Friday, July 10, 2009

Marco's Razor & Marcomony

Occam's Razor states that "entities should not be multiplied uneccessarily".When competing hypotheses are equal in other respects, the principle recommends selection of the hypothesis that introduces the fewest assumptions and postulates the fewest entities while still sufficiently answering the question.

In science, parsimony is preference for the least complex explanation for an observation. This is generally regarded as good when judging hypotheses.

Marco's razor states that in cases of imperfect and only circumstantial evidence, entities should be added/multiplied to obtain hypotheses that explain the most unusual aspects of the limited circumstantial evidence available.

Marcomony is a preference for explanations which explain the most about the variability of observations, even if the observations fit within a simpler explanation.

These rules of thumb are ideal for game theory in the analysis of geopolitics. These are the reasons:
1) There really are more entities in these situations. By introducing "representative" entities one at a time to the simplest models, more of the circumstancial evidence can make a best fit.

2)With some entities actively keeping secrets and there being interdependencies more of the information is partial/uncertain.

For game theory in the analysis of geopolitics eg. the Israel-Palestine conflict is simplified to a two entity game in game theory and most arm-chair strategy discussions. A simple addition of a third entity keeps the overall strategy simple enough for the layperson to examine and analyse while explaining better the unusual aspects of the conflict (in this case, the multi-generational longevity of the conflict, the lack of success from mediators)

In evolutionary science occam's razor seems to have been liberally used to simplify explanation of evolution. These have evolved to a raft of "rules", such as the Weismann Barrier, Central Dogma of evolution etc. which are taken as gospel within scientific papers. Challenges to these "rules" are taken as challenges to the overall theory of evolution, which they need not be.

The main motivation with these improved rules of thumb is to:
a) Avoid cliched science such that researchers do not truthify the oversimplifications that are the norm in the reporting of science.
b) Avoid the human instinct to extend the presumption of innocence/truth-telling to dicussions irrelevant to the enforcement or judgement of law.


Another example is the choice between exogenesis and geogenesis theories. All we have as evidence that pertains to this is circumstancial and imperfect. The usual rules of thumb dictate that geogenesis is the least complex explanation. There is no real "data" to fit into a model, but exogenesis explains more of the unusual aspects of the properties of dna based life.

Monday, July 06, 2009

Light of Other Days read

Rather than review this book formally, now that I've read it, I'll just ramble on various thoughts I have until they become coherent to me.

Arthur Clarke and co-author, more than a novel, seem to have sketched a kind of Utopian vision - where technology and circumstance conspire to result in a radically altered World, free of crime, violence and treachery in the face of impending doom. The basic precept can be summed up by the clamly phrase thus "Good information = Increased crime... Perfect information = No crime".

Although I have reservations about the simplicity of this concept, evidence about the effects of good information on crime and war are encouraging. The rise of the internet and video phones everywhere has considerable curtailed certain types of crime/war. New crimes that rely on this new technology have risen as well, but arguably, these new crimes have less scope to be as organised to be destructive to society as the old ones were.

My two objections existing from before I read the novel remain even given the assumptions of possibility within the book:
1) Perfect information is impossible. At the margin the limit of the speed of light will mean that one person will be able to know and react to something before another, and "first-mover advantage" will always be relevant and a source of moral hazard. Even the amazing technology of perfect vision of past and current events evident in the novel doesn't meet, say, the perfection of knowledge attributable to a higher being. The utopian vision, however, makes this point moot, as the available information becomes as good as it needs to be.
2) Both commerce and organised crime rely on an imbalance of information to some extent. If a customer knows as much as a supplier of a product, suppliers will not be able to make a reliable profit. At the margins, this could be considered criminal exploitation, but practically all profitable commerce relies heavily on trade secrets if not outright intellectual property. Not to mention the issue of electronic commerce if secure keys cannot be kept secure enough for people to trust it as a store of value.