The virtue of excellence

Sunday, August 29, 2010

Aretavianism in 1 post

  • The question of truth is poorly defined. Ask about methods of prediction, and then facts applied to the methods. The probability of being true is less than one ( p <>
  • For ALL useful non-statistical predictions outside the physical sciences (F=ma), the probability of their being true is much less than 1 ( commonly abbreviated p <<>
  • Example: The space shuttle, one of the harder pieces of engineering in recent memory, failed on 2/130 missions...a nearly 2% failure rate, which is itself a better rate than the Apollo program.
  • Since problems involving biology, people, or social systems are MUCH harder than problems with engineering, it's wise to consider the space shuttle's 98% success rate an unattainable high point...and hope to beat 50%, after lots of tests.
  • Because error is ubiquitous, the primary issue in ALL systems is error-management.
  • Error-management is almost identical to feedback systems. How well do systems respond and get better WHEN (not if) the errors happen. (And how well do the feedback systems get better).
  • Mostly, this is unrecognized because people do not have statistical analysis engines, but instead status-driven monkeybrains.
  • Monkeybrains equate failure with lower status, and so do not admit of failures, in order to protect status rather than improving.
  • Wealth is a measure of how much one can get of what one wants. It is also basically equivalent to total crystallized practical (useful) knowledge.
  • Wealth is the social system success metric (wealth can be exchanged for ALMOST anything, happiness included. Often it buys something else instead of happiness.)
    • When people are wealthier, they do spend more of their wealth on things besides material goods, including increased health, decreased violence, increased environmental protection, increased care for the poor, and such.
  • Therefore, increasing wealth is better than ALL other social options, almost all the time, particularly given that growth rates are exponential, not linear.
  • In social/economic systems, the god-particle of feedback is exit. (I choose to buy from Comcast because AT&T sucks...or else vice versa)
  • High feedback allows consumers to switch to higher-value choices, thus creating further gains from trade, but also leads to high business-turnover.
  • Monopoly/Cartel prevents exit, and is the prime evil. Basically all social ills are caused by monopoly power preventing exit.
  • Force/Violence is the primary cause of monopolies, and the almost exclusive cause of cartels.
  • Government is an effective monopoly over force, and therefore both the normal cause of monopoly, and almost the only cause of cartel (CAFE, etc. killed the chance for revolutionary small automakers).
  • Government's activity is best understood using public choice analysis, and the nominal form of government really doesn't predict much.
  • Prediction: ALL government legislation everywhere (+/- 3%) decreases competition by circumventing the natural feedback system, thus increasing the profits of the incumbents in an industry. Especially true of rules regarding political action.
  • Education is no different. Feedback dominates.
    • If the time between acting and getting information about success is long, nearly no learning happens.
    • If the teacher gets no feedback on method, nearly no improvement happens.
    • If the school system has no feedback (via exit), the system never improves.
  • Healthcare is no different. Feedback dominates.
    • If the doctor isn't compensated based on success, success doesn't improve.
    • If the consumer doesn't have different costs for different choices, costs don't decrease.

1 comment:

drpat said...

One slight disagreement: compensation is not the only form of feedback that doctors receive.

Many/most doctors also receive personal satisfaction from helping the sick. And they receive increased status from being successful at helping the sick.

In practice, this doesn't change your policy prescriptions much.