* From: "Ray Murphy" <ray...@[EMAIL PROTECTED]
>
* Newsgroups: alt.astrology.moderated
* Subject: Virgos play basketball more than Taureans
* Date: Sat, 19 Apr 2008 00:53:33 -0500
* Message-ID: <480982f0$1_4@[EMAIL PROTECTED]
>
*
* Hi All,
*
* If people who are born under the sign of Taurus and Virgo have equal
* access to the s****t of Basketball, then those with Sun in Virgo elect
* to play that s****t at a much higher rate than the Sun in Taurus
* people.
*
* See the attached progressive-graph in date order for the 502 people
* listed in Astrodatabank Version 3 who were born in either of those
* signs. There are 285 Virgo and 217 Taurus people born during a 66 year
* period from 1913 to 1979.
*
* The observation is statistically significant at about the 0.03 level
* when a comparison is made between those 2 signs in isolation.
*
* Noon times have been used for all births.
*
* [Graph has been posted in the Files section of the Google Group
* "Tropical astrology Research"]
*
http://tropical-astrology-research.googlegroups.com/web/Basketballers.gif?gda=U2MV2kUAAAAR4o_U4Mb1ar5u1N08e_lzoV3BSWdITZ_Q3wqNGgyUGOG2sVwGLF-i21hNBTL4y6OkNTLimnrPyTN8pl8xhyRsrnbirdOKW0UubKy7Gu2GBQ&gsc=1rI-yAsAAABS4Ee1kLUl647TiZkjb1-K
*
* Ray
For all it appears, this post claims to present a significant indication
that Virgos become more often basketballers than Taureans.
As usual in RMs posts, things are not 100% specific. The first sentence
of the post, for example, reminds me personally a lot of "if the moon is
made of green cheese, I can fly around the moon". In my experience and
personal perception, RM usually replies only to part of what someone else
wrote in each sentence, and very often interpreting that part differently
from what apparently the original poster intended to say. Moreover, in
my personal experience, also RMs replies are kept ambivalent by either
leaving away some details or making statements that are contradictory or
undecided from a purely logical point of view. That usually leaves the
other side, in my view and personal perception, facing a messy heap of
things that appear to be necessary to clarify. With each attempt to do so,
however, things usually tend to grow worse by more subsequent replies by
RM, almost exponentially so. But that is factually just RMs style, and
I leave it to the discretion of the reader how to handle this or not... :)
That made it somewhat more difficult to isolate the essential facts in the
case of the basketballer graph, but for all it seems, I got them from RMs
replies in the end.
So, in the following I will focus on that, even though more philosophical
questions like what is "random", how does it relate to "predictability",
"complexity", "white noise", "meaning of a message" would be interesting,
too, but were maybe more interesting with Pluto in Sagittarius. Instead I
would like in the following to focus on whether the presented graph, plus
how it was obtained, represent a scientifically significant result, a
formal scientific proof or maybe just an indication, a "signpost", that
Virgos would in fact in general be more likely to become basketballers
than Taureans. In other words, would one expect if the study was extended
to everybody on earth to find significantly more Virgos than Taureans who
become baskeball players? Or is this just an expected peak in noise, i.e.
almost certainly no reason to focus on this instead of on other things?
The graph looks as follows (in case the graph should no longer be online
some years from now and somebody should read this):
- rectangle
- left scale: 0 - 300 (labeled 0, 50, 100, 150, 200, 250, 300)
- bottom scale: 0 - 502 (labeled 0, 251, 502)
- top scale: 1913 - 1978 (labeled 0, 1951, 1978)
- numbers on each scale a equal relative distance within each scale
- two curves
- a blue one from bottom left (0,0) to top right (502,217)
- labeled "TAURUS (217)"
- a violet one from bottom left (0,0) to top right (502,285)
- labeled "VIRGO (285)"
- the two curves are relatively close to being two straight lines
- text at top left corner: "Basketballers in Astrodatabank, Virgo
consistently more frequent than Taurus for 66 years of birth"
Now let me gather some more facts:
* From: Ray Murphy <ray...@[EMAIL PROTECTED]
>
* Newsgroups: alt.astrology.moderated
* Subject: Re: Virgos play basketball more than Taureans
* Date: Sat, 26 Apr 2008 17:23:07 -0500
* Message-ID:
<143c64a4-6a43-4a8b-9b90-3b842096926d@[EMAIL PROTECTED]
>
[...]
