Improbability Principle

My father's technique was to watch the wheels for an hour or so before beginning play. He said that he almost always found one that was slightly out of calibration. Of course, he only played at a few casinos and never won too much, so they didn't throw him out on the street.

~~ Paul

Heh, I guess if you could actually accurately detect that it might work. Not sure it’s that easy!

Heh, I guess if you could actually accurately detect that it might work. Not sure it’s that easy!

He wrote notes as he watched. But I wonder if he just convinced himself?

~~ Paul

He wrote notes as he watched. But I wonder if he just convinced himself?

~~ Paul

Didn't a bunch of MIT students do that in the 70's or something and get in a lot of trouble for succeeding which forced a bunch of casinos to change how they handle roulette tables?

I thought a post on Skeptiko by Jim Smith was an interesting example of these issues:
http://www.skeptiko-forum.com/threads/wh...ost-122746

He linked to a YouTube video entitled "Smart Crow uses cars to crack nuts in Akita, Japan", and also to a scientific paper entitled "Crows Do Not Use Automobiles as Nutcrackers: Putting an Anecdote to the Test". The paper argued that the observations might be an incidental by-product of the crows' habit of dropping nuts on to hard surfaces, and described observations of crow behaviour in the presence and absence of cars. The conclusion was based on the absence of a statistically significant difference in behaviour, interpreted in the light of an arbitrary assumption about the size the effect would have if it existed.

I thought a post on Skeptiko by Jim Smith was an interesting example of these issues:
http://www.skeptiko-forum.com/threads/wh...ost-122746

He linked to a YouTube video entitled "Smart Crow uses cars to crack nuts in Akita, Japan", and also to a scientific paper entitled "Crows Do Not Use Automobiles as Nutcrackers: Putting an Anecdote to the Test". The paper argued that the observations might be an incidental by-product of the crows' habit of dropping nuts on to hard surfaces, and described observations of crow behaviour in the presence and absence of cars. The conclusion was based on the absence of a statistically significant different in behaviour, interpreted in the light of an arbitrary assumption about the size the effect would have if it existed.

Seems to me that if the crow is clever enough to crack a nut by dropping it on to a hard surface, it is probably clever enough to learn that cars running over the nuts gets the desired result. Taken together with the other extremely clever things crows are known to do, I don't think it is an unreasonable conclusion. Disclaimer, however: I haven't watched the video or read the paper - just an off-the-cuff observation based on your post.

Didn't a bunch of MIT students do that in the 70's or something and get in a lot of trouble for succeeding which forced a bunch of casinos to change how they handle roulette tables?

Here. https://www.forbes.com/sites/startswitha...a4c06b3c39

Seems to me that if the crow is clever enough to crack a nut by dropping it on to a hard surface, it is probably clever enough to learn that cars running over the nuts gets the desired result. Taken together with the other extremely clever things crows are known to do, I don't think it is an unreasonable conclusion. Disclaimer, however: I haven't watched the video or read the paper - just an off-the-cuff observation based on your post.

I felt it was actually quite a close parallel to the psi debate. 

Individual crows have been seen doing something that looks very much like using cars to crack nuts. But the sceptical idea is that there are so many crows in the world carrying nuts around near roads that this kind of thing will happen occasionally by chance. So some people did an experiment to test the idea, and because they didn't see statistically significant evidence in favour of the effect, they concluded that it didn't exist (at least in the title of the paper; the conclusion in the body of the paper is less dogmatic). 

But their reasoning, in turn, is open to criticism. Apart from the fundamental difficulty of proving a negative, they didn't have any way of knowing how big the size of the effect was if it existed. So in estimating the power of their experiment they relied on an assumption they admitted was arbitrary. It might just have been that the behaviour was less common than they assumed, and they didn't watch the crows for long enough for it to become evident. Or maybe the crows were thinking more in the long term - if they left a nut in the road for a while, sooner or later it would get run over. (In the YouTube video, quite a lot of cars go past before one drives over the nut.) 

It seems to me the authors of the paper reached a conclusion on insufficient evidence, just like the people who relied on anecdotal observations.

I felt it was actually quite a close parallel to the psi debate. 

Individual crows have been seen doing something that looks very much like using cars to crack nuts. But the sceptical idea is that there are so many crows in the world carrying nuts around near roads that this kind of thing will happen occasionally by chance. So some people did an experiment to test the idea, and because they didn't see statistically significant evidence in favour of the effect, they concluded that it didn't exist (at least in the title of the paper; the conclusion in the body of the paper is less dogmatic). 

But their reasoning, in turn, is open to criticism. Apart from the fundamental difficulty of proving a negative, they didn't have any way of knowing how big the size of the effect was if it existed. So in estimating the power of their experiment they relied on an assumption they admitted was arbitrary. It might just have been that the behaviour was less common than they assumed, and they didn't watch the crows for long enough for it to become evident. Or maybe the crows were thinking more in the long term - if they left a nut in the road for a while, sooner or later it would get run over. (In the YouTube video, quite a lot of cars go past before one drives over the nut.) 

It seems to me the authors of the paper reached a conclusion on insufficient evidence, just like the people who relied on anecdotal observations.

I have to admit to being ignorant of statistical calculations and how they arrive at conclusions but I have long thought that skeptics tend to use chance as a default and proclaim it decisive. In other words, if something could have happened by chance, then it must have happened by chance. I don't see why this should be the case. 

So, in your crow example, the anecdotal evidence is that some crows have been observed waiting for cars to crack their dropped nuts. How can anyone prove, by statistics, that the crow had no such intent? Even if some events were coincidental, some crows might have figured it out. As I said earlier, crows are pretty clever birds and they are known to figure out more complex problems than this one.

A similar avenue of logic is the law of big numbers: a finely tuned universe is inevitable if the number of universes is large enough. Or the mysteries of quantum mechanics go away if you adopt the many worlds interpretation. Abiogenesis again opts for chance combinations because the alternative is unthinkable. Again, even if something could have happened by chance does not prove that it did happen by chance. There is bias in the underlying assumptions.

I have to admit to being ignorant of statistical calculations and how they arrive at conclusions but I have long thought that skeptics tend to use chance as a default and proclaim it decisive. In other words, if something could have happened by chance, then it must have happened by chance. I don't see why this should be the case. 

So, in your crow example, the anecdotal evidence is that some crows have been observed waiting for cars to crack their dropped nuts. How can anyone prove, by statistics, that the crow had no such intent? Even if some events were coincidental, some crows might have figured it out. As I said earlier, crows are pretty clever birds and they are known to figure out more complex problems than this one.

A similar avenue of logic is the law of big numbers: a finely tuned universe is inevitable if the number of universes is large enough. Or the mysteries of quantum mechanics go away if you adopt the many worlds interpretation. Abiogenesis again opts for chance combinations because the alternative is unthinkable. Again, even if something could have happened by chance does not prove that it did happen by chance. There is bias in the underlying assumptions.

Concerning big numbers. http://m.wisegeek.com/how-many-chemicals-are-there.htm
What can be inferred with this knowledge?