Hi All,
In the interests of future astrological research discussions I'll
explain what's been happening in the various threads connected with
the "basketballers" observation and a few other observations like it.
The whole saga has been smoke + mirrors and strawman arguments
which have nothing at all to do with the original observations that I
presented. Mind games like this are a big part of usenet but they
retard astrological research if they are carried on for too long
without retraction and without finally agreeing to talk about the
same issues.
The first (irrelevant) argument was that 11 degrees of freedom should
have been used instead on 1.
It is impossible to calculate a p-score with 11 degrees of freedom
from a Chi square score that was derived from 2 factors. Such a
concept makes no sense at all in statistics. Sure, one could DO the
math, but it would be just as irrational as using your grandmother's
house number in the eqaution.
Any result obtained with flakey numbers is not actually a p-score
(hence the impossible nature of it) because a p-score has a definite
meaning in ordinary language. In my presentation of the basketballers
data the p-score means this in the English language --
"The chances of getting a graph with those high and low scores with
random data was 3 in 100". There was nothing to dispute here.
The p-score I provided, implied no reference to the regularity of the
data throughout the plotting period of 66 years - where the Virgo
score was higher all the way. A variable like that needs a different
method of calculation (for that extra rarity) and it must yield a
better p-score than 0.03. Random data simply cannot generate that
sort of "consistent looking" result as frequently as 3 in 100 tries.
So what I presented was:
(a) A p-score that has a specific meaning in English, PLUS
(b) Additional commentary which pointed to the fact that if we
.. . .consider the constancy of the progressive scores, then the
.. . .p-score is in reality much better than 0.03 but we cannot
.. . .calculate it with any known formuala. The only solution is to run
.. . .tests to find out what it is..
What ~others~ here have presented, was a series different statements
that have no bearing on what my 2 statements were.
Hermes was saying that we cannot find a result like the Virgo-Taurus
thing without looking in a lot of places, so we cannot treat the 0.03
result seriously and it must be "devalued" by using 11 degrees of
freedom to knock the p-score down to 0.9487. Richard came in
and repeated the same erroneous claim. In effect both were
suggesting that with a p-score of 0.9847 (derived from 11 degrees
of freedom) we would have a situation where the Virgo-Taurus result
could be found 984 times out of every 1000 graphs we look
at - i.e. "nearly all the time with graphs for 502 people"
Can anyone SEE how absurd that is yet - trying to knock my
observation down from 30 in 1000 to 984 in 1000? Then we had
Hermes arguing with himslf in the opposite direction by saying hat
a result like mine could be found 1 in 10 times we drew a graph with
random data -- in other words 100 in 1000 times and NOT
1 on 984 times.
Now which version is actually true -- can we expect to find a
graph like mine very frequently (984 out of 1000) or (30 out of 1000)
or (100 in 1000 times). The answer is clear - it's what I said on
day one (or 100 posts ago -- 3 in 100 times, or less tha 3 if we look
for the same consistent "shape").
Until Hemes presents his random data for *exactly* 502 people
PLUS the p-scores for each graph, the real answer will not be proved
and will remains a mystery to those who cannot already see what's
been going on here.
Ray
|
