[Ntrails]

His underlying point is broadly right - "past performance is no guarantee of future success" and as such historical (which includes current at the time of making an investment decision) is a fairly poor predictor.

The argument that the fund research sellers peddle is that "we recommend based on the investment methodology in use, the processes, risk management, the people, etc etc". Otherwise what would they be doing that a simple google search couldn't?

Again, arguably far better metrics to find an active fund that will on average out-perform. They're not wrong.