$30k challenge to Interpreter’s “Team Bayes”

[Philo Sofee]

I have constructed a mirror argument to show that Kyler's logic cuts both ways. He starts with an extreme skeptic and multiplies probabilities to move the skeptic's belief towards the authenticity of the Book of Mormon. One can just as easily start with an extreme believer and multiply probabilities to move the believer's belief towards that Book of Mormon being a fraud.

Hey, that’s cherry picking. No fair!

No, it is FAIR......get it? FAIR - Foundation for Apologetic Information and Research.......

[Lem]

I don't think KR wants to admit how biased his approach is. Here's a recent response to someone noting the bias:

Daved6:

....I'm sorry but the book doesn't look more ancient because it's unlikely Joseph could have written it. It doesn't look ancient because the number of words per section (or book within it) are larger than the number words per book in the Bible.

----
KR:
....Here's a different analogy, one tying a little bit closer to how it's framed in the essay. Say that I have a random friend living in the early 19th century. He comes to me, having no history of publishing books, or indeed of having written much at all. He says "I've written an 876 page book." Based on the data I've collected, it's clear that such would be a very uncommon feat.

On the basis of probability, I reply "Yeah right. Prove it." If he did write it, he could quite easily furnish a bunch of evidence to support his claim: stuff like a manuscript and early manuscript drafts written in his hand.

That type of evidence would quickly and dramatically overwhelm my p = .0006-level skepticism of his original claim, and I'd be forced to admit that he was correct.

http://disq.us/p/2i3vb1p

So, using KR's own analogy, a reasonable response to his hypothesis re: ancientness and length would be:


"He comes to me, having no history of publishing books, or indeed of having written much at all. He says "I've received gold plates from an angel, and used my seer stone to record their contents into a book." Based on the data I've collected, it's clear that such would be a very uncommon feat.

On the basis of probability, I reply "Yeah right. Prove it." If an angel really did bring gold plates and he used a seer stone to get the contents, he could quite easily furnish a bunch of evidence to support his claim: stuff like showing me the gold plates, evidence supporting the existence of angels, evidence from language experts that not only were the writings in the plates were translated correctly, but that the language inscribed on the plates was a real language from a real group of people, earlier success with his seer stone working, evidence showing the existence of the Book of Mormon peoples in real world history, etc...


That type of evidence would quickly and dramatically overwhelm my p = .0006-level skepticism of his original claim, and I'd be forced to admit that he was correct."


On the basis of probability, KR is asking for proper evidence of 19th century writing, while his evidence relies only on coming to the conclusion that he assumed in his starting conditions, that length supports ancientness.

[Philo Sofee]

In other words, he is using Bayes "accurately"....

Lem, your grasp of Kyler's assumptions and how he is so badly bungling things is stellar. It is not a thankless job posting your observations. I for one am seriously grateful for you, as well as so many others, posting and sharing insights into the strenuous difficulty apologists have even attempting to be objective. Apparently they just can't do it.

[Dr Moore]

I've been in touch with Kyler Rasmussen. He's made it clear that he has no interest in bringing his statistical analysis to the teller window to collect $10,000. I've attempted to appeal to him privately with examples where his statistical assumptions are likely to break down, as if we were a team, but rather than consider my examples, he accuses me of not engaging a good faith discussion because I haven't yet looked at all of his arguments. Kyler further says that what I'm proposing isn't feasible or worthwhile, would require (at best) a herculean effort with minimal payoff.

Well, rather than "give up" I'm going to try again. I hope something will connect. If these guys won't do their homework, then moving the Mopologetic enterprise into the hard sciences with statistics to bolster their arguments will surely backfire, making their work even less credible and less relevant than it already is.

Below are 2 reasons why.

First: rule #1 of probability multiplication is that you cannot multiply probabilities if they aren't independent functions

What we see in these astronomical odds arguments in support of Joseph Smith restoring genuine ancient texts by John Gee, Bruce & Brian Dale, and now Kyler Rasmussen, is, at its core, multiplication of many probabilities without ANY effort at establishing independence among those probabilities.

At the risk of being too wordy, the rule is simple: the probability of N "events" happening together is

P(A AND B AND C AND ... AND N) = P(A) * P(B) * P(C) * ... * P(N)

IF AND ONLY IF

A, B, C, ... N are each independent processes.

I've offered $10,000 to Interpreter if Gee, Dales and/or Kyler will do the work to explain/show/demonstrate/argue why and how their N probabilities are independent from one another.

