"/>

Aug 06 2012

# New Bayesian Calculator

Thanks to Cam Spiers (who has produced an interesting selection of free javascript Bayesian Calculators), I have updated my own Bayesian Calculator page using the most basic of those. This might be updated again in coming months. Right now it only allows running calculations with two-digit probabilities from .01 to .99 (or 1% and 99%), so you can’t use it for odds outside that range (for example, you can’t see what happens when the prior is 1 in 1,000 or 1 in 1,000,000 or when a consequent is even closer to 100% than 99%). But future versions of the page might have those features.

For people new to the whole idea of Bayes’ Theorem and Bayesian reasoning, you should first check out my talk at Skepticon last year: Bayes’ Theorem: Lust for Glory! For a more thorough treatment (using historical reasoning as a running example), which is also aimed as much as possible at lay readers, there is now of course my book: Proving History: Bayes’s Theorem and the Quest for the Historical Jesus.

1. 1
##### Brian

Richard, I’ve been checking Amazon for a few months now, but still no kindle version….:(

1. 1.1
##### Richard Carrier

I know. I’m as annoyed as you are.

2. 1.2
##### Brian

Not your fault Richard. I should’ve just bit the bullet and bought the dead tree version. How’d the July course which used the book among others go?

3. 1.3
##### Richard Carrier

July course went superbly.

And I just got word the kindle edition will drop in the next few weeks.

2. 2
##### F [is for failure to emerge]

I sit and await a rebuttal.

3. 3
##### Michael Borland

For those who use Android devices, I have a no-cost, ad-free app that you can use. In addition to the simple calculation, it allows you to provide minimum and maximum values for each probability, based on which it gives you the minimum and maximum values of P(H|EB). Just search for my name on Google Play.

1. 3.1
##### Michael Borland

Richard, after checking out your on-line calculator, I made some improvements to my Android app by adding sliders. To deal with the slider resolution issue (which I guess is what is limiting you to values of 0.01-1.00), I added *10 and /10 buttons that rescale the sliders. E.g., the slider defaults to a range of 0.01-1.00. Pressing the /10 button makes it 0.001-0.100, then 0.0001-0.0100, etc. The *10 button reverses this.

I was also toying with the idea of a logarithmic slider. The top end could be 1 and the bottom 10^-6, for example. However, I was concerned that 10^-6 wasn’t small enough for theological computations.

Hope this is useful. Sorry to bore the others with coding talk.

2. 3.2
##### Richard Carrier

That’s a great idea. I’ll look into doing the same.

4. 4
##### Hjalti

….there is now of course my book: Proving History: Bayesâ€™s Theorem and the Quest for the Historical Jesus.

James MacGrath has posted a review of your book: Review of Richard C. Carrier, Proving History.

1. 4.1
##### Richard Carrier

Thanks. I was aware of that, but I might not have been, so I appreciate bringing it to my attention. MacGrath’s review is both praising and critical, and his critical remarks generally misunderstand my arguments and Bayesian reasoning, but are nevertheless quite interesting and intelligent and will make good examples to teach from. I’ll blog about it in coming weeks (I’ll be largely AFK for a while, first at the Atheist Film Festival, then in studio for a couple weeks for an audiobook publisher, recording my books for them…so anyone out there who has been waiting for that to happen, it’s very soon!).

5. 5
##### Steven Carr

If Jesus came from a really small place, doesn’t that reduce the prior odds of his existing?

Isn’t that something to be factored into any Bayesian calculation of his historicity?

Suppose we are told that Sally has very strong views of female equality.

Which is more likely?

Sally is a bank teller.

Sally is a bank teller and a feminist.

Sally is a bank teller and a feminist and was born in Red Bank, New Jersey.

1. 5.1
##### Richard Carrier

You are referring to the conjunction fallacy. It does not apply here, because prior probability is a relative probability, not an absolute probability. The prior probability that “I was struck by lightning five times” is true is not the probability of being struck by lightning five times, but the probability that someone who says they were is telling the truth (see Proving History, p. 51, with n. 11, p. 301).

Similarly, the prior probability that Jesus is historical is the probability that any Judean is historical, which will be the same as the probability that anyone in the world is historical, unless there is an unusual frequency of fictional people in Judea, or an unusual dearth of them there. We have no reason to believe anything unusual about Judea in this respect, so the probability is the same (if, worldwide, 1 in 1000 persons claimed to be historical are not, then all else being equal, in Judea 1 in 1000 persons claimed to be historical are not).

IMO, this is a mistake made by Michael Martin in The Empty Tomb (where his calculation of priors in chapter 2 is simply invalid).

Switch to our mobile site