Bayes Theorem. Is that still a thing?

Where have we been?
User avatar
Lausten
Persistent Poster
Posts: 3917
Joined: Sun Dec 06, 2009 6:33 pm
Location: Northern Minnesota

Bayes Theorem. Is that still a thing?

Post by Lausten » Tue Mar 12, 2019 12:47 am

I haven't seen a reference to Bayes 2012. So, it's about time. This article was written in 2014 and offers a good intro. It begins with the statement, ".. if you are claiming something is the case (or is not the case), you cannot argue that claim is true by looking around for evidence that supports it." Which I had to read twice and when I read the next statement, was glad he called that counter-intuitive. Of the many things Bayes does, it helps us to see how we naturally consider if something is true or not, in Carrier's case, he's talking about true in history. Although we generally start out presenting our evidence for something, we actually have a lot of prior assumptions about what we're saying and we have melded all our earlier assumptions into a general idea of how confident we are of whatever the claim is. So although in our head we might be thinking we are 82% sure, it might come out our mouth's as "we believe that".

There are lot's of reason for this in casual conversation, so I'm not recommending using Bayes for every forum post, but what's interesting about the formula is that it captures all our assumptions, and it can be used to understand when someone missed a logical step, whether they did it accidentally, or are meaning to be deceptive. I don't want to rewrite the whole blog, so check it out.

I'll also address one objection that has already been made; human thoughts are not math. That's true. To use Bayes when taking something a historian claims, like "I'm pretty sure about that", then translating that into 95%, we are not attempting to put a number on thoughts, or to even come to a result that is something you could bank on. The point of the formula is to track when you make an assumption, what it was, and what evidence you have to go with it. Then, when someone refutes your claim, they know they can challenge your assumption or your evidence, but there is no point in challenging the Bayes formula, that's neutral.
A sermon helper that doesn't tell you what to believe: http://www.milepost100.com

bobbo_the_Pragmatist
Has No Life
Posts: 19276
Joined: Wed Sep 09, 2015 9:39 am
Custom Title: After being pimped comes-----

Re: Bayes Theorem. Is that still a thing?

Post by bobbo_the_Pragmatist » Tue Mar 12, 2019 12:54 am

Most claims that are counter intuitive are wrong. probablility is high at .85.

Putting math on Bs is still BS. People who love dressing BS up thinking they have some secret can't be stopped from doing it.
Real Name: bobbo the contrarian existential pragmatic evangelical anti-theist and Class Warrior.
Asking: What is the most good for the most people?
Sample Issue: Should the Feds provide all babies with free diapers?

User avatar
Lausten
Persistent Poster
Posts: 3917
Joined: Sun Dec 06, 2009 6:33 pm
Location: Northern Minnesota

Re: Bayes Theorem. Is that still a thing?

Post by Lausten » Tue Mar 12, 2019 3:02 am

You comment is barely coherent.
A sermon helper that doesn't tell you what to believe: http://www.milepost100.com

bobbo_the_Pragmatist
Has No Life
Posts: 19276
Joined: Wed Sep 09, 2015 9:39 am
Custom Title: After being pimped comes-----

Re: Bayes Theorem. Is that still a thing?

Post by bobbo_the_Pragmatist » Tue Mar 12, 2019 9:10 am

Lausten wrote:
Tue Mar 12, 2019 3:02 am
You comment is barely coherent.
What probability of it being correct or wrong would you give it? What factors are you including and excluding??
Real Name: bobbo the contrarian existential pragmatic evangelical anti-theist and Class Warrior.
Asking: What is the most good for the most people?
Sample Issue: Should the Feds provide all babies with free diapers?

User avatar
Lausten
Persistent Poster
Posts: 3917
Joined: Sun Dec 06, 2009 6:33 pm
Location: Northern Minnesota

Re: Bayes Theorem. Is that still a thing?

Post by Lausten » Tue Mar 12, 2019 11:21 am

bobbo_the_Pragmatist wrote:
Tue Mar 12, 2019 9:10 am
Lausten wrote:
Tue Mar 12, 2019 3:02 am
You comment is barely coherent.
What probability of it being correct or wrong would you give it? What factors are you including and excluding??
Given your past posts and a few facts about grammar, not looking good for you.
A sermon helper that doesn't tell you what to believe: http://www.milepost100.com

bobbo_the_Pragmatist
Has No Life
Posts: 19276
Joined: Wed Sep 09, 2015 9:39 am
Custom Title: After being pimped comes-----

Re: Bayes Theorem. Is that still a thing?

