Value stocks, as "properly" defined here (in this piece) match the market. So do growth stocks. But value stocks, as "traditionally" defined, beat the market.

Some years ago, one of the value houses (I think it was Tweedy Browne) divided their stock universe into five quintiles from cheapest to most expensive. With one exception (and its ramifications), the quintiles performed over a period of some years in the order that a value investor would expect.

The cheapest quintile was the best-performing one and the second cheapest was the second-best performing one. So far, so good.

But it was the MOST expensive quintile that was a market performer. Meaning that the third and fourth cheapest quintiles underperformed.

The most expensive quintile was (almost by definition), the growth quintile. The authors of the study found that this group, in the AGGREGATE, had different characteristics from the rest of the universe (a double growth rate). All the other four quintiles were similar in their composition, with a mix of "dogs," low growers, and "sleepers" (future growth stocks).

But the high-growth quintile was only a market performer, because the higher growth prospects were just about offset by the higher valuations. This is just a straightforward application of the efficient markets hypothesis.

If the remaining four quintiles had similar operating characteristics, how would they perform? The two cheapest ones would be outperformers and two more expensive ones would be underperformers, based on regression to the mean.

The cheap 40% is the outperforming "deep value" universe, which is why I am a deep value investor. The expensive 40% is what I call the false value, or false growth, or "limbo" universe, which underperforms.

But the "traditional" value-growth split is 50-50. The "value" universe is the 40% "deep value" part, plus 10% (out of 40%) of the underperforming "limbo" stocks.

The "growth" universe, on the other hand, consists of the market-performing 20% of growth stocks and the remaining underperforming 30% "limbo" stocks. That's why "value" outperforms "growth," as traditionally defined.