Background

I wanted to follow up my simulation article with one on statistics because the two are intimately connected by one key common factor: the fact that the universe we experience may not be random and everything that presupposes that it is, is built on sand.   

I have written before about the problem of astrology and statistics. But I want to return to it here because the message cannot be repeated often enough.

The majority of scientists will say that astrology is a load of rubbish and that they know this as it can’t be proven scientifically- or rather by statistical experimentation. Though, curiously, although they supposedly have the tools, they have no interest in seriously trying. (Though silly, non -scientific tests, are often loved by them)

In social science it is little better. Psychologists, often prepared to work outside pure scientifically proven boundaries, will sometimes utilise astrology but with reservations. Carl Jung, for example, who incorporated a certain amount of astrology into his work, bemoaned the lack of statistical studies in his letter to Professor Raman[i].

“What I miss in astrological literature is chiefly the statistical method by which certain fundamental facts could be scientifically established”.

Even some of the more dogmatic astrologers will also bemoan the fact that there should be more focus by astrologers and by regional astrological bodies in conducting statistical study.

However, when studies are undertaken, such as those by Gauquelin, who was originally a sceptic, but found evidence for some effects (e.g.sports people having angular Mars etc.,) people cannot wait to find every reason to tear apart their findings. I particularly like the following quote[ii] :

“Baffling planetary puzzles
Gauquelin's work simultaneously discredited astrology and replaced it with weak planetary effects of no practical use…… planetary effects attracted a huge controversy about whether they were real and (if real) what caused them…. But planetary effects create far more intriguing puzzles than this, and….. are just as puzzling for astrology as they are for science.”

It is hilarious because it basically says they found planetary correlations but they can safely be completely ignored because they weren’t the ones already written about. So Science can only recognise something if there is sufficient documentation about it before it is discovered? I doubt that. Anyway the statement can only be made someone who understands a limited amount of astrology as in fact they are written about in literature but are only a small part of a much bigger picture.

Actually, the average astrologer will tell you that the Gauquelin research while interesting and while indicating that there are some planetary correlations with elements of peoples’ lives, is limited and further statistical analysis all but impossible because astrology is difficult to demonstrate because of its complexity.

The extent of the problem

That is a reasonable argument. Let’s look at the basics. Assuming only signs and houses count (i.e. assuming aspects don’t and degrees don’t- so 25 degrees Capricorn is the same as 5 degrees Capricorn), we can look at the potential combinations for a random person:

Say person A is picks a person B but we know nothing about them, even age or DOB or what they look like. Let’s restrict it to someone who could be living, say aged 0 to 100

Our job is to guess the position of planets in their chart; sign and house.

We can start with the sun and moon: We can have the following permutations.

12 Sun signs x 12 houses x 12 moon signs x12 houses. Approx. 20,000

Now add Mercury *– it can be only 1 sign away (either side) from sun but that adds x3x3

We are at 186,000

Now add Venus * - it can be two signs (either side) away x5x5

Now we are at 4.6million.

Mars* moves on average 6 signs a year so x6x6

168million

Now let’s add Jupiter, it travels a sign a year, Saturn, takes 2 ½ years, In 100 years we can most of the combinations-being conservative let’s say 80, Adding Uranus and Neptune takes us up to more than 100 and that is ignoring house positions.

Ok so now we have 16billion combinations.

*- these are exaggerations as in any given year fewer permutations are possible – but over 100 years and with the outer planets only partially added it is a reasonable assessment.

You’ll be a while trying to get the right combination for the unknown person. Add close aspects to this and frankly you’ll be there guessing for the next 100 years. And I haven’t even bothered with degrees or other bodies such as fixed stars.

So there are clearly a lot of permutations, each one slightly different from the next.

What this means for statistical analysis is that it isn’t going to be easy to extract a single factor (e.g. ,Mars) from someone born on 1 July 1970 and compare with someone born on 15 April 1995 and come to a conclusion how that manifests in the person’s life without reference to any of the other factors . So, yes, those astrologers who say this makes statistics hard to apply to astrology would be right.

But separating factors from the birth-chart is like separating bits of DNA. With DNA there are certain syndromes or illnesses which require just one genetic mutation. These can now be isolated and identified from the millions of lines of genetic information. However, there are still plenty that can only be partially explained; cancers that have a genetic and lifestyle component, musical ability, food intolerances, intelligence. Get a report for yourself and you will find that the genetic picture presented for you will be qualified by the fact that only 50% for example of a trait can be explained by known genes.  A bit liek astrological characteristics then!

Yet, with the benefit of huge computing capacities, and more and more data, the science moves on day by day. One day they’ll get there.

An alternative comparison is that astrology is like playing chess, with astrologers as the players. And chess has pretty much been mastered now by computing. Not entirely, but enough that one can make predictions about almost all outcomes from it. Computing power wins again.

In astrology, we can now get astrological reports that are computer generated that will tell you all about the individual combinations in your chart- though they still don’t attempt to synthesise all the full combinations.

