The Interpreter; Bayes Theorem; Nephites and Mayans

[_Simon Southerton]

I disagree, Simon, because the flaws all work together to produce the result. The reason for the imbalance you describe is, in large part, due to their exclusion of many, many pieces of evidence that would result in at least weak evidence supporting a non-historical Book of Mormon.

In what way do we disagree? The imbalance of 131 supportive evidences against 18 negative evidences is the major problem. It is the major cause of their achieving the one in a billion, billion, billion, I could go on.......result. Tweaking the strength of a few positive or negative correspondences only makes a difference of a few billion. The whole thing was set up to succeed by starting with 131 for and 18 against.

[_Lemmie]

I disagree, Simon, because the flaws all work together to produce the result. The reason for the imbalance you describe is, in large part, due to their exclusion of many, many pieces of evidence that would result in at least weak evidence supporting a non-historical Book of Mormon.

In what way do we disagree? The imbalance of 131 supportive evidences against 18 negative evidences is the major problem. It is the major cause of their achieving the one in a billion, billion, billion, I could go on.......result. Tweaking the strength of a few positive or negative correspondences only makes a difference of a few billion. The whole thing was set up to succeed by starting with 131 for and 18 against.

Actually, it is the assumption that every single piece of evidence is independent of every other that causes that particular imbalance. That independence assumption is not in any way a rational or acceptable mathematical approach. It is laughable that this journal would say this paper passed a peer review by statisticians without addressing and correcting that independence claim. .

However, even that would be overcome if the author's assumptions about facts in The Maya and definitions regarding the likelihood ratios were applied consistently. For example, if the calculations were done correctly any one of the negative evidences should completely outweigh all 131 positive evidences, because the likelihood ratio denominator in the negative evidence case must be arbitrarily close to zero ( I've explained why in previous posts), thus making the LR = infinity. infinity divided by 10 to the 32 power is still infinity, and the posterior odds that the Book of Mormon is fiction would be infinitely large.

Correction of any one of the massive statistical missteps made by the Dales and exacerbated by the Interpreter peer review process would render this paper nonsensical.

[_Res Ipsa]

I disagree, Simon, because the flaws all work together to produce the result. The reason for the imbalance you describe is, in large part, due to their exclusion of many, many pieces of evidence that would result in at least weak evidence supporting a non-historical Book of Mormon.

In what way do we disagree? The imbalance of 131 supportive evidences against 18 negative evidences is the major problem. It is the major cause of their achieving the one in a billion, billion, billion, I could go on.......result. Tweaking the strength of a few positive or negative correspondences only makes a difference of a few billion. The whole thing was set up to succeed by starting with 131 for and 18 against.

We disagree, I think, because I don’t find it implausible that there are questions out there in the world where there are 113 pieces of evidence that support one side and eighteen the other. And in some of those cases, 113 very weak pieces of evidence taken together do not outweigh 18 pieces of very strong evidence.

So, I don’t see the 113 v. 18 as the fundamental problem. I see it as an artifact of the flaws in the methodology.

[_honorentheos]

I have a copy of the 9th edition of The Maya, and not the 10th. But unless Dr. Coe really went in a different direction with him latest edition I think it's giving the Dales' too much credit to say they were comparing what The Maya with what they think the Book of Mormon says. It seems like they took extensive liberties in interpreting both for their exercise

The difference is, they quote the language from the Maya. They don’t quote the language from the Book of Mormon — they only present their interpretation of the language. If they were to quote the actual Book of Mormon language, the weakness of the correspondences would be much more apparent.

I tend to agree with Analytics that their treatment of The Maya, including providing "quotes" involved so much selective reading it's not much different. Either way, they don't treat their sources with care, if they clearly treat the Book of Mormon with favoritism.

It doesn't take being a scholar of the Book of Mormon or even beastie-level knowledge of Mesoamerica to see on first reading of the Appendixes that they were calling hits that should have been misses. I recall multiple people calling out the parallelmania as blatant. In a sense it strikes me like Mormonism itself: it wasn't passing the smell test so the more complicated aspects of the paper seemed irrelevant. But that also holds true for those who view the Book of Mormon as sacred and find the results confirming if unlikely to change their view of it in a meaningful way.

But I think the work to point out the issues with the methodology and abuse of Bayes actually makes important points. When you read Brant Gardners comments it seems clear he views the correct place for making comparisons between the Book of Mormon and Mesoamerica as being in the kinds of books he had written. It seems moving it to statistics suddenly meant something wasn't passing the smell test for him, either. He is just confusing it coming from the Book of Mormon with it being the Dales' methodology. In that regard, I think the paper actually exposes how vulnerable the common apologetics are by showing how thin the evidence actually is.

When I made my half joking pass at calculating LR for a couple of their correspondences it really drove home how bullheaded a person had to be to ignore how often a "pencils down" point had to be ignored. The sample size and data sources prevent calculating honest probabilities. Like Lemmie pointed out early on, it runs into superlative all-or-nothing moments where you are dividing infinity by zero. Of course the Dales' decided not to do any of that as a result, and instead picked swag LRs because its impossible to do what they claimed to want to do and have defensible results. And in a sense I think it undermines LGT Book of Mormon apologetics because it doesn't take too much familiarity with the math to realise this. I hope your attempt to address it in depth and with rigor is successful. It could topple a few notions of how unassailable the Mesoamerican setting is once one accepts modern prophets aren't the authority on the subject.