I am not making a pedantic, esoteric challenge here. I'm asking these guys to take their hundreds of pages of analysis and add the rigor, whether with data or with clear logical arguments, of showing that independence. Doing so would have a HUGE payoff (sorry Kyler, you are wrong) for at least two reasons. First, it would qualify you to use probability multiplication in the first place. Second, and more importantly, it will make each of your "event" probabilities, whether Bayesian or something else, qualify as statistically valid data points.

Ahem, for what you're trying to do -- take a near-impossible proposition and show why it's actually mathematically probable -- NOTHING ELSE WOULD GIVE YOU A HIGHER PAYOFF.

This is especially vital for attempts at utilizing Bayesian approaches, like Kyler is doing here. Why? Because each point probability, P(A), P(B) ... P(N) is itself a multiplication of probabilities!!! Because of that, buried in the components of every one of Kyler's Bayesian point conditionals is one thing: what was and was not in Joseph's mind due to his environment, exposure to books, exposure to stories, exposure to languages, the creativity of his imagination, his intelligence, his charisma, his self-image, and the strength of his memory.

Because of that, which by the way has yet to be addressed in any sort of statistically relevant context by any of Team Bayes, I do agree with Kyler's realization -- that it is not feasible to assert anything statistical about what was and was not in Joseph's mind and therefore it is a waste of time to assert probabilistic assumptions about most of Joseph wrote. But that's kind of my point: if these clowns would even attempt to show the rigor demanded by probability multiplication and Bayesian analysis in the first place, they would immediately expose the uselessness of these probabilistic tools for their purposes.


Second: statistical analysis requires controlling for all information; otherwise you have no basis to assert a valid conclusion

I'm not going to spend a lot of time on this topic, because >60 years of Mopologetics makes it clear that control, control, control is not the goal.

But I will show a graphic, created using a nifty online tool by yours truly.

What Team Bayes are really doing is even more intellectually bankrupt than skipping steps in the operation of advanced statistical machinery. What they're doing is carefully cherry picking data to demonstrate that for N dice rolls, some number close to N turn up "sixes" in Joseph Smith's favor. There may be a few "misses" included for a patina of credulity, but ultimately it's Texas Sharpshooter, as our Dean already pointed out.

Here's the graphic, to illustrate. Notice the "misses" included in the top right corner. This, in very essence, is what Team Bayes are doing.



And now I'm sure Kyler will rush to defend himself, and so will the Dales and Gee. "We look at all the evidence!" they'll protest.

Okay, let's just address the protest preemptively for Kyler then. What counts as 1, 2, 3, 4 or 5 in this game of high-N statistical dice analysis?

Kyler's premise question: what are the odds of the Book of Mormon being true? And he will start with 1 in 10^40 odds -- near impossible.

And to be sure, Kyler will use all of the "sixes" he can find, plus a few "ones" or "twos" for appearances. But here is the money question: will Kyler include ALL of the "ones" "twos" "threes" "fours" and "fives" he can find? Such as, just for starters, each of the following?

• Odds of building ships and making successful transoceanic journey with the resources at Nephi's (and Jaredites) disposal?
• Odds of fully-developed Christianity existing in the American continent before Jesus was born?
• Odds of an ancient author transcribing passages from Isaiah which had not yet been written?
• Odds the Tower of Babel was an actual historical event?
• Odds that ancient American Indians received dark skin as a curse from God, not through natural genetic inheritance?
• Odds that Joseph of old prophesized about the restoration by Joseph whose father's name was also Joseph, with no written record of that prophecy?
• Odds that Jesus visited America?
• Odds that Hebrew language and Egyptian writing were commonplace in the Americas?
• Odds of millions of armed and armored warriors were killed in battle between 400 BC and 400 AD in the Americas (and not been discovered)?
• Odds of a decapitated man doing a push up and gasping for air?
• Odds of each historical anachronism, including materials, animals, crops, New Testament verses and teachings?
• Odds of 2000 inexperienced warriors fighting in prolonged gruesome battle and not one dying?
• Odds of an ancient compass being used for navigation, before invention of the compass?
• Odds of chariots being made and used by ancient Americans?
• Odds of a temple built and fashioned after Solomon's temple being built in the Americas between 600-500 BC?
• Odds that a translator of one ancient document would fail miserably when translating a second, far shorter, ancient document?
• Odds of Ancient American population growth rates on the order described in the Book of Mormon?
• Odds of crafting a sufficiently hot bellows out of animal skins and other materials found in nature, to make molten ore from rocks?
• Odds of golden plates contained in a readily-accessible stone box surviving undiscovered for >1,400 years on the hill Cumorah?

