Linda: What must be repeatable? The event that was foreseen? Or the ability to foresee? Because neither of those seem to be reasonable criteria. I think it would be almost impossible to fulfill your second and third criteria without repetition, but I don't otherwise see justification for this requirement.
The ability to foresee of course.
Just think about it: Somebody dreams over years every night. Sometimes he has nightmares. There is never a relationship between his dreams and a real even. And then once one of these nightmares has some common details with a real catastrophe (which may be next day or four weeks later). Foreseeing?
There are 6 billion people in the world, and most of them are dreaming every night, and one (or five or twohundred) dream something which has some similarity to a real catastophic event.
Take the Zener experiments as an example. A candidate scores just average and the experimentators are close to give up. But suddenly they find out, that in some runs the candidate guessed the _next_ card significantly well, on other runs it is the one past next and in others the previous. Come on - just flipping coins and mining the results will give you similar results.
Linda: This is the part that makes most testimony inadequate - rarely is the available information detailed enough to make a determination. And this is one way repeatability is useful. A repeatable event can be subject to more thorough investigation (one isn't caught off guard).
This inadequate testimonies are the next problem.
Let me give an example (clearvoyance, not foreseeing):
The german (Mannheim) "medium" Artur Orlop tried to describe the village where a to him unknown Student (Geir Vilhlalmsson) lived. Orlop pruduced a sketchy drawing of the village and added some comments. His results could to about 50% be (with very good will) matched with the real village. Alone within a distance of 50 miles from my home there are two villages with more matches. So what.
Somewhere in the web there is also a detailed description of a clearvoyant describing the surface of Jupiter. It was absolutely ridiculous wrong (30000 ft high mountains beyond other stupidity), and confronted with that fact the clearvoyant claimed he might have seen another planet. ROTFL, this one brought me up for a whole day
Linda: I think this part speaks to the improbability factor - it should be highly improbable that this vision would occur in the absence of premonition.
No. In most cases a person trying to guess the next roulétte number will be wrong - and forget it. One time he will guess right - and remember it forever. Don't stumble into the very common trap of probability. Probability does _not_ tell you anything about what will happen next. No matter how improbable something is - that it happens once does tell us absolutely nothing. Not even if it happens twice or several times. if it is not repeatable statistics is out. Period.
Linda: You are simply talking about various levels of improbability. A single correct vision of the outcome of a specific lottery is so improbable as to be suggestive (but not proof) by itself. Foreseeing the next number in roulette would need to occur repeatedly to acheive the same level of improbability.
Linda: One can determine the probability related to the premonition of a unique event. The trick is to appropriately model the context - i.e. select the relevant numerator/denominator.
That is data mining. Forget it.
