otherwise!

( )


After listening to the podcast, my confidence that Bayes’ Theorem can help me live my day to day life has not improved. On the other hand, my confidence in my own ability to understand complex important ideas has dropped sligjtly. I may just need more data.

(On the gripping hand, it seems a Bayesian approach to something like the Drake equation should prove most productive, expecially as we add more exoplanets to the catalog)


They’re Patreon backed, and their website says that the Patreon goal for providing transcripts has already been met.

They’re just [a little behind] ( with the transcriptions.


I definitely misread this title as the much more exciting:

“How to get the most out of realizing you are wrong to use Bayes’ Theorem to update your beliefs”

I’d be super interested in that.


I taught math for many years, and this is the illustration I always used for Bayes’ Theorem:
Suppose 5% of the football team uses drugs, and they start using a drug test that is 95% accurate. If someone tests positive, what is the probability that they actually use drugs?
I won’t bore you with the actual formula, but imagine 10,000 people are tested. Of those, 5%, or 500 people, are using drugs. If 95% of them test positive, that’s 475 rightfully accused people. But, of the 9,500 non-users, 5% will also test positive (the test is only 95% accurate). So the actual probability that someone who tests positive on this 95% accurate drug test is actually using drugs is only 50%, not 95% as most people assume.
Good enough to ruin someone’s life over?


Your scenario assumes that false positives and false negatives are detected at a rate of 5%. I don’t think that it’s often the case that the false negative rate is the same as the false positive rate.


It doesn’t really matter, though, the real point is that you can’t take “95% accurate” and go, “Oh, I can trust it 95% of the time.”

Well, the real point of Bayes theorem is that P(A|B) = P(B|A) * P(A) / P(B).

But for most people, I’d probably simplify the whole thing down to, “Don’t trust the media when they say ‘percent’.”


Bayes theorem is useful for statistics (though the advocates would do well to stop misrepresenting frequentist statistics when they advocate for it), but I too was hoping the thing might be about the Cult of Bayes that uses the theorem as a magic bullet for domains it doesn’t apply well to, and as a way to sneak biases into assumptions about priors so that they eventually justify reasoning themselves into all kinds of strange places like obsession on the Singularity.


I’m usually always wrong, mostly…


Then again, I’m always right…

'Tis a paradox.


I suppose I should have specified more carefully. I actually started to, but thought more of brevity.
You’re probably right about the false positives and negatives, but it was a thought experiment after all. IRL, they would just do another test.


I misread it in a slightly different way:
"How to get the most out of realizing you using the wrong Bayes’ Theorem to update your beliefs"
and I was wondering what the other theorem was…


Maybe you’ll find it in “Divine Benevolence, or an Attempt to Prove That the Principal End of the Divine Providence and Government is the Happiness of His Creatures”


This topic was automatically closed after 5 days. New replies are no longer allowed.