## Tuesday, January 14, 2014

### The Oak Island Mystery Show...and Statistics

The History Channel has taken a break from Hitler and Pawn shows to have a series on the Oak Island Mystery, specifically "The Money Pit". I remember reading about this pit as a child, and much like stories of the Loch Ness Monster and alien spaceships, there haven't been any real breakthroughs in the last 40ish years.

The short discussion is the Money Pit is this mysterious pit that supposedly has treasure in the bottom of it. Nobody knows, of course, because supposedly booby traps were set, and and now it's flooded with salt water. There's a whole mythology around the pit, I won't bore any with the details...but they actually kinda-sorta found something, and had some statistics on the show.

One of the many stories about the island is a mysterious cache of coconut fiber that was supposedly on the island, years ago. The nearest coconut trees are a thousand miles away, so finding such fiber would be a sign that someone (or a really bad storm) brought coconut fiber (common on ships in the sailing vessel days, to stabilize cargoes) to the island.

Sure enough, they did some digging, and found some coconut fiber. A natural question, of course, is to date the fiber, to get an idea of when it might have been brought to the island. I used to teach carbon-14 dating in class, before admin told me it was too hard for students (a shame, since I learned it in 10th grade algebra...).

Anyway, they get it dated, and the dating lab said they were 95% confident that the fiber was from the 12th to the 14th century.

Hey, look, statistics! What they've just said here, is, using a technique that (according to theory) works 95% of the time, they're confident the fiber dates between 1200 and 1400 AD (that means a 100 year margin of error, and the sample mean likely dates to 1300 AD). That's a "confidence interval."

Now, the PROBABILITY that the true age of the coconut fiber is between 1200 and 1400 AD? That probability is 0 or 1...we'll never know. But the technique works 95% of the time, and that's good enough. Even if the technique didn't work, it's not outrageous to suppose the fiber is from 1199 or 1401 or something. Anyway, that's why they call it "confidence" as opposed to probability.

And it's kinda neat they finally found a tiny little bit of something that indicates that maybe that hole is quite old, indeed...at least if you want to make the leap to think that the hole is related to the coconut fiber.

1. There are some subtleties at play here. The true age is some number -- and not a random variable. So we can't talk about the probability that it's from between 1200 and 1400 AD. And if we construct a specific 95% confidence interval either the true age is somewhere within the interval -- or it's not. All we can say is that *were* we to construct an arbitrary 95% confidence interval, the chances are 95% that the true age will be somewhere in this interval.

2. Exactly, once the interval is created, probability is no longer involved (i.e., the probability is either 0 or 1).

All the % does is indicate the level of confidence, based on what is known. Of course, if the assumption of "what is known" is false, then it all becomes meaningless. This could be a real problem in carbon dating, since there's evidence now that radioactive decay isn't a constant (an assumption made for dating).

1. Since when was radioactive decay considered constant? At the atomic level, it is a sort of a stochastic process. All we can say is that, over time, every such atom will undergo the process to the limit of our detection capability; but we cannot predict when that process will occur for any given atom. At the bulk level, the half-time is approximately an intensive property. As the 'bulk' decays, the tail will naturally show more variability, more so if the decay products also undergo radioactive decay that confound the measurement. Add to that issue, in terms of dating, what other processes that could add or remove the radioactive element we are analyzing from the sample, either at random one-time events, Poisson events, or periodic events.

2. Radioactive decay is presented as a constant in schools, and even wiki (har, I know) still refers to it as a constant.

3. A key evidence as proof to who may have built the Oak Island Money Pit lies with the use of the Coconut Fiber found within the Main Shaft and Flood Tunnels
My theory for “The (Untold) Story of the Oak Island Money Pit” does not support the carbon dating of the Coconut Fiber to be between 1260 and 1400, with 1330 being between those dates
This time frame has led to many latter dated theories of Vikings, Knight Templars, and Druids with all of them using the Coconut Fiber during this period for their construction.
Investigating into the history of the Coconut Tree shows that although it originated in Malaysia during the latter 1200ad, it did not cross over to Africa, Europe, South and Central America until mid-1550 with the Coconut Fiber not available for use until the latter 1600’s.