Interpretering Bayesian Analysis

[kyzabee]

Thanks for the chat Dr. Moore!

You SERIOUSLY *HAVE* to engage his comments. He is correct. If you are just having fun, that is one thing, but if this is any kind of serious, and you really expect others to view you as credible, then you have to be credible. This flippant attitude here concerns me. Am I to even bother taking what you say seriously then? Sure Sic et Non will love you and praise you in high brainwashed fashion, but is that low bar all you go for? Really?! Aren't you even mildly curious about the truth of any of this?

I shall know far better after you bring out your DNA materials and probability analysis of this issue. You need to be seriously good here because there is a lot of ways to cheat, and being here.... you WILL get caught cheating if you do so. Are you up for doing it correctly when that is discovered? I have seen apologists constantly cheating on this issue either with the evidence, or misstating the background knowledge. I sincerely hope you are a better man than they, and yes, I include the FARMS and FAIR idiots who have tried and failed. You truly need to branch out past that to do anything of significance, please. Please do so and be real about this I really want to see your number, and your work showing how and why you get the numbers you do on the DNA stuff. You don't strike me as an idiot like most Mormon apologists are, please don't become one here with the DNA analysis.

Eh, I've done my best to answer his questions in as straightforward a manner as I can. The points he's brought forward I've tried to address (and there was nothing particularly new in that post). If he's done engaging, he's done engaging, and that's fine. He knows where to find me.

You were doing such a great job of making me feel welcome. It's starting to feel a bit less so.

[kyzabee]

While I do appreciate the effort to be affable around so much critical feedback, it’s already very obvious to me that Kyler is wasting everyone’s time.

With the volume of critical feedback he’s already received, we have seen zero willingness to adjust process, acknowledge deeply erroneous assumptions, or even to accept an offer to pay for free input from a paid consultant from BYU. Think about how ridiculous that would be in ANY academic setting. “Thanks for your questions but I’m right and you’re biased” is what his messages here amount to.

So… I’m not going to waste the time engaging with him anymore. The whole project might be fun for him, but it clearly isn’t serious.

I’d like to see Dr. Rasmussen take one item from his set of values, and ‘run it through’ a Bayesian analysis right here, so Cassius faculty can discuss his process, in a contextual and pragmatic manner. We need to see how one hit or miss is used within Dr. Rasmussen’s process.

- Doc

Kind of feels like the episodes already run you through it pretty clearly, so you wouldn't necessarily need me for that. But if you'd like to have a piece-by-piece discussion for the DNA one when it comes out, I could probably be persuaded.

[Doctor CamNC4Me]

Well, yes. To break this down in the simplest manner I can, I don’t see why you can’t identify a ‘hit’ off the top of your head, and then simply reproduce the analysis in a post. At that point you and the folks here can discuss each step and its reasoning with regard to your finding. For example:

1. ‘Cat’ is a hit for the Book of Mormon.

2. Here’s my reasoning.

3. Here’s the math. x+y+z = cat is a hit

4. Discuss x, y, z, reasoning, and result

- Doc

[Philo Sofee]

You SERIOUSLY *HAVE* to engage his comments. He is correct. If you are just having fun, that is one thing, but if this is any kind of serious, and you really expect others to view you as credible, then you have to be credible. This flippant attitude here concerns me. Am I to even bother taking what you say seriously then? Sure Sic et Non will love you and praise you in high brainwashed fashion, but is that low bar all you go for? Really?! Aren't you even mildly curious about the truth of any of this?

I shall know far better after you bring out your DNA materials and probability analysis of this issue. You need to be seriously good here because there is a lot of ways to cheat, and being here.... you WILL get caught cheating if you do so. Are you up for doing it correctly when that is discovered? I have seen apologists constantly cheating on this issue either with the evidence, or misstating the background knowledge. I sincerely hope you are a better man than they, and yes, I include the FARMS and FAIR idiots who have tried and failed. You truly need to branch out past that to do anything of significance, please. Please do so and be real about this I really want to see your number, and your work showing how and why you get the numbers you do on the DNA stuff. You don't strike me as an idiot like most Mormon apologists are, please don't become one here with the DNA analysis.

Eh, I've done my best to answer his questions in as straightforward a manner as I can. The points he's brought forward I've tried to address (and there was nothing particularly new in that post). If he's done engaging, he's done engaging, and that's fine. He knows where to find me.

You were doing such a great job of making me feel welcome. It's starting to feel a bit less so.

That's because your last few comments appear to me you have already come to conclusions and I haven't even seen your evidence yet, let alone any analysis. I am looking VERY forward to it, but like I say if all you are going to do is use Bayes to come to a testimony conclusion that's a major fail. I really want to see Bayes used accurately here, not in an apologetic manner. I am looking forward to your DNA materials and other areas concerning the Book of Mormon. Just remember, if some of us bring out other background materials you have not thought of or other evidences you have not assessed, it ALL HAS to go into the hopper, there is no skipping it if it is a probability you are looking for and not just supporting a conclusion you hope for. That's why it is an ongoing investigation that could take a while. We are in no rush whatever. I just simply demand you use Bayes properly. I am sick of seeing the cheating from the apologetics side. You are better than going that route.