[_Physics Guy]

The Bayesian methodology is meant to cover a multitude of sins in the individual pieces of evidence, partly just by snowing critics with mathematical mumbo jumbo, but mainly by generating such a fantastically small chance for the Book of Mormon being fake that even huge give-or-take factors in individual items of evidence will seem unable to change the final conclusion.

And indeed as long as one allows the long product of likelihood ratios to stand as a methodology, the item-by-item argument over evidence will be a long task with little to gain. Perhaps after pages of fierce dispute one might compel the Dales to revise a few of their LRs, perhaps even so much as to win a factor of a million. Normally if you get an opponent to concede a million-to-one odds in your favor, it's time to acknowledge their intellectual honesty and share beer because you've won the whole argument. In this case you'll still have a factor of 10^126 left to go, so you'll have to repeat that struggle twenty-one times.

The bogus Bayesian methodology is a pure illusion but the illusion has to be dispelled before it's worth even talking about anything else.

[_DrW]

Given the way in which Bruce and Brant seem to have tacitly acknowledged the massive problems with the Bayesian methodology in the the Greatest Guesser paper, and chosen instead to double down regarding defense of the correspondences, one is left to wonder if either of them really care about young Brian.

Looking at a few of the other publications on which Brian Dale is a co-author, there seems to be no indication of misuse of Bayesian Inference. As he stated in his biosketch and some of us have confirmed, Brian is co-author (and now and then first author) on original research and review papers having to do with MRI imaging. (The physics of MRI imaging are not easily mastered, as I'm sure Physics Guy would attest). I also have it on good authority that the statistics in a sampling of mainstream literature papers on which Brian is a co-author are well done and highly credible.

When working in a multidisciplinary group, there is sometimes the temptation to overestimate one's competence outside one's field because of the familiarity with processes and subject matter one gains from working closely with collaborators from those other fields.

In these situations, one can pick up the terminology and basic concepts of other fields, and may be able to carry on a reasonable discussion regarding the data and conclusions. Chances are good, however, that said individual lacks the fundamental understanding and hands-on experience to be able do the original work. (For our collaborative groups, the two most valued consultants were #1 - the statistician, and #2 - the inhouse editor - either one of whom would have sent the Greatest Guesser paper back to the drawing board in short order.)

Looks as if young Brian may have simply gotten in over his head trying to mimic the work of the real statisticians in the group - this in an attempt to help his father underscore the imagined significance of the Book of Mormon - Maya "correspondences" or "hits" on which Bruce and Brant now wish to focus.

Alternatively, Brian may have simply advised his father that 'all one needs to do is multiply', and Dad happily did so.

Great for the senior folks like Bruce, Brant, DCP, and Midgley. Not so great for young Brian, who it appears will be saddled with his name on this pseudoscience disaster for a long time to come.

[_honorentheos]

And indeed as long as one allows the long product of likelihood ratios to stand as a methodology, the item-by-item argument over evidence will be a long task with little to gain. Perhaps after pages of fierce dispute one might compel the Dales to revise a few of their LRs, perhaps even so much as to win a factor of a million. Normally if you get an opponent to concede a million-to-one odds in your favor, it's time to acknowledge their intellectual honesty and share beer because you've won the whole argument. In this case you'll still have a factor of 10^126 left to go, so you'll have to repeat that struggle twenty-one times.

The bogus Bayesian methodology is a pure illusion but the illusion has to be dispelled before it's worth even talking about anything else.

It seems like Bruce enjoys debating the veracity of the Book of Mormon, which his blog I linked to earlier shows is a hobby of his in which he's invested a fair amount of time. The paper has provided an opportunity for doing so in a venue tightly controlled by, well, someone. I noticed at least one comment from a poster that made it past moderation that was later taken down. It had challenged the call to focus on the correspondences. My guess is that the post was inoffensive enough to not warrant moderation but someone, I'd guess Bruce Dale, had it taken down to enforce the current aim of the comment section. People taking notice and debating him on friendly turf about the Book of Mormon's truth claims appears to be all the win he needs.

His blog post notes something about his early understanding of Bayesian analysis, and in that it appeared Brian may have provided him with some literature and information but Bruce had not developed a detailed understanding of it. It certainly left me with an impression that Brian was helping his dad out, but defending the Book of Mormon was not a hobby of his or a place he seemed to invest nearly as much as his father has. I'm not surprised to see Brian's involvement in the comments be minimal, and mostly superficial touches on the background behind some of their thinking but no attempt to address any criticisms in detail or actively provide information that was brought up as flawed. I don't think he will, either. But maybe I'll be surprised and he'll ruin a Memorial Day weekend to stick it to some internet critics of their paper.

