🚀 go-pugleaf

On 8 Jun 1997 17:27:58 GMT, da...@taic.net wrote:
>	However, there are still a whole lot of categories of information
>which seem "really really relevant" after the fact, but if they are not
>spottable before the fact, are meaningless. Like the "Amir" situation. For
>even though one finds the name "Amir" next to the Rabin assassination
>phrases, if one also finds 200 other Hebrew names, then this find is
>nothing. Not useful before or after. However, for example, if one finds
>the phrase "Sus cachol" next to the Rabin Assassination phrasings, even
>though this phrase is totally meaningless and useless to someone seeing it
>(if he had a computer) 200 years ago, if it turns out that the assassin of
>Rabin fired while traveling in a blue Chevrolet, then this is an example
>of something highly significant and relevant after the fact, even though
>useless and irrelevant before the fact (no predictive value).

Now here, I think you'll find yourself in complete agreement with the
statisticians whose words I posted. _None_ of them believe that
references to Amir or murderer around Rabin have any demonstrable
relevance.

The actual stastical study, otoh, carefully laid out a hypothesis in
advance, and tested against a null hypothesis, and came up with
(honestly) phenomenal results. These _were_ demonstrated to be
statistically relevant, even though the whole business remains without
predictive value for the future.

YM