[Doctor Scratch]

Greetings, Dr. Rasmussen, and welcome! I have to say, your project is somewhat interesting, though to be honest, I think it's misguided. You mentioned elsewhere that all you really wanted was to generate "amicable" conversation, which is great and all, but why not aim higher? For example, why don't you do an analysis to determine whether the Book of Mormon took place in Latin America versus the Heartland? There is talk that the Heartlanders are going to cause a "schism" in the Church, so you'd think that addressing the issue is a rather important matter. You already have belief in the basic narrative of the Church as common ground, but if Bayesian analysis can be used to show that it's more likely that the Book of Mormon took place in the Yucatan, why not do that?

You also mentioned that you were inspired by Dr. Peterson. Perhaps you also know that one of the issues that upsets him the most is the accusation that he has (at times) been paid for his apologetic work. Why not do a Bayesian analysis to determine the probability of whether he was paid or not? You could take the official IRS documents as one piece of evidence, and weigh it against the thousands of denials he's made over the years. Just think what a relief it would be to him if you were able to show, via your statistical work, that it's highly unlikely that he ever collected any compensation.

[Philo Sofee]

I like the recent suggestions kyzabee.

Why don't you go ahead and give us one of your evidences for the Book of Mormon and against it, and do a Bayesian analysis of it right now? That can get the wheel turning... I think Scratch has a dang fabulous suggestion on the Mesoamerican vs the Heartland model of Book of Mormon geography myself. That IS critically important actually.

I mean chiasmus also would be interesting. Or DNA, or whatever. But pick one and take it through the math and show us your work. Lets get started discussing a specific.

[kyzabee]

You already have belief in the basic narrative of the Church as common ground, but if Bayesian analysis can be used to show that it's more likely that the Book of Mormon took place in the Yucatan, why not do that?

If this interests you, you should pay close attention to Episode 15.

[kyzabee]

Well, yes. To break this down in the simplest manner I can, I don’t see why you can’t identify a ‘hit’ off the top of your head, and then simply reproduce the analysis in a post. At that point you and the folks here can discuss each step and its reasoning with regard to your finding. For example:

1. ‘Cat’ is a hit for the Book of Mormon.

2. Here’s my reasoning.

3. Here’s the math. x+y+z = cat is a hit

4. Discuss x, y, z, reasoning, and result

- Doc

We can do that for the DNA analysis right now, as it ended up being relatively simple on the math side.

The lack of Lehite DNA represents evidence against an authentic Book of Mormon with an evidence score of -3 (decreasing the probability of an authentic Book of Mormon by 3 orders of magnitude).

In terms of how likely we would be to observe that lack of DNA if the book was a fraud, I used p = 1. I have a line that if the book were fraudulent, we'd be as likely to expect Lehite DNA in the Americas as Hobbit DNA in Wales.

In terms of how likely we'd be to observe the lack of DNA if the book were authentic, I split it up into separate estimates for Autosomal and Sex-Linked DNA, which I treated as independent.

For autosomal DNA, I took a look at a study used by Simon Southerton tracking historical migrations worldwide. The simulation data for that study found that 6% of real migrations would've been missed. Despite Lehi's genetic contribution likely having a much smaller proportion of DNA than average, and was at a greater time depth than average (both of which would've made it easier to miss), I used p = .06 as the estimate for the likelihood of not observing Lehi's DNA if his migration was real.

In terms of sex-linked DNA, as faithful, critical, and neutral scholars all seemed to agree that the sex-linked DNA could've been lost to time due to genetic drift and population bottlenecks, and as there are quite a few examples of this occurring in real life, it didn't seem appropriate to use an astronomically low estimate. I made an attempt at modeling it myself using python, and learned some interesting things, but none of my attempts appeared to give me anything realistic. I ended up assigning the lowest Bayes factor as used by the Dales (and secular scholars), which was p = .02 (note that the Dales' own analysis gave DNA as a whole a p = .5).

Doing the math, you can multiply .06 by .02 to make p = .0012, which, when taking the log10 of that value (after dividing by our probability value for the hypothesis of fraud of 1), gives us an evidence score of -3.

[drumdude]

I think this needs to be discussed further.

10^-42 is the bullseye drawn around the arrow. There’s no reason it couldn’t be -40, -30, -50, -60 etc. -42 is completely arbitrary.

You can spend all your time describing the flight of the arrow, air resistance, gravity, but at the end of the day it doesn’t change the fact that all you have accomplished is drawing a bullseye around the arrow you fired.

You're right. The prior can be whatever you want it to be. I figured a value that represented double the orders of magnitude that others had applied to things like time travel would be sufficient here, but you can pick whatever value matches your level of belief.

What order of magnitude do you end up with when you're finished?

If you're crossing the same distance in order of magnitude as there is between atoms and galaxies, I think this indicates you have gone very wrong.

[drumdude]

Also, to level set, here's the order of magnitude of confidence we have in quantum mechanics. It is widely cited as the most accurate model of reality that has been created by man:

Quantum physics is probably the most precise scientific discipline ever devised by humankind. It can predict certain properties with extreme accuracy, to 10 decimal places, which later experiments confirm exactly.

https://phys.org/news/2019-04-common-my ... ysics.html

Somehow, you will prove that the Book of Mormon is at least 30 orders of magnitude more likely to be true than our theories of quantum mechanics. Can you begin to see how ridiculous this all is?