That's definitely a school of thought, but practically the opposite of the Swiss, which basically considers all players as tied at every stage, and the goal is to progressively induce conditions that will break ties (first with match wins, then buchholz, and so forth)Very surprising to see capoch in.
As far as the system goes, I am a bit of a fan of determining tie breaks with "actual" wins ideally but there's no doubt that a 2-2 score with wins vs lierrey and daut, for instance, would/should count a lot higher than wins vs barles and slam. Ideally there's a system in place that makes tie breaks extremely uncommon though. For instance, I would prefer such a system to give an advantage to a player who beats Lierrey, e.g., vs by having him face barles next game for the decider, rather than facing Viper. Such that a tiebreak didn't directly determine a player to "move on" but instead did so indirectly.