* RM: The horizontal axis represents time in 2 ways simultaneously --
* firstly as the accumulating sum of birth counts [in sequence as they
* occurred] and secondly as time [divided into 2 halves according to the
* scores]. For example in that graph 251 TAU + VIR people had been born
* in the first 33 years up until 1946 and another 251 were born in the
* 33 years from 1946 until 1979. In this case there was 33 years in each
* half but in other graphs for different projects there are more years
* in one half than in the other.
[...]
* RM: For the same time span (1913 to 1979) there were 2409
* basketballers in the database for the other 10 signs.
*
* 239 -- ARI
* 217 -- TAU
* 224 - GEM
* 277 - CAN
* 268 - LEO
* 285 - VIR
* 242 - LIB
* 230 - SCO
* 228 - SAG
* 219 - CAP
* 246 - AQU
* 236 - PIS
* ----------------
* 2911 - SUM
* ----------------
[...]
* These were my steps:
*
* (1) I had a database of 41,249 people in 100 categories.
* (2) Filtered for all Virgo's and found 3347
* (3) Printed out all the scores and %'s and X^2 etc for each category.
* (4) Looked for the highest Chi squared result (Basketballers was 10.03)
* (5) Looked at the basketballers' Sun signs in isolation in Jigsaw
* ... to find the lowest scoring Sun sign, which was Taurus.
* (6) Isolated the TAU + VIR basketballers in a new file.
* (7) Sorted them into birth order.
* (8) Made a "progressive graph" which showed the scores for
* ... both TAU + VIR as each person was born. The reason for producing
* ... that style of graph was to ascertain if there was any consistency
during
* ... the 66 year period or whether it was just a wild fluctuation in
random data.
and
* From: "Ray Murphy" <ray...@[EMAIL PROTECTED]
>
* Newsgroups: alt.astrology.moderated
* Subject: Re: Virgos play basketball more than Taureans
* Date: Mon, 28 Apr 2008 11:24:38 -0500
* Message-ID: <4815f567@[EMAIL PROTECTED]
>
[...]
* [Correction]
* In my reply I said that half of the data was from 1913 to 1946 but it
* was 1951.
and
* From: "Ray Murphy" <ray...@[EMAIL PROTECTED]
>
* Newsgroups: alt.astrology.moderated
* Subject: Re: Virgos play basketball more than Taureans
* Date: Mon, 28 Apr 2008 10:30:22 -0500
* Message-ID: <4815e6bd@[EMAIL PROTECTED]
>
[...]
* RM: That's effectively what the progressive graphs are showing - that
* there were more Virgo basketballers than Taurus basketballers born in
* time blocks for the whole 66 year period. The reason why I introduced
* the progressive graphs is to smooth out the fluctuations that
* naturally occur with tiny amounts of data.
and finally
* From: "Ray Murphy" <ray...@[EMAIL PROTECTED]
>
* Newsgroups: alt.astrology.moderated
* Subject: Astrology Statistics for Dummies
* Date: Wed, 30 Apr 2008 19:08:59 -0500
* Message-ID: <4818f905$1@[EMAIL PROTECTED]
>
[...]
* Astrology Statistics for dummies:
* [Finding Chi squared with only 2 bars in a graph]
*
* (abbreviations commonly used are X^2 -- or -- Chi2)
*
* Example below is from the Basketballers thread:
*
* Virgo score . = 285
* Taurus score = 217
* ----------------------
* SUM .......... = 502
* ----------------------
* Expected score for Virgo = 251
* Expected score for Taurus = 251
*
* Formula:
* X^2 = (Obs score - Exp score) squared, divided by Expected score
*
* X^2 = (285 - 251) = 34 squared = 1156 divided by 251 = 4.6056
*
* To find the p-score (probability of this occurring by chance) go to
* this online calculator and enter the Chi squared value of 4.6056 with
* ONE degree of freedom. Answer is p = 0.0319
* http://www.anu.edu.au/nceph/surfstat/surfstat-home/tables/chi.php
*
* To convert the p-score into intelligible probability numbers for
* astrologers we divide it into 1.
