This idea has a lot of merit, and stands well assuming the size of the investment to be allocated is sizeable (perhaps $10 m +). For anything smaller, it pays more to avoid it: rather pick and hold specific companies in different industries, in the asset class better suited to those companies. Think NiPiLo (Nit Picking for the Long Term).

Why?

NiPiLo is not readily/widely replicable. Any latest proven market beating, shotgun effect (think "diversification") strategies, if successful, are replicated and cease to be outperforming.

It is the only legal way to hold to information asymmetry to one´s advantage (one takes the time and effort to dig deep into the fundamentals of non public entities, gets a piece of the action, and holds on).

It does not ignore wider market fundamentals, yet drills down to a level that goes undetected by the bigger investors and investment managers with loads of money to invest, but not enough resources to cover small capital hungry entities.

Of course, this is not meant to replicate the Penny Stocks strategies of the ´80s and beyond, but a sound back-to-basics approach for smaller sized investment portfolios.