Beyond the signal-to-noise

There’s a hump in the curve, way, way out beyond the six sigma point …

Shea and Wilson were almost on it when the pointed out that a ‘conspiracy’ with five vectors is more than most people can handle.  Well not any more if you use computers, have good sampling and know how
to use autocorrelation to extract signal from noise. Communication their says that you can extract energy – and by implication information – from the heat death of the universe if you do it properly. However
seeing how we usually don’t act until we are at the watershed I suspect we won’t set up that structure for political reasons.

Do some research on sampling vs noise in that arena, look up the 1/f theorems and how widely they apply. Its either amazing or frightening depending on how you view it; sunspots, disease and terrorist attacks all follow the same pattern. We can dress it up, as people do nowadays by calling it “fractal theory”, but the maths I learnt, the maths that Shannon worked out in 1948, still hold regardless.

Well OK, we’ve been pushing the Shannon limit and cheating it for a long time now, but so what? In the lab as a teenager I was shown how to send signals[1] faster than light.

The “Delphi Method”  is one way of dressing up the “beyond six sigma” phenomena, but a little off the point. It assumes a coherency of expertise that defies the parable about the five economists having seven different opinions on some matter. It defies – or possible falls into a trap depending how you look at it – a couple of other things too.

The first is that it doesn’t allow for a “minority report”. You might call this a “Dissident report” or even “out of the box thinking”.  High Court Judges are not firebrand radicals by definition. Don’t expect revolutionary or innovative thinking from this method.

The second is the same thing by another means – call it filtering.
Suppose you do have an open discussion forum seeking a solution. How do you tell a radical and innovative idea from pure noise?

This is the real problem with what are termed “evolutionary algorithms”.

Pierre Teilhard de Chardin
Image via Wikipedia

Its not that they don’t work, its that they are easily fooled into loops or “local maxima“. Evolution is not as Pierre Teilhard de Chardin envisioned it with an Omega Point but rather a response to local pressure. A purely local optimization, not driven by any ambition or objective.

The real problem with a lot of “fundamental analysis” out of the B-Schools approach that get drilled into MBAs, and hence emerges in the advice firms like McKinley give is

a) the optimization curve is flat
b) really, its all about people

The “people” are key in a number of ways; the great salesmen, the great inventors, the great visionaries, the great leaders, the great support staff… and the great customers. Its not about what the
anthropologists call “Big Man” since all these are common and few are purely political. They are really part of teams and may be the catalyst rather than the dictator.
What this is about it recognizing “humanity” rather than the “technical” aspects of the process.

This emerges in many ways. Anyone experienced in security will recognise that “people are out greatest asset” also means they are the greatest risk, and there are many aspects to that risk. It is also easy for a great “asset” to flip over and become a great “liability”.

[1] Shannon would not agree. Einstein would laugh.
The definitions get tricky. Einstein would not laugh at Alcubierre.
This despite some physics:

Enhanced by Zemanta

About the author

Security Evangelist

Leave a Reply