Weighted Factor Analysis

Weighted factor analysis scores each factor of an option, multiplies each score by a weight for its importance, and adds the products into a single “overall goodness” number, so options can be ranked. It appears as weighted pro/con lists, weighted matrices, school grades, college and car rankings, and product reviews (RTINGS scores a TV from 50+ weighted sub-factors). CF treats it as the dominant — and mistaken — model of multi-factor decision making.

CF’s core objection is mathematical. Factors live in different dimensions (price, cuteness, safety), so they are unlike terms that cannot be added; summing them is like adding acres, hours, and grams. Weights are really unit-conversion factors between dimensions, but there is no objective rate converting, say, cuteness into dollars. People convert each factor to a made-up “goodness” dimension, but the weights are arbitrary — typically reverse-engineered to match a conclusion already held (rank Harvard high, then weight whatever Harvard does well).

A deeper objection: weighted sums are a maximizer strategy that makes every tiny factor shift the total, yielding fragile conclusions hypersensitive to noise and rounding. CF wants robust judgments instead.

The proposed replacement is binary evaluation: convert each factor into a good-enough pass/fail sub-goal, then combine them by multiplying binaries — one failure fails the whole. This is the same move as Goldratt’s pro/con method of solving every con. It focuses analysis on a few key factors and breakpoints rather than blending dozens.