Post by bobbo_the_Pragmatist » Tue Mar 12, 2019 12:51 pm

Thats not a probability. Sure you understand the subject?
Real Name: bobbo the contrarian existential pragmatic evangelical anti-theist and Class Warrior.
Asking: What is the most good for the most people?
Sample Issue: Should the Feds provide all babies with free diapers?

User avatar
Lausten
Persistent Poster
Posts: 3917
Joined: Sun Dec 06, 2009 6:33 pm
Location: Northern Minnesota

Re: Bayes Theorem. Is that still a thing?

Post by Lausten » Tue Mar 12, 2019 4:09 pm

bobbo_the_Pragmatist wrote:
Tue Mar 12, 2019 12:51 pm
Thats not a probability. Sure you understand the subject?
It is. That's the entire point. You can assign "not looking good" a number, but that doesn't change the meaning. If you can't get that simple point, you can't get started with this topic.
A sermon helper that doesn't tell you what to believe: http://www.milepost100.com

bobbo_the_Pragmatist
Has No Life
Posts: 19276
Joined: Wed Sep 09, 2015 9:39 am
Custom Title: After being pimped comes-----

Re: Bayes Theorem. Is that still a thing?

Post by bobbo_the_Pragmatist » Tue Mar 12, 2019 4:23 pm

and in what sense is that assigned number a probability as defined in mathematics?
Real Name: bobbo the contrarian existential pragmatic evangelical anti-theist and Class Warrior.
Asking: What is the most good for the most people?
Sample Issue: Should the Feds provide all babies with free diapers?

User avatar
landrew
Has No Life
Posts: 11208
Joined: Sat Jul 14, 2007 5:51 am
Location: Fox Meadows

Re: Bayes Theorem. Is that still a thing?

Post by landrew » Tue Mar 12, 2019 5:03 pm

If a probability has a large margin of error, is it still useful? Many eggheads think so. One of the advisors on my graduate committee was stuck on the idea that all the effects observed in my research should be reducible to a formula of some kind. He fought for it for awhile, but he was over-ruled by the other professors. Formulas are often out of place in the life sciences because of a large number of inconsistent and unknown variables. The results are often wildly inaccurate. Most life scientists know this and refrain from using formulas unless appropriate.

Angelina Jolie had two healthy breasts removed, no doubt due to some Bayes theorem calculation by some egghead doctor. Stupid woman.
The job of a skeptic is to investigate the unexplained; not to explain the uninvestigated.

User avatar
Lausten
Persistent Poster
Posts: 3917
Joined: Sun Dec 06, 2009 6:33 pm
Location: Northern Minnesota

Re: Bayes Theorem. Is that still a thing?

Post by Lausten » Tue Mar 12, 2019 5:25 pm

bobbo_the_Pragmatist wrote:
Tue Mar 12, 2019 4:23 pm
and in what sense is that assigned number a probability as defined in mathematics?
That question doesn't make sense. A person assigns a number to the statement, i.e. very likely = 99%, very unlikely = 1%. Math doesn't have a rule for that.
A sermon helper that doesn't tell you what to believe: http://www.milepost100.com

User avatar
Lausten
Persistent Poster
Posts: 3917
Joined: Sun Dec 06, 2009 6:33 pm
Location: Northern Minnesota

Re: Bayes Theorem. Is that still a thing?

Post by Lausten » Tue Mar 12, 2019 5:28 pm

I should have said at the outset that I am aware of opposition to this theorem. I was hoping for some objections that I could grapple with. Something better than, "you can't do numbers on that".
A sermon helper that doesn't tell you what to believe: http://www.milepost100.com

User avatar
landrew
Has No Life
Posts: 11208
Joined: Sat Jul 14, 2007 5:51 am
Location: Fox Meadows

Re: Bayes Theorem. Is that still a thing?

Post by landrew » Tue Mar 12, 2019 6:22 pm

Lausten wrote:
Tue Mar 12, 2019 5:28 pm
I should have said at the outset that I am aware of opposition to this theorem. I was hoping for some objections that I could grapple with. Something better than, "you can't do numbers on that".
I have stated my opinion, but I doubt you will find some universal final answer to that question, as you guys seem wont to believe. I have little faith in climate predictions as many others do, but this argument keeps circling back to the notion that such predictions are credible, when clearly they have not been so.
The job of a skeptic is to investigate the unexplained; not to explain the uninvestigated.

