Popular Post

Tuesday, December 6, 2011

Lies, damn lies and then there is statistics

There are lies, damn lies and then, there are statistics. If I had received $100,000 every time I heard this, I’d be rich today. As a retired professional statistical analysis, having worked for large capitalist corporations, I took this fiction as a personal affront. It is time to set the story straight.
To prove the validly of statistical prediction, I under took a rigorous, scientifically sound analysis of flipping coins. To insure a valid study, a large number of trials needed to be studied (The Law of Large Numbers*). Using a team of trained professionals to supervise the volunteers, we flipped 10 billion coins. To avoid any possible selection bias*, a random assortment of coins were employed and no single coin was used more than once. To avoid physical bias* half of the volunteers were left handed and half right handed. The person selecting the coin was blind folded and the person flipping the coin was also blind folded (Double Blind Test **). The result of the flip was recorded by robotic vision recognition hardware (no human participate was allowed to observe the coin to remove possible Human bias*), and then all results were electronically transmitted to a secret CIA data base.
The results were mathematically computed using software designed by MIT professors and double checked by professors from The University of Heidelberg. All supervised by the Lucasian Chair of Mathematics at Cambridge University.
The results are indisputable. 50% heads with 50% tails, the average coin toss will land on its edge and balance there. I can safely predict that any future coin toss will land on edge, and balance there, being half head and half tail.
This Schrodinger coin toss state, where the coin is simultaneously head and tail, exist as long as you do not look at the coin. Proving that statistical claims are true, providing that you do not look.
This hard science approach validates those statistics you will hear on cable news shows. I don’t wish to hear anymore foolishness about statistics being falsely used to prove some unsupportable ideological claim.

* Terms of art for people who like to Google stuff.
** OK, this isn’t really a double blind test, but, if this was an op-ed in the Wall Street Journal it would fly.

No comments: