Dale and Dale's 'World's Greatest Guesser' and Bayesian apologetics

A 2019 Interpreter paper claims odds of about 10^-111 that the Book of Mormon is fiction; the number comes from multiplying 131 subjective guesses as if independent.

What the Church claims

Bayesian analysis of correspondences between the Book of Mormon and Michael Coe's The Maya shows overwhelming statistical evidence of ancient Mesoamerican origin.

Once the evidence is in, this claim: Untenable. The claim rests on a single calculation whose headline number is statistically meaningless, because the correspondences it multiplies together are not independent and the sponsoring organization's own summary effectively concedes the point. Nothing survives the arithmetic being wrong, which is why this fails outright rather than merely weakening.

The record

  • Bruce E. Dale and Brian M. Dale, Interpreter 32 (2019), compare 131 positive correspondences with Coe's The Maya.
  • Each correspondence received a likelihood ratio of 0.5, 0.1 or 0.02, and a billion-to-one skeptical prior was overwhelmed to a posterior of about 1.03 x 10^-111. The two published tallies of negative correspondences disagree: Scripture Central counts six (horses, elephants, iron, steel, copper, precious metals), the critique at Wheat and Tares counts eighteen.
  • Scripture Central's own summary concedes the study was not grounded in empirical probabilities and that the ratios rest on intuitive judgments, so different assessors could reverse the result.
  • Kyler Rasmussen's 23-part 'Estimating the Evidence' series in Interpreter applies the same Bayesian approach across many Book of Mormon topics.

The moving target

The Church has never cited the Dale and Dale analysis, and has never endorsed a statistical case for the Book of Mormon in any essay, manual or general conference talk. The Interpreter Foundation, which published it, is independent of the Church and says so. That matters in both directions: no member is obliged to accept the odds figure, and if the calculation is wrong, nothing the Church has taught falls with it.

The exchange

Dale and Dale; promoted by Scripture Central, LDS Living, Daniel Peterson

Even with every assumption tilted against the book, the cumulative weight of 131 hits is decisive.

The answer: The items are not independent (a society with kings has palaces, tribute and wars), so multiplying them is invalid. Most 'hits' are true of many complex societies and of Joseph Smith's Bible, so their likelihood ratio is near 1, not 0.02. The scoring also capped how much weight any negative correspondence could carry, although one secure anachronism is far more diagnostic than one generic parallel. The same method would 'prove' almost any historical novel.

What they get right: The authors published their scoring openly, which is why it can be checked.

Bruce E. Dale and Brian Dale, Interpreter Foundation

Bayes' theorem is the standard tool of inference in science, medicine and law. An objection to using it here is an objection to the conclusion, not to the method.

The answer: Nobody disputes the theorem. The dispute is over what is fed into it. Every one of the 131 likelihood ratios was assigned by the authors themselves, with no blind coding, no second rater and no published reliability check, and the product then treats the items as independent when a single assumption — a Mesoamerican setting — makes dozens of them move together. Multiplying 131 self-assigned, correlated numbers returns whatever the assigner put in, which is why the same machinery run by a skeptic on the same book produces the opposite answer with comparable confidence.

What they get right: Making the priors explicit is better practice than the vague 'too many hits to be coincidence' argument it replaced, because it can at least be audited.

Bruce E. Dale and Brian Dale; Kyler Rasmussen

Negative evidence was included and scored against the book, so the analysis is not cherry-picked.

The answer: The negatives are scored at mild odds while positives run to a thousand to one each, and the list of 131 items was assembled by advocates after the fact. With no pre-registered item list, the decision about what counts as an item is already the result. A procedure in which the analyst selects the evidence, scores the evidence and declares it independent cannot fail, and the authors have never stated what finding would have moved the number the other way.

Both cases at full strength

The skeptics' case

The method treats a list of resemblances as independent draws, assigns each a guessed probability, and multiplies. Any long, detailed narrative about an agrarian kingdom will score astronomically against any real agrarian civilization under that procedure. Reviewers inside and outside the Church said so within weeks, and Scripture Central now attaches a caveat that amounts to a retraction of the number.

The apologetic case

Informal Bayesian reasoning is how historians actually weigh evidence, and the Dales at least made their judgments explicit and invited re-scoring. A skeptic who thinks the hits are worthless should show it item by item.

What would change this

A pre-registered scoring by independent Mesoamericanists with a dependence model and empirically derived base rates.

Closely related

Sources

An independent project. Not affiliated with The Church of Jesus Christ of Latter-day Saints or with any organization cited here. Summaries are written here; short quotations are attributed and linked. Every page names its sources so that it can be corrected, and every page can be challenged.