I invite all of my colleagues to contribute additional questions to help Kyler identify the "ones" "twos" "threes" "fours" and "fives" because unfortunately with his paltry 23 evidences, he's damn sure missing a lot of them.

Kyler, are you considering the likelihood of the miracle stories in the Book of Mormon too? Because, brother, if Early Modern English gets air time in episode 9 (as I'm told will happen), then you have opened up a rabbit hole of evidence that you need to consider? Otherwise, as the saying goes, garbage-in, garbage-out.

[Doctor CamNC4Me]

Well. He’s going to have to introduce the concept of the multiverse, for sure. <- Is that anymore desperate than this or the Ghost Committee? The only possible way the Book of Mormon narrative happens is that in an alternate universe, and much in the same way the plates are taken out of this universe back to their dimensional home, everything can fall into place if the Mopologists start employing sciencebabble to explain iStones linked by subspace arrays, tempads (Loki reference) opening portals so messengers can visit prophets, and other things fantastical.

- Doc

[Gadianton]

re: Dr. Moore. Just a note that it was PG who brought up sharpshooting, I was just trying to figure out his logic (and make fun of apologetics).

As for independence, it's kind of hit and miss. For instance, here was his note about the next episode:

Next week, our skeptic will encounter Joseph’s multiple accounts of the First Vision, and we’ll estimate the probability of producing highly disparate and even contradictory accounts when telling a story years apart in different settings.

This is an example of 'evidence' that in my opinion is likely independent of book length or chapter length. However, 'evidence' of what? Neither length nor Joseph's accounts of the first vision have anything to do with whether the Book of Mormon is ancient. To the extent that Kyler is finding independent evidence for his hypothesis, it's because his evidence is actually irrelevant to his hypothesis. But that's his main chance to do the multiplication cheat. The relative strength of the Dales analysis was that they were more or less talking about things that would be evidence of the Book of Mormon being ancient. Of course, that greatly increased the chances of their variables being correlated.

what about book length and chiasmus? (pretend both are evidence that it's ancient)

At first glance, they are independent. Take the Bible. Its propensity to have chiasmus is unlikely related to the length of the Bible. If we know how many words a book has, we probably can't make any guesses about chiasmus content. So does that mean that lending the benefit of the doubt, and that each is evidence of ancient, that he can multiply them together because they are independent?

The problem, as Dr. Moore has said, is that they are likely correlated when considering the alternative hypothesis. In this case, since Joseph Smith was copying, mimicking, and plagiarizing the KJV, any number of things that make the KJV 'ancient' will transfer to the Book of Mormon. Dear god, it's like, if I paint a picture of a woman and call it the Mono Leeza, Kyler would make the Mona Lisa his target, take something by Jackson Pollack as his modern alt hypothesis, and conclude it's ancient; profoundly so.

[drumdude]

I can’t wait for the Interpreter article where an apologist with a doctorate in music theory proves the Book of Mormon is true using quantum electro dynamic theory.

[Philo Sofee]

Multiple First Visions?!? What has that to do with an ancient Book of Mormon?!? He is already lost. He can't even focus for crying out loud. So what next, how many times Joseph prayed? And then what? How many revelations did he receive for Orson Pratt? And then what? How many miles he traveled as a prophet? And then what? How many temples he built?

Just get to the damn Book of Mormon evidences, AND ***ALL*** the background of responses, the refutations, etc., showing that the high probability of them demonstrating its antiquity have been refuted over and over again, taking it right back down to the very low prior of 1 x 10^40.... Because he ain't getting away from that prior if he properly uses ALL the background (as he must if he is using Bayes), and ALL the evidence both pro and con, and all the attempts, mistranslations, guesses, and scholarly disagreement with Mormons with Mormons, AND with Mormons and scholars, AND with Mormons and Ex-Mormons, AND with Mormons and non-Mormons.
Hey, I am not over-doing it, he is the one who wishes to use Bayes, then let him use it properly and show his work...

[Dr Moore]

The problem, as Dr. Moore has said, is that they are likely correlated when considering the alternative hypothesis. In this case, since Joseph Smith was copying, mimicking, and plagiarizing the KJV, any number of things that make the KJV 'ancient' will transfer to the Book of Mormon.

Right. If he was trying to convince people that he spoke to God and had the powers he claimed, then we would expect (Bayesian at work here) something like a Biblical fan text, accounts of seeing God and angels… and to separate that correlated causal aspect from the prior probability in KR’s Bayesian is, in a word, impossible.

[Dr Moore]

Also, in my dice roll analogy above, Mopologists love to flip the dice from 5 to 6 when no one is looking. A good control would solve, but of course that defeats their purpose.