Jayson
Why do you characterize those who disagree you as "spitting" when they do their work? They used to be "running." What we would actually be doing in the situations you describe is called "modeling." No bodily fluids, netiher sweat nor spit, are necessarily involved.
I wasn't characterizing anyone with slander; that was not the intention.
I have mentioned previously that part of my day job is programming (which also means that I get tasked with anything to do with computers or models of any form).
If my manager comes to me and asks about a particular prediction model that we have for call volume to staff capacity, and wants to see how a given change might affect the outcome, then I would say the vernacular to her of something like, "I'll go run the computations and see what values spit out".
So I wasn't intending to belittle anything; I would use the same slang for myself (and have) regarding the results (I picked up this term of "spit" from my Mother's computer teacher back in the early 80's, who had been working on computer systems since back when they used punch cards, and he referred to the output as "spitting" in reference to a printed out result, which tended to be pushed out at the end of the printing with a final "spit"-like action).
I apologize if this offended you; it surely was not the intention.
The interpretation of probability which Bayesian methods implement is especially apt for unique, unprecedented and otherwise irreproducible eventss. Even the most determined Bayesian critic, Karl Popper, conceded the usefulness of probability for such events. He rejected the Bayesian view, of course, but championed a very similar propensitist interpretation.
Even a frequentist would point out that it is not their teaching that an event is literally repeated, but that the event belongs to a "reference class" of events, plural. Every event which can be uniquely identified is necessarily unique. Frequentists do not imagine that Groundhog Day was a documentary.
Yes, and the reference class is what I am referring to.
You need more than one event in the reference class if you want a reasonable number in the output.
I am not stating that you cannot derive a value, only that the value is not reliable beyond the limitations of the data entered into it.
I don't need a BT model to understand, for instance, that the Jesus tome is limited in like data to enter into the reference class.
The entire Jesus phenomenon is like one bit of data.
The way Carrier solves this problem is by comparing the data within the accounts against known data outside of the accounts.
That's one solution, but that assumes that the text's data indicates whether or not there was or was not a Jesus.
Instead, it would be far more helpful if we had 3 such like figures and some verified standing on their historicity, in which then the question we would be capable of asking would be whether it is likely that the central figure of these texts existed, given that a body of theological texts arose around the central figure.
Unfortunately, Buddha is about the only other figure with a similar phenomenon surrounding the figure, and Buddha is no better attested than Jesus when critically examined beyond historical inheritance.
And what method can, Jayson, in your view?
I didn't offer that there must be one.
I don't think there must be a best choice when the data is too low to suffice for such processes.
If we only have one shot fired in artillery before the high-tech sensory age (or in the absence of such availability), then we don't know where the target is.
We can't force a solution from one shot; as much as I would appreciate such.
We must wait for more shots, or we must begin to make shots our selves to narrow the probability regarding the location of the target (assuming that it is a mostly stationary target).
Or to put this another way, and I am not attempting to be snide or rude, BT cannot tell you where an opponents ship is with one shot in Battleship.
It can give you a probability, but the individual must keep in mind that the first probability is not the probability to rest on, and that more data will need to be entered later to refine the probability.
Jesus, having only really Buddha in like fashion, is almost exactly akin to a single shot probability.
The only way we'll get a 'second shot' pass on the model is if we find out that Buddha did or did not exist as an entirely solid proven with no room for debate (let's say we find his skeleton, or something of that caliber whereby it is Archaeological proof that verifies the matter; or we find a text talking about the creation of the Buddha myth prior to the theological text's dates).
Then we could run (hope that doesn't bother you) the probability through a second pass.
That would help refine the probability, but it would still not rule out the error of the probability due to limited data.
Turing, for example, was rather successful, in part, due to the shear amount of bits of data available for multiple passes and refining the probabilities.
Turing did not run a single bit and let it then determine what the final probability was.
Does this mean BT is useless?
No.
Does this mean that BT is a bad tool?
No.
Do I think more folks within the Historical society could benefit from applying BT?
Yes.
Do I think BT should outright replace the entire Historical Method?
No.
I'm not equating anyone to anything, nor attempting to belittle anyone for appreciating BT.
I am, instead, stating that what was earlier offered was a bit off the mark in just resting the Jesus case of historicity on BT entirely.
I would hold the same about the application of BT for the historicity of Buddha, and I would say that the same could be argued regarding Hipocrates and nearly all Kings of Egypt of the 13th dynasty; though these figures are not directly comparable to Jesus and Buddha.