Re: psi

Damien Broderick (
Fri, 08 May 1998 01:46:15 -0700

At 11:29 PM 5/6/98 -0700, Pelagius wrote:

>I certainly don't think more experiments are needed in precognition because
>every week a huge experiment on it is conducted under the most rigorous
>controls imaginable. Every week in the USA millions of people buy lottery
>tickets and have been doing so for many years. On every Saturday 6 numbers
>are picked at random, each number can be 1 to 50, if you guess all 6 numbers
>you win millions of dollars, if you guess 3 you win about $10. If there were
>anything to ESP the number of winners should be larger that what you would
>expect from chance, but it is not, and if the effect is too small for even a
>huge experiment like the lottery to detect then The Princeton Engineering
>Anomalies Research sure as hell won't find it.

People keep handing me these terrific straight lines... :)

As it happens, I published a book in 1992, full of otherwise unobtainable
data aand terrific charts, called THE LOTTO EFFECT (Melbourne, Australia:
Hudson, ISBN 0-949873-3-41-1), which is based on exactly this notion. I
examined some three-quarters of a billion guesses in 20 consecutive draws
of the national Australian Lotto game. First I normalised the guesses at
each of the numbers in the range 1-45, because populations hold different
preferences and biases toward each of these options - 7 and 13 always get a
markedly enhanced vote, e.g., whereas all the `non-birthday' numbers over
31 get a reduced vote. Then I did a linear regression to iron out certain
long term irrelevant trends. Then I compared the normalised votes for each
candidate number *when it was a target* against *when it wasn't*.

If psi precognition were operating, one would expect a significantly
elevated (or perhaps a significantly depressed) vote for the six numbers
popping out of the randomiser. I didn't find that overall, to my chagrin -
but I did locate a very weird effect in which the *most drastically deviant
positive residuals correlated with targets to a far greater extent than one
would expect by chance*. Opinions differ over whether this post hoc
finding supports the psi hypothesis or is an artefact. I think that when
the suggested artefact is factored in, there is still evidence for
precognition in the data.

(BTW, while it is true that `If there were anything to ESP the number of
winners should be larger that what you would expect from chance', the
expected impact isn't as obvious as it seems. Leaving aside a truly
immense amount of stereotyped entry strategies - the same birthdays every
week, the same `lucky numbers', etc - I assume that each guess is
independent, and therefore each perfect entry requires 6 `psi-mediated
events' (other than those which get some numbers by chance, as some must).
There's a kind of pyramid generated, in which almost all the fairly
infrequent psi events you'd expect in such a population will be wasted on
people who would otherwise by chance get only zero, one, two, or three
right. This still leaves some opportunity at the top, but it wouldn't be
as dramatic as one might intuitively suppose. I discuss all this in
monstrous detail in my book.)

Why isn't this result widely known among New Age fruitcakes and skeptics
alike? Beats me. Dr Dean Radin does reference my book in his THE

Why haven't I just done a simple replication run or two and settled the
matter? Because it took me about a decade to gouge the goddam data out of
the lottery company, and they are not interested in giving me (or several
interested university statisticians) any more. Sigh. But one of these
days, when someone gets the Nobel Prize for this, remember that you saw it
here first...

(And yes, I know, Mike Lorrey and I should get together and create a
psi-powered Dean [Radin] Machine, ho ho bloody ho.)

Damien Broderick