[From Bill Powers (980805.1040 MDT)]
Bruce Gregory (980805.0940 EDT)--
Thanks. I'm very comforted to know that all the things I thought I had to
decide, such as how to allocate my retirement funds, will take place without
my need to "decide" anything. I'll just wait for all the conflicts to
resolve themselves in the constant flux of life. What a relief!
Suppose you have to decide whether to put your retirement funds in Plan A
or Plan B. Let's say that you are using various data in reaching this
decision: the annual interest rate, the estimated security of the
investments, the age at which you can start withdrawing, and so forth. Each
plan has some advantages and some disadvantages relative to the other plan.
Plan A might, for example, grow faster than Plan B. So if you were
controlling for best growth rate, you would choose Plan A. On the other
hand, Plan B starts paying out sooner than the Plan A, so if you were
controlling for earliest payout, you would choose Plan B. If only one of
these criteria, such as early payout, were being used, you would have no
decision to make. You choose the plan that makes the error the smallest.
There's no alternative to consider.
But when you set up _two_ criteria, earliest payout and best growth rate,
you discover that simple logic stops working: there is a conflict. If you
control for best growth rate, you can't have earliest payout, and vice versa.
Various methods for resolving conflicts like this have been worked out. For
example, you could assign weights to each of the conflicting factors, and
sum up the scores of all the plans, picking the one with the largest score.
But once you've worked out that formula, there is no longer any decision to
make. You apply the formula to the perceptions, and out comes the answer:
Plan B. Given the weighting algorithm, there are no longer any alternatives
to choose or decide between. There can be only one answer. The same goes
for flipping a coin. Once you've accepted that method of reaching a
decision, you no longer have any decision to make. You flip the coin and it
indicates the alternative: tails, it's Plan B. If it's tails, Plan A never
even comes up for consideration.
To sum up: if there's no conflict, you just control for the thing you want.
If there is conflict, you have to choose or decide between the mutually
exclusive alternatives. If you already have a formula for making such
choices or decisions, you simply apply it and do as the result indicates;
there's no decision for you to make, because the formula gives only one
answer. If you don't have any formula, you have to reorganize until there
is some basis for deciding. That "basis," whatever it is, removes the
necessity to choose or decide between alternatives.