So...what would happen if someone developed and posted an honest attempt at a calculation of a likelihood ratio for one of the 20-odd correspondences Bruce wants to debate? I don't think the problem is one that can be talked away. Otherwise, it would have gone away already. I do think it needs demonstrated away. Like one has to with a not-too-bright student, sometimes walking through the problem up to the point their approach stops working and letting them come to that realization may be the only way to get through. If the Dales could be enticed to do even one LR calculation in full, I suspect it would advance the conversation on the methodology rapidly towards revision.

[_Dr Exiled]

In their abstract, Drs Dale say the following:

There is overwhelming evidence that the Book of Mormon has physical, political, geographical, religious, military, technological, and cultural roots in ancient Mesoamerica. As a control, we have also analyzed two other books dealing with ancient American Indians: View of the Hebrews and Manuscript Found.

These are pretty big claims, obviously. Yet, aside from DNA, another problem I see (and forgive me if this has already been discussed) is that the Dales limit their control group to just the View of the Hebrews and Manuscript Found. They don't do any comparisons to Joseph Smith's outside world that was known to him. For example, in the latest smug challenge from Dale Sr., he demands to see if anyone can refute his calendaring comparison. Yet, I don't see this as remarkable given that the world used a calendaring system with days, months and years. Also, the way the earth rotates every 24 hours and rotates around the sun, lends to having a calendaring system like the world does. So, is it really that remarkable that the Maya used such a system that Joseph Smith was familiar? At a certain point, it devolves into well the Maya liked to drink water and so did the Book of Mormon people, so ......

[_SteelHead]

In their abstract, Drs Dale say the following:

There is overwhelming evidence that the Book of Mormon has physical, political, geographical, religious, military, technological, and cultural roots in ancient Mesoamerica. As a control, we have also analyzed two other books dealing with ancient American Indians: View of the Hebrews and Manuscript Found.

These are pretty big claims, obviously. Yet, aside from DNA, another problem I see (and forgive me if this has already been discussed) is that the Dales limit their control group to just the View of the Hebrews and Manuscript Found. They don't do any comparisons to Joseph Smith's outside world that was known to him. For example, in the latest smug challenge from Dale Sr., he demands to see if anyone can refute his calendaring comparison. Yet, I don't see this as remarkable given that the world used a calendaring system with days, months and years. Also, the way the earth rotates every 24 hours and rotates around the sun, lends to having a calendaring system like the world does. So, is it really that remarkable that the Maya used such a system that Joseph Smith was familiar? At a certain point, it devolves into well the Maya liked to drink water and so did the Book of Mormon people, so ......

As stated on reddit at the beginning of the saga - "Parallelmania in Bayesian drag".
Now:
"Parallelmania in Bayesian drag"+ GIGO = a polished turd.

[Guest]

As noted by many, the independence issue overwhelmingly skews the results the Dales reach. It seems common sense to note that individual statements in a book, by a single author, about a single story line, cannot be rendered as independent events, but the Dales have persisted, so I thought I would try explaining it using their own words.

Looking at their list of statements that do not support the Book of Mormon being historical, we find this on page 92:

We were able to find six such points of disagreement between The Maya and the Book of Mormon, namely the existence of (1) horses, (2) elephants, (3) iron, (4) steel, (5) copper and (6) refined gold and silver.

Note that #6 includes two items grouped together: refined gold and refined silver. Their explanation for grouping the two together immediately follows the quote above:

(We combine refined gold and refined silver instead of considering them individually because gold and silver are usually found together, and thus to refine gold is also to refine silver.)

In other words, if a statement in a book discusses refined gold, they consider it to NOT be a statement independent of those statements discussing silver, so at least in this case, they admit that there is a dependency among at least some statements.

Additionally, a mention of gold should be used in only one independent correspondence, any other mentions of gold and or silver would constitute statements that are NOT completely independent, by virtue of depending upon the related element. In violation of their own independency definition, however, gold and/or silver is mentioned in support of at least 5 different correspondences in Appendix A and B, thus removing the independency from those 5 likelihood ratios. So just in this one quick check, using the authors own words, five LRs are found to be dependent.

Going further, note the way the 131 "independent" statements from the Maya are defined:

Specifically, we have found 131 such correspondences. We divide these correspondences into six separate categories:
• Political (33 correspondences)
• Cultural/social (31 correspondences)
• Religion (19 correspondences)
• Military/warfare (12 correspondences)
• Physical/geographical (13 correspondences)
• Technological/miscellaneous (23 correspondences)

In each category, the authors are asserting that every single correspondence is considered as occurring independently of every other correspondence both within and between groups, but they group them by commonalities that by default suggest dependencies!

In order for two events to be independent, the existence of one event has to have no influence on the existence of the other, so to show how dependent statements within a common group can be, consider two examples of relatedness: if someone was writing a story where they describe "warfare with ambushes and traps [group 4, # 7]," what are the odds that they might also describe "raids to take captives/slaves [group 4, # 8]" ? What would be the likelihood that telling a story where "fighting with 'darts' [group 4, # 5] " might also include a description of "thick clothing used as armor [group 4, # 4]" ??

Their assertion of independence of each correspondence is not a rational position.