New episode out tonight. DB Cooper’s safety deposit box with Tim Collins aka SafecrackingPLF.
You are not allowed to view links.
Register or Login
Enjoy!
I’ve got some other great guests coming up, but I’m still waiting on a few of you guys to come on. If you’re on the fence I’d say maybe ask some one that’s been on, how their experience was. This show will last a pretty long time, but I’m not going to be doing it much longer. I’m talking to you Georger, Flyjack, Parrothead, shutter, and snowmman.
3 hours is a lot to listen to. Anyone know when he talks about Bayes or if he talks about Max Gunther? Here are some comments on the parts I've listened to, and response to some of the posts on DZ.
Tim seems like a reasonable person. I like the approach of narrowing the field so to say, but I question some of the math. He starts off taking a total US population of 225 million, then the next step is taking into account male smokers, then somewhere in there is wavy hair, pocket knife ownership, etc. I don't think you can multiply the percentage of male smokers by the total population. You need to multiply the percentage of male smokers by the total male population.
The flaw in this is the overlap of what I would call conditions. There is double counting going on. If I take all the pocket knife owners, with wavy hair, with swarthy complexions, who smoked, then I might have more than the total population. It would be like me adding together all the people in the US with high blood pressure and then adding in all the people who are overweight and who don't exercise. These symptoms overlap, in that someone who has high blood pressure is also likely to not exercise, so I can only really count them once. I learned this in detail when looking at populations to recruit from for military service. We found that much of American youth had multiple disqualifying conditions, and that 18 year olds with criminal records often had poor grades and tested poorly on aptitude tests. So even if we got them to prepare for the aptitude test and do well, it did not change the fact that they had other issues. Comorbid would be the medical term.
I do believe in narrowing the field, but would look at it a little differently. Something in this order:
1. Take the total male population in the US around 1970. Make the assumption he was American. Add in Canada if you want. He was not female.
2. Take his age. In 1970 there were about 25M males between the age of 35 and 60.
3. Take out those who did not serve in the military (again an assumption)
4. Take out East Asian, African American, and any other groups that he certainly was not a member of
5. Now try to figure in smokers. Take out PhDs, maybe take out Masters level. Take out eye colors if you'd like.
It gets tricky when you start looking at overlaps, like brown eyes who served in the military who are of the right age. But, you can definitely whittle down the total number. But you can't get it down to 19 people like the podcast guest did. Duane Weber is one of thousands of American males, maybe 10s of thousands that fit Cooper's description. How does this help right now? It probably does not. But in the future with enough databases, the field can be narrowed. Ancestry.com can do a lot simply by using the census data. How many males, aged 35-60, of European descent, who lived in a certain area, who served in the military, etc. It can be done. And that then gives you the list of who's DNA to look at. Possible, but not probable given our privacy laws.
Georger: Good points on DZ about probability. Simple probability works. How many times did a plane flying South on that route make the turn to the West? To use probability we need to know what distribution to use, as in uniform, normal, Poisson, etc. It is a complicated field, but it is often used to overcomplicate things. The parachute probably opened. It probably got him to the ground. But if we want to use Bayes and figure out what the probability of the chute getting him to the ground given the probability that it opened, then great.
I'm all for using math like in the podcast, but there needs to be a balance with common sense. There are a lot of basic facts in this case that don't need to be made more complex. I have graduate level classes in probability for engineers, and I can say that it is a very complex field. Maybe not to the level of physics, but complex. So it is easy for people to throw out a bunch of concepts that confuse or make people disengage. I just don't see the need for deep complex math here. Maybe some algebra, but not calculus so to say. Well maybe a little calculus.
Last thing. The podcast mentions left hand smoker. Were there nicotine stains on both hands or just one, and if one, then which hand?