Originally Posted by tanya

Howdy Doug, no, sorry to write that I think you missed my point, entirely.

The citation from Detrano et al was not haphazard, as it may have appeared.

They concluded that Bayes' Theorem (which they did employ properly) was very useful, in predicting the existence of significant coronary artery disease. They based that optimistic conclusion on the fact that Bayes' theorem accurately predicted the existence of disease in MOST patients PRIOR to their undergoing confirmation by coronary angiography.

My point then, was simple: Bayes' theorem is very useful in predicting probabilities where the consequence of error is outweighed by the larger quantity of cases in which the theorem yields predictions which turn out to be accurate. In the case of heart disease, of course, one would never rely upon Bayes' theorem, because it is imperfect.

More folks are accurately diagnosed, using Bayes' theorem, than are inaccurately diagnosed using Bayes' theorem, however, if one's coronary arteries are blocked, one must ascertain, with certainty, prior to surgery, the extent and nature of the blockage. Having a high probability is nifty, if one is betting at the racetrack, but not if one is betting one's life.

So, no, Detrano et al did not discover anything that disproves Bayes' theorem, if anything, on the contrary, they showed that it is quite useful in predicting heart disease. However, its reliability is insufficient to base a clinical decision on the outcome of this computation.

So, then the question arises, of what utility could employment of Bayes' theorem provide, investigators attempting to elucidate the origins of Christianity?

I reply to that question: NONE

Why? For me, and perhaps, singularly me, the probability that Irenaeus wrote xyz, or Clement of Rome wrote abc, is not interesting. I want to know, as with coronary artery disease, precisely, exactly, who wrote what, when.

Assigning a probability of 87% to the likelihood that the LAD is 100% obstructed, is fine, but if it is Sheshbazzar's LAD that is blocked, then, I want to know that fact, definitively, not probably. I require certainty, not likelihood. One needs to establish, before cutting, that the surgical approach proposed, to unblock LAD, will resolve Sheshbazzar's problem.

The history of Christianity's origin will not be revealed, no matter how many arithmetic manipulations are performed.

Excavation, supervised recovery, transparent, fully televised, archaeological inquiry, offers the key to learning the origins of Christianity. Put the mathematics books away. They are useless for this inquiry.

Originally Posted by ApostateAbe

...
tanya, for me, it is all about probabilities, and, if we could apply Bayes' Theorem with its numerical percentages of probability, then that would be the best way to go. The main problem seems to be simply that we can't, at least without pulling an ass-load of numbers completely out of our crap piles.

Instead, I think it is best to make decisions with estimates of relative probability, i.e. this interpretation is more probable then that interpretation, having in mind the patterns, the context, and the historical plausibility. Those things you simply cannot effectively quantify with numbers.

Originally Posted by Toto

But here you are pulling your relative probabilities out of your own crap pile, as you put it. Using Bayesian statistics would force you to be more rigorous, instead of just squinting at the data and proclaiming your own theory the winner.

Originally Posted by ApostateAbe

...
So what is your opinion for why such an assertion is not yet demonstrated? How is the proposal for the application of Bayesian statistics to history better than the proposal for a time machine?

Originally Posted by Toto

What assertion are you talking about?

Richard Carrier is in the process of trying to convince professional historians that they need to use Baysian statistics (or that they have been using it all along without realizing it, if they are reasoning correctly.) He's been working at this for only a few years. Progress takes time.

Originally Posted by Toto

Richard Carrier has written a book on the subject which is now being peer reviewed and is scheduled to be published by Prometheus Books.

He started this project after he got his PhD. The complications have been the usual ones of finding a publisher, etc. I don't think you can describe Carrier as "unemployed." But your entire post makes little sense.

Originally Posted by ApostateAbe

...
I didn't get that news. Who is reviewing the book?

Originally Posted by tanya

My point then, was simple: Bayes' theorem is very useful in predicting probabilities