Thursday, January 22, 2015

Desultorily Recursive

Fantasy Football

Here's an image of some scattered dots. Each dot represents a "catch" in American football, with the X and the Y axes representing different variables involved in the physical act of catching the football:



As you can easily tell, this first graph--the "football catches"--is utterly random. The dispersal of outcomes across the X/Y axes occur in a haphazard, unguided fashion. A catch may occur with a certain degree of grip intensity, or less grip intensity. A ball may be dropped, or not dropped. It's all random, and the spread of possible outcomes is not influenced in any meaningful way by the weather, the players, the astroturf, the coach, the particular game or season--none of that. Only once the ball has been caught does a pattern begin to emerge. Before then, all is chaos.

Next, here's another image. It shows the rate of change over time in the "interceptor" strain of the genetic code of bonobo monkey populations between 1923 and 2013:



This second graph, equally easy to analyze, represents the ways that natural selection "guides" the bonobo monkey population through its development. The "interceptor" strain of the bonobos' genetic code takes effect after each bonobo offspring's birth, creating a tidy spread that produces the shape of a "bell curve." Note how the total spread of possible results is limited within certain confines: 0.1% to 5.0% rates of change. This is the effect that randomized natural selection produces: impartially limiting the spread of variables into a narrow hallway.

The Other Shoe

The pictures, if you haven't realized through reading the dross this one generated above, are reversed for the purposes of illustration. This picture is actually produced by a set of randomly generated numbers:



As you can see, it's random. That's what random looks like: it's formless; purposeless; nothing in particular. It can appear anywhere at any time, in any shape or form.

This picture is one of interception rates for quarterbacks in American football:



As you can see, it's not random. It's not random because, by isolating the most skilled "American football quarterbacks" on the planet, and graphing their behavior during an activity in which their body and mind is specifically focused on a goal, we produce a bell curve effect. This effect indicates intent: a visible proof that the quarterbacks are trying not to throw interceptions, and that some of them are more or less successful at this over time.

A graph of the entire human population throwing American-football passes would likely produce a similar bell curve, spread out over a larger interception rate than 5% (depending, of course, on the quality and effort of the interceptors).

Were this not American football, though, but a random game of being blindfolded, deposited on a piece of astroturf chosen by a computer's random number generation program, and then to throw the ball uwpards in a high wind, the places where the ball fell on that field would look more like the first graph--a random distribution.

When we see an output shaped like a bell curve, we know that non-random variables are involved. We may not know what those variables are, and we may currently lack the technology to identify said variables, but the presence of the bell curve is, like a traveler's cairn on the tundra, a decided form of communication.

A bell curve is a Rorschach test for the Hobbesian faithful, posing the question, "Do you believe this result could be produced by accident?" Like most popular Rorschach tests, the patterns feign randomness. In order to produce the desired result (some kind of judgmentally-friendly response from the subject, which the scientist can then publish about), the inkblot has to be meddled by someone exercising purposeful intent: Rorschach inkblots are symmetrized, producing shapes--non randomly--from which so-called scientists then draw conclusions as though the shapes were, in fact, random.

Mirroring the inkblot in order to produce a symmetrical image, which induces test subjects to draw comparisons between the symmetrical relationship and other things that have symmetrical relationships--lightformed symmetrical organisms, or human-built products--is a way for the scientist to hide her role in the process. The "scientist" wishes to draw conclusions about a person's origins or behavior, but not to believe that anything is responsible for those conclusions (much less her own actions), ergo she seeks information that will justify randomness. When randomness proves nothing, though, the scientistic priest must guide randomness--make it non-random--by imposing a guidance standard upon the project. Inkblot tests, then, become not random, but symmetrized, while in another feat of mental gymnastics, roughly-symmetrical bell curve spreads are ascribed to random, unintentional factors.

Look again at what "random" really looks like:

2 comments:

  1. I'm no statistics whiz, but I remember that random variables can and are often normally distributed. i don't know why or how, but I wager the notion of randomness has more specific meaning in statistics, that, as often happens in math, is in no apparent relation with the intuitive everyday idea

    still, it is interesting question. The scatter plot essentially shows that each value has equal probability of occurring. A normal distribution assumes that different values have different probability of occurring.

    ReplyDelete
    Replies
    1. No, you're exactly right--random distributions regularly produce "shapes" and other things that appear to have been created by a Prime Mover. The overwhelming majority of distribution matrices appear as fields of disconnected dots, but it is possible for a very small percentage of random matrices to display a geometric shape (or a picture of Admiral Ackbar saying, "It's a trap!" or other recognizable stuff).

      Delete