4 ideas
Market efficiency can never be tested alone, because judging whether returns are 'abnormal' requires a model of what normal expected returns should be. Any anomaly is therefore evidence that either the market is inefficient or the asset-pricing model is wrong, and the data cannot say which. This makes every claim of beating the market, or of market irrationality, conditional on an assumed risk model.
Fama and French sorted NYSE, AMEX and NASDAQ stocks from 1963 to 1990 by size, book-to-market and market beta. Once size was controlled for, beta showed essentially no relation to average returns, contradicting the CAPM's core prediction.
Firm size and book-to-market ratio capture the cross-section of average stock returns better than beta. Because prices are assumed rational, a high book-to-market ratio signals a firm the market judges distressed or risky. Its higher average return is therefore read as compensation for bearing that risk, not as mispricing.
When a strategy earns persistently higher returns, first ask whether it is being paid for bearing a risk the benchmark model omits, rather than crediting skill. Under this lens, value investors buying cheap, high book-to-market stocks may be harvesting a systematic premium for holding unloved, fragile firms. An appropriate multi-factor benchmark can make apparent outperformance disappear.