Benford's Law And A Theory of Everything
A new relationship between Benford's Law and the statistics of fundamental physics may hint at a deeper theory of everything
kfc 05/07/2010
- 57 Comments
In 1938, the physicist Frank Benford made an extraordinary discovery about numbers. He found that in many lists of numbers drawn from real data, the leading digit is far more likely to be a 1 than a 9. In fact, the distribution of first digits follows a logarithmic law. So the first digit is likely to be 1 about 30 per cent of time while the number 9 appears only five per cent of the time.
That's an unsettling and counterintuitive discovery. Why aren't numbers evenly distributed in such lists? One answer is that if numbers have this type of distribution then it must be scale invariant. So switching a data set measured in inches to one measured in centimetres should not change the distribution. If that's the case, then the only form such a distribution can take is logarithmic.
But while this is a powerful argument, it does nothing to explan the existence of the distribution in the first place.
Then there is the fact that Benford Law seems to apply only to certain types of data. Physicists have found that it crops up in an amazing variety of data sets. Here are just a few: the areas of lakes, the lengths of rivers, the physical constants, stock market indices, file sizes in a personal computer and so on.
However, there are many data sets that do not follow Benford's law, such as lottery and telephone numbers.
What's the difference between these data sets that makes Benford's law apply or not? It's hard to escape the feeling that something deeper must be going on.
Today, Lijing Shao and Bo-Qiang Ma at Peking University in China provide a new insight into the nature of Benford's law. They examine how Benford's law applies to three kinds of statistical distributions widely used in physics.
These are: the Boltzmann-Gibbs distribution which is a probability measure used to describe the distribution of the states of a system; the Fermi-Dirac distribution which is a measure of the energies of single particles that obey the Pauli exclusion principle (ie fermions); and finally the Bose-Einstein distribution, a measure of the energies of single particles that do not obey the Pauli exclusion principle (ie bosons).
Lijing and Bo-Qiang say that the Boltzmann-Gibbs and Fermi-Dirac distributions distributions both fluctuate in a periodic manner around the Benford distribution with respect to the temperature of the system. The Bose Einstein distribution, on the other hand, conforms to benford's Law exactly whatever the temperature is.
What to make of this discovery? Lijing and Bo-Qiang say that logarithmic distributions are a general feature of statistical physics and so "might be a more fundamental principle behind the complexity of the nature".
That's an intriguing idea. Could it be that Benford's law hints at some kind underlying theory that governs the nature of many physical systems? Perhaps.
But what then of data sets that do not conform to Benford's law? Any decent explanation will need to explain why some data sets follow the law and others don't and it seems that Lijing and Bo-Qiang are as far as ever from this.
Ref: arxiv.org/abs/1005.0660: The Significant Digit Law In Statistical Physics
On Twitter
Become a Fan on Facebook
Subscribe to the Feed
lumidek2
13 Comments
Silly
It's completely silly to assign this law with some mystery. Every good high-school student should be able to derive Benford's law.
Any quantity whose value may a priori be very different - differing by orders of magnitude - inevitably has a distribution that is quasi-uniform on the logarithmic scale.
It's pretty much equally likely for such a "universal" quantity to be between 0.1 and 1, as between 1 and 10. The multiplicative factors are just a matter of convention. That's surely the case e.g. for the price of stocks where the price of the "minimum, one stock" always depends on coincidences.
So because log(x) is distributed along a distribution, and the distribution extends by several decades, and is inevitably quasi-continuous or smooth inside, it follows that the probabilities for the first digits are fixed by the logarithmic law.
On the log axis, the period between 1 and 10 is copied many times. Log((N+1)/N) / Log(10/1) is the fraction covered by the first digit's being N. It's repeated everywhere, so it's clear that e.g. the probability of the first digit's being 4 is log(5/4) / log(10/1).
The most likely first digit is clearly 1, with probability log(2/1)/log(10/1). Intuitively, you may imagine that the numbers 12-19 are "almost as likely as "2-9" because they're just "shifted by ten" and there's not much difference between the behavior of "5" and "15" at the beginning. So "1" is more likely at the beginning than e.g. 2.
Benford's law proved above gives the exact result for the probability. Because the proof only uses elementary mathematics, we can't extract anything that would go beyond elementary mathematics out of it.
Reply
syb
32 Comments
Re: Silly
I'm no mathmatician so help me understand this. I interpret your explanation to mean it's a basic property of logarithmic numbers that this ditribution falls out, so there is nothing of interest here to discover and no underlying mystery. But if that were true, it should work with any data set, not just some datasets. It seems that if this were universal then it wouldn't require further examination, but it is selective, indicating there is some other element at work here.
Reply
johnswentworth
1 Comment
Re: Silly
Benford's law only applies to distributions over countably infinite sets, where we cannot construct a uniform distribution. So for telephone numbers, for instance, we have finitely many possibilities so we can easily choose 10 digits uniformly. But for almost all useful distributions over the integers (which implicitly includes distributions over the reals with finite uncertainty), Benford's law applies.
Reply
socrates8
3 Comments
real numbers
...i thought real numbers were not countably infinite?
Reply
Lanny Budd
1 Comment
Re: Silly
I thought that the question was that if Benford's Law applies to countably infinite sets and does not apply to the reverse, then does the fit of a set of data to the Law describe the data as countably infinite or not?
If so, then what does that say about the distributions in the article? And why should countability and noncountability vary with temperature?
Reply
mcg11
1 Comment
Re: Silly
One property that is different for EB and FD statistics is the concept of distinguishable vs indistinguishable electrons. This might be the underlying difference of the data sets. Possibly similar to countable or uncountable data sets or selection with or without replacement?
Reply
shazam
1 Comment
Re: Silly
lumidek2 say - "It's pretty much equally likely for such a "universal" quantity to be between 0.1 and 1, as between 1 and 10." Well, that's pretty much the question - by assuming that "universal quantities" are exactly those that fulfill this scale invariance, then the question is: why do so many phenomenon give rise to these sorts of (in your parlance) universal quantities? The article was pretty clear on the point. But if you simply take the widely occurring existence of scale invariant quantities for granted, then the nature of the question is not going to be understood.
Another way of looking at it: how/why do scale invariant quantities arise from physical laws?
Reply
dpfitz
1 Comment
Are all the fundamental forces nothing but simple PHASE relationships?
Benford may have been on to something but I think G. Lisi and S. Wolfram are too.
Stephen Wolfram's "A New kind of Science" gives us these three (3) important facts:
1. Math can only explain simple things.
2. We need a model to explain complicated things.
3. But - a simple model can explain complicated things.
I ask myself this: Is G.Lisi's model similar to Dr. Milo Wolff's spherical, standing wave model?
While Wolff's model is only of the electron, Lisi's E8 model may be a similar model of a more complete quark-electron-gravitational setup.
If so then you must ask:
Are all the fundamental forces nothing but simple PHASE relationships?
"in phase attract"
Type those three words - above - into Google (Include quote marks) to learn not only the basics of electricity but how all the fundamental forces work.
Or click the following link that will give you the same page in Google. http://www.google.com/search?hl=en&source=hp&q=%22in+phase+attract%22&btnG=Google+Search
Not only do all electric motors obey these phase laws but this entire universe does as well.
(Click link below.)
http://www.amperefitz.com/in.phase.attract.htm
Reply