User avatar
Lausten
Persistent Poster
Posts: 3917
Joined: Sun Dec 06, 2009 6:33 pm
Location: Northern Minnesota

Re: Bayes Theorem. Is that still a thing?

Post by Lausten » Tue Mar 12, 2019 7:24 pm

landrew wrote:
Tue Mar 12, 2019 6:22 pm
Lausten wrote:
Tue Mar 12, 2019 5:28 pm
I should have said at the outset that I am aware of opposition to this theorem. I was hoping for some objections that I could grapple with. Something better than, "you can't do numbers on that".
I have stated my opinion, but I doubt you will find some universal final answer to that question, as you guys seem wont to believe. I have little faith in climate predictions as many others do, but this argument keeps circling back to the notion that such predictions are credible, when clearly they have not been so.
Who said anything about climate change in this post? We have a whole 'nother forum for that. And who said I was looking for a universal final answer? I don't even know what you mean by that. And who are "you guys"?
A sermon helper that doesn't tell you what to believe: http://www.milepost100.com

User avatar
landrew
Has No Life
Posts: 11208
Joined: Sat Jul 14, 2007 5:51 am
Location: Fox Meadows

Re: Bayes Theorem. Is that still a thing?

Post by landrew » Tue Mar 12, 2019 7:32 pm

Lausten wrote:
Tue Mar 12, 2019 7:24 pm
landrew wrote:
Tue Mar 12, 2019 6:22 pm
Lausten wrote:
Tue Mar 12, 2019 5:28 pm
I should have said at the outset that I am aware of opposition to this theorem. I was hoping for some objections that I could grapple with. Something better than, "you can't do numbers on that".
I have stated my opinion, but I doubt you will find some universal final answer to that question, as you guys seem wont to believe. I have little faith in climate predictions as many others do, but this argument keeps circling back to the notion that such predictions are credible, when clearly they have not been so.
Who said anything about climate change in this post? We have a whole 'nother forum for that. And who said I was looking for a universal final answer? I don't even know what you mean by that. And who are "you guys"?
Deflect, deflect, but never reflect.
The job of a skeptic is to investigate the unexplained; not to explain the uninvestigated.

User avatar
Lausten
Persistent Poster
Posts: 3917
Joined: Sun Dec 06, 2009 6:33 pm
Location: Northern Minnesota

Re: Bayes Theorem. Is that still a thing?

Post by Lausten » Tue Mar 12, 2019 8:52 pm

landrew wrote:
Tue Mar 12, 2019 7:32 pm
Lausten wrote:
Tue Mar 12, 2019 7:24 pm
landrew wrote:
Tue Mar 12, 2019 6:22 pm
Lausten wrote:
Tue Mar 12, 2019 5:28 pm
I should have said at the outset that I am aware of opposition to this theorem. I was hoping for some objections that I could grapple with. Something better than, "you can't do numbers on that".
I have stated my opinion, but I doubt you will find some universal final answer to that question, as you guys seem wont to believe. I have little faith in climate predictions as many others do, but this argument keeps circling back to the notion that such predictions are credible, when clearly they have not been so.
Who said anything about climate change in this post? We have a whole 'nother forum for that. And who said I was looking for a universal final answer? I don't even know what you mean by that. And who are "you guys"?
Deflect, deflect, but never reflect.
Do you actually have anything worthwhile to contribute?
A sermon helper that doesn't tell you what to believe: http://www.milepost100.com

User avatar
Lausten
Persistent Poster
Posts: 3917
Joined: Sun Dec 06, 2009 6:33 pm
Location: Northern Minnesota

Re: Bayes Theorem. Is that still a thing?

Post by Lausten » Sat Mar 16, 2019 3:58 pm

Here's another good intro

He starts with miracle claims from ancient history, cuz that's what he does. About half way through he introduces Bayes. About 3/4 of the way he points out that saying “expected evidence” is code for probability.

That is, naturally, we expect evidence. We use terms like that in our speech all the time. Rarely do we translate that into numbers, but that's what historians do. It makes it easier to spot where you missed something and easier for someone checking your work to see what you left out or miscalculated.
A sermon helper that doesn't tell you what to believe: http://www.milepost100.com