The advances in computing power mean it is no longer theoretically impossible to program all the possibilities in, even if practically no one has yet written all the interpretations.  So could it be that we have reached the point when astrology can or cannot be analysed by statistics?

Perhaps we have. Perhaps in a few years someone will try.

However, we are still focusing on the “what in astrology can be proven by statistics” question. And I contend that this is the wrong question.

The real question

The real question is "if astrology is works, can we actually rely on statistics?"

I don’t mean, “can we rely on statistics to verify astrology?”. I mean “can we rely on statistics at all?”

Here I am specifically talking about the use of random sampling and probability.  

My scepticism does not lie with the mathematics of statistical theory, which, rather like the mathematics of classical economics, appears to be complete and correct within itself. No, my problem is in its application to practical everyday problems.

Astrology is all about time.  In astrology every moment has very specific characteristics associated with it as described by the planets etc.  Aside from birth-charts, astrologers use this in the practice of horary where the answer to a question is provided by the conditions at the time the question is asked, and in electional astrology when a date and time is picked for a specific action.

The key is that astrology says that every single moment is unique.

Statistics relies on the opposite: very moment being the same. Statistics says if you pick a random sample from a given population at 3pm on Thursday and pick another one at 10am on Friday, those samples will be equally representative of that population. Statistics says keep picking random samples and you will eventually fully reflect the population.

Astrology says: Naaah. Those samples will be representative of the moment you picked them.  Even if you keep picking samples, there will always be a time bias to them.

THERE IS NO SUCH THING AS RANDOM.

In practice people don’t keep picking samples anyway, they pick one samples and use know mathematical confidence limits to explain to what extent that describes the population.

But it doesn’t even stop there, because someone picks the sample (even if all that involves is pressing a random number generator). And that someone has as birth-chart, so what you get is a sample that incorporates the time of picking and that birth-chart. If the birth-chart says the person will fail then (they will pick a bad time and) chances are the study will fail to show the expected result.

This is important because it means that the conclusion from a study using statistics is dependent on the time the sample is selected. That conclusion is valid only for the moment of the selection. So it is potentially worthless in any other context.

Even the answers to interviews will differ according to the time they are undertaken.

This is a problem. Heck, it’s a mega problem!  Damn it, if you are conducting clinical trials it is more than that: it is a disaster. Indeed why even conduct the study? Instead of picking a sample just cast a chart and it will tell you exactly the same (biased) answer!!!

Of course I am exaggerating. Very long term repeated studies should remove most of the short term planetary influences, but we should still be aware that they are, for example, being conducted under a Pluto – Neptune sextile and will reflect all the characteristics of that. However this might not matter as the conclusions will also be applied under the same planetary configurations. The important thing, as in all statistical results, is to understand the limitations and bias inherent in the study.

An astrological example of that comes from Jung:

Jung did conduct a major study and demonstrated that otherwise inexplicable astrological patterns did arise, but it was not through individual charts but through pairings. He looked at the Sun and Moon signs of hundreds of long term couples and he found that certain combinations (e.g. Moon conjunct Moon), were far more common than could be explained by chance.[iii]

Although of course I would be contradicting myself if I did not note that the moment he commenced the study would have contributed to his results, it is interesting that i) the candidates were selected over a long period and ii) what was being studied involved long relationships so short term planetary factors were of less importance. Nevertheless, we cannot fully understand the results without knowing when the study started etc.

And, of course this does not mean that every study is flawed; only those that rely on statistical sampling and inference/extrapolation. If I take blood from everyone in a room of 10000 I can confidently state what percentage of people in that room have which blood group. But if I take blood from only 10 or 20% of them I cannot rely on the inference.

A really good example of how this works is an astrological study I did myself. In around 2005 I did a full study of the zodiac signs of the Chairs, CEO and CFO of the FTSE 100 Companies. As I looked at all 100 companies I had all the population data at that date. Some of the results were highly “significant”. But I realised that it wasn’t possible to extrapolate to say whether or not Gemini’s (there were noticeably fewer of those) were poor directors or even generally self-selected to do something else. All it did was describe the nature of what big company boards were looking for in the UK in those years. A study today would probably(? ha, Ed!) look very different (though privacy regulations mean I can’t undertake it as Companies House no longer gives exact birthdates!!!!!).

So there we are. Not only can you not prove or disprove astrology with statistics, but you need to be aware that you may not really be able to prove or rely on very much else with statistics either. 

Black Swans Rule (when it is their moment)

  [i] Letter from C.G. Jung – Astroview

Jung on Astrology - Jungian Center for the Spiritual Sciences

20 Carl Jung Quotes On Astrology | (astroligion.com)

[ii] History of Gauquelin planetary effects that created baffling puzzles (Abstract+Article ) (astrology-a The Gauquelin Controversy - Astrology News Servicend-science.com)

[iii] This did not class as a standard statistical study- candidates were not randomly selected from a population for example. So it was very easy for others to question the validity anyway. I use it as an example of the problem and related conditions only.