Olli Järviniemi comments on Distinguish worst-case analysis from instrumental training-gaming