Overall Probability of the Proposition that Earhart Landed on Nikumaroro

[Jon Romig]

I hope this offering is not unwelcome or untimely, but I have an enduring interest in determining the combined/overall probability of our hypothesis.

TIGHAR has identified MANY items of evidence, none of which is a "smoking gun" that, when all items are combined, result in a near certainty. Rick has stated this many times in interviews and writings. Although it may be foolish to try to put actual numbers on so many items for which probabilities are difficult to ascertain, it seems to me that, using wide ranges of probabilities for each item, and then combining many items, we could perhaps get closer to certainty and reinforce Rick's claim. The principle is that many items - even that are of very low probability and nominally useless - when combined result in something of real value, and where the initial high individual uncertainties average or cancel each other out (I am no statistician but I believe that is correct).

Attached is a sample (Excel file) of how we might approach putting actual numbers to the combined probability for the hypothesis, based upon all of the evidence that has been amassed by TIGHAR. If it has not already been done, this effort might also provide a record in one place of all the items of evidence (I have wanted this list for a long time). Finally, this worksheet could be a living document that can be updated as new information or analysis is acquired.

Please let me know if you think this is doable and worthwhile.


Thanks for your interest!

Jon

[Martin X. Moleski, SJ]

I am sympathetic with your desire for quantification.

That is one aspect of the hard sciences (physics, chemistry, biology) that has driven great advances.

For like-minded people, the numbers may be consoling.

My problem is that every number entered in the spreadsheet is a judgment call, as is the decision to take some items as originating from the aircraft or crew in the first place. The numbers come from a prejudgment that the item is evidential: the judgment does not come from the numbers.

I worked very hard on tracking every sextant number that we could find on the internet, thinking that the odds against the sextant box not being related to all of the other artifacts at the Seven site were astronomical. I was wrong. The numbers on the sextant box eventually did tell a tale, but it was not the Niku narrative.

I do salute your effort to make a list. That in itself may prove helpful in grasping how much TIGHAR has discovered. I have held the original bones file in my own hands in Fiji, and I saw a tiny sample of Ric's mapping of low tides at Niku against the timing of the transmissions that were probably from AE and FN. Of course, I went to New Zealand and to TIGHAR headquarters because I was already persuaded that the hypothesis is true not on the basis of assigning numbers to elements of the case but by a personal judgment that the many strands of TIGHAR's case do add up, in a non-mathematical sense, to create a rope or a cable that is stronger than any of its components that are weak and inconclusive in isolation.

[Bill Mahoskey Jr]

As a long time follower of this topic, and a fan of those who share a similiar passion, I commend you for re-igniting the discussion in the forums.
I have a unique connection to the topic. In 1997 I worked at a private aircraft hangar (FBO) at Honolulu Internation Airport (Circle Rainbow Aviation). Linda Finch recreateed Amelia's flight and stored her aircraft (a lockheed electra E) at our facility for a few days.
I pointed out a small vibration related crack in one of her engine cowlings. She let me drill a 'drill start' at the terminus of the crack (it was a fairly short crack).
As a former airframer on aircraft, I was pretty pleased with myself lol.

[Arthur Rypinski]

Mr. Romig is adopting a Bayesian approach to the problem of Earhart's final resting place.  We have a bunch of artifacts that might (or might not) originate with Amelia Earhart.

One can assign a subjective probability to the likelihood a particular artifact being from Earhart and/or her plane.  That probability, as Marty points out, is an opinion. There are a range of views on what that probability might be. With each new fact we uncover about each item, one's opinion is likely to change.

However, Mr. Romig's essential insight is that Bayes' Theorem shows that if it is one's opinion that there are several artifacts that each have a non-zero (subjective) probability of originating with Amelia Earhart, the (subjective) probability that AE landed on Niku is greater than the probability of any particular artifact originating with her, and that more plausible artifacts raise the overall probability of the Niku Hypothesis being correct.

The Niku Hypothesis does not stand or fall on the provenance of any particular artifact, and conversely, if any one artifact originates with Earhart with certainty, then the Niku Hypothesis is true.

Bayes analysis has several uses.  If one constructs a decision tree, it helps define which uncertainties have the biggest impact on the probability of the Niku Hypothesis being correct, which can be useful for resource allocation. It can also illustrate the impact of new information on overall (subjective) probability of the truth of the Niku Hypothesis.

[Martin X. Moleski, SJ]

However, Mr. Romig's essential insight is that Bayes' Theorem shows that if it is one's opinion that there are several artifacts that each have a non-zero (subjective) probability of originating with Amelia Earhart, the (subjective) probability that AE landed on Niku is greater than the probability of any particular artifact originating with her, and that more plausible artifacts raise the overall probability of the Niku Hypothesis being correct.

Thanks so much for the clarification of how people can use Bayes' Theorem and decision trees to guide their research.

I was persuaded by the Niku Hypothesis on my first reading of the TIGHAR website in 2000. Without doing any diagrams of what convinced me, I just had an overall feeling that the hypothesis made good sense. In the 25 years since then, there have been many changes in the data available to support or discredit the hypothesis. 

I understand that this kind of circumstantial case, in which every piece of alleged evidence is dubitable, is very weak. The fact that it is a weak case does not mean that it is false. In 2009 I had high hope that underwater searches would turn up the Any Idiot Artifact, but now I understand how misplaced that expectation was. The search area under and around Niku is not like that around the Titanic. If there are any pieces left, they are almost certainly lost to us, given the state of the art in saltwater metal detecting.

Life is like that.

The navigation logic and the pattern of the radio transmissions give me a great deal of confidence. Neither of these had been studied in detail in 2000. I put a lot of trust in these to lines of argument, even though they amount only to probabilities, not certitudes.