Employing databases and statistical skills, academics compute with precision the 'beta' of a stock... then build arcane investment and capital-allocation theories around this calculation. In their hunger for a single statistic to measure risk they forget a fundamental principle: It is better to be approximately right than precisely wrong.