* 1 divided by 0.0319 = 33.3480
*
* This means that if we repeat the test with 33 more sets of 502
* VIR + TAU basketballers we might get the same good p-score
* again just by chance.
Phew!
OK, first a word about the double horizontal scales. For all that it
appears, the lower horizontal scale is the one that really counts,
i.e. it is sum of counts of Virgo plus Taurus basketballers in the
order in which they were born. The years indicated on top are just
indications of the corresponding birth years for 3 specific values
on the bottom scale. (This is also obvious from the fact that 1951
is not (1979+1913)/2; 1946 would be arithmetically in the middle.)
Now, that way of drawing data is already a bit unusual because it
mixes the scales. The vertical scale, counts of Virgo resp. Taurus
basketballers versus sum of both in the horizontal scale. This has
the overall effect of smoothing out some fluctuations. It is also
not unexpected that the two curves are close to straight lines, very
simply because straight lines are more probable than ones that do
get closer and get away from each other or would even cross. Note
that because the horizontal scale is the sum of the counts for the
two curves, each step of one curve is mirrored by a corresponding
step of the lower curve.
So, overall, the fact that one sees two straight lines is not an
unusual or unexpected thing. Remains the question of whether the
difference at the end is in any way significant or an indication
of that sort.
Apparently the graph was obtained as follows:
1) Choice to look at Virgos.
2) Selection of basketballers from 100 different categories. Not 100%
sure how exactly the Virgos were selected, not sure what exactly
point (3) in RMs reply above means. But it is sure that there was
a selection process used that aimed to create a maximal difference
between counts of Virgo and another sign. (In more detail: The way
that RM calculated a Chi2 value in the end was (diff)^2/(sum/2),
which aims to maximize the difference relative to the sum, which
is for each category separately roughly equivalent to maximizing
the difference.)
Conclusion: A factor of about 100 must be taken into account when
estimating the significance of the result.
3) The sign with the count with the largest difference was selected.
Again, more precisely, the Chi2 value in the way RM calculated it
was maximized, i.e. (285-count)2/(sum/2). Since in this case, the
count of Virgos is above the expected mean, 2911/12 = 242.6, the
chances of a value lower than the average to give a maximal Chi2
value are bigger than for one bigger than the average (Note that
a count for a second sign means a smaller sum which means a larger
Chi2 value).
Conclusion: A factor of 11 must be taken into account when
estimating the significance of the result.
The selection of one of 11 signs already makes the probability of
getting a difference of 0 extremely small (counts for all 12 signs
would have to be exactly equal). A computer simulation shows that a
difference of about 30 is roughly what has the maximal chance of
happening. Smaller and larger differences are less likely. As the
simulation also shows, the chance of finding a difference of as much
as the count in the basketballers graph, 284-217 = 68, is close to
0.02 = 2% = 2 in 100.
Since this graph was effectively selected from 100 possible ones,
and selected essentially to give maximal difference, it is not at
all astoni****ng to see this graph and thus provides no scientific
proof, not even the slightest indication, that Virgos might in fact
in general become basketballers more often than Taureans.
In terms of divination, the whole thing is certainly interesting,
but one should not overlook that in terms of divination, almost
every graph is likely to show something that makes it special, that
fits with the signs and the category involved. One graph might show
actually lines that do cross - also something not very likely to
happen, but maybe likely to encounter in a set of 100 categories and
12 signs. It is thus simply and conclusively unfair to ignore the
others and just to focus on the difference of counts at the end.
I guess I have no choice but to leave this post archived to keep
the message in one piece... :)
PS: Note Sedna transit across my natal north node and now joined by
transiting sun (and of course, Mars has entered Leo). There is maybe
a bit too much attention on Eris and Pluto these days, while Sedna
is still the biggest object in its category and realm. In that sense,
current times, like the ones after discoveries of Uranus, Neptune,
Pluto (and Lilith, now in early Sag, trine Mars), and maybe Cheiron,
might be more influenced by Sedna than by Eris and Pluto.
)o+
--
http://www.exactphilosophy.net/
Something different...
|
