The 62% violent pac-man

I'm going to have to side with the game reviewer over the Harvard PhD in this one. Specifically, I agree with the statement that the study, which is now notorious for rating Pac-Man as 62% violent, while an interesting piece of research and potentially useful as a basis for other kinds of quantitative research, should never in a million years be used as a basis of policy or a metric in rating games. "Violence" is far too vague and subjective a concept to legalistically quantify by absurd metrics like "number of deaths", "deaths per minute," etc. If you're hellbent on providing a violence rating for a game, you have to have reviewers review it and score it. No stupid rules, no pointless micro-metrics, just a simple, "On a scale from 1 to 10, how violent is this game?" Get enough people to score it like that and you have a rough idea of how violent other people will perceive it to be.

Now, if you can highly correlate those legalistic micro-metrics to the essentially subjective metrics, fantastic. Go for it. But until you do, they have no business in policy.