Paranormal ... (but I bet you knew that before I typed it!!)

Myksyk

An interesting discussion was beginning between rde and meatproduct (on the 'New Scientist' thread), on psi and other alleged paranormal abilities. It was suggested that a new thread was needed as it was somewhat off topic where it began. So here we are.

Mention was made of "sufficient number of controlled experiments" having been carried out to demonstrate that psi existed. Just wondering what people's opinion of this claim is.

BTW, this thread is about supposed paranormal abilities in general but psi is as good a place to start as anywhere.

watty

I guess they got the Idea for the "Jonathan Creek" series because "stage magicians" have been 100% succesfull at debunking supposed succesful experiments. I saw a very convincing documentary of a Russian ESP experiment that was positive.

The stage magician came in and in hours the results had evaporated. A slight error of technique and they unwittingly "fooled" themselves.

What was interesting was that there was no suggestion of delibrate colusion or fraud.

Longer ago "Houdini" was famous for exposing "Mediums" as frauds.

These "stage magicians" become expert at "misdirection" and are total cynics in reality that do entertainment that depends on the innate human capcity for guilability.

Houdini had quite admired Sir Arthur Conan Doyle (Sherlock Holms etc), but was upset when he discovered how guillable his "hero" was about the paranormal.

In fact no properly controlled experiment has produced conclusive results, though frequently you get claims there are.

Also rolling a dice and getting sixs more than 1/6th of the time can mean two things NOT connected with PSI powers:

1) The dice is not perfect

2) You haven't done it long enough!

If you study two branches of Math *beyond* ordinary schoolbook Probability and Statistics, you won't try to have a system for the Lottery or Roulette*

a) Stochastic Theory
b) Random walks

I'll not try exactly to explain Stochastic Theory. I wrecked my head with it designing a Pan European Paging System.... But it is the relationship between resources, queues, probability and Statistics so you can build a model of a system and predict its loading or traffic. Believe it or not, that *DOES* apply to interpretation of ESP experiment results.


Random walks are how long the results can go on deviating from the probable mean line and how far.

As experiement you log the data from a "test" with known probability of each result. You then can plot the random walk. Then you get very clever and have lots of math to explain this.

Suffice to say that with two outcomes the random walk length mostly is short and deviation is fixed. With a dice it is sometimes much longer. With the 37 position Roulette wheel you *can* get unbelievable deviation and length of walk. Sometimes.


What this means for ESP testing, is that the results have to be repeatable on multiple occasions and a higher margin that you would intuitively think, or else in fact you are looking at "one of" the possible "random walks". Like a "lucky night" at the wheel, Losing and winning streaks are perfectly mathematically predictable size of "random walks".

You can't use the math to beat the system as you never know "which" random walk (losing or winning streak) you are on.


*Roulette in a Casino *HAS* been beaten, But not by a system the way you would imagine. They had a computer connected to toe operated switches (hidden in shoe) and the player keyed in a guess as to spin speed and where ball first hit, etc. The computer sent back a predicted number. It wasn't perfect, but enough to beat the odds. Of course the management of a casino know the difference between a "winning streak" and something else. Eventually they got caught.

Similarly people that can memorise exact sequences of cards etc their identity is circularised and they are refused entry... There are some Humans as good as Data. See the sig below.

watty

I'm not saying ESP doesn't exist. But proving it is a little more complex than holding up the patterned cards.

Your results might not mean what you think they mean.