Forums

[CE80]

Originally Posted by Gazoo So, across an entire season, don't "the numbers help you decide" who the best teams are? (This assumes the point of the tournament is to select the best teams, which may not be your basis for selection.) In your scenario, you seem to be clearly painting a picture that the better team lost. It's unlikely that happened 25 times in a season, so across a large enough sample size the numbers should point you to the better teams. But you would instead select worse teams for the tournament that under-played their opponents and won on flukes? Or do we use the eye test when the stats fail to show us the expected answer? Just playing devil's advocate here, but, it seems like you either trust the metrics or you don't. Or, you trust a few of them but not others, which might be fair but would require a lot longer discussion than "sometimes yes and other times no" because there would need to be case-by-case reasons why you don't trust those metrics.

I would probably go along with you if everyone in the pool played everyone an equal number of times. Without that, I don't have enough trust in that type of system to differentiate between teams playing and therefore being evaluated on different competition. Ultimately, it is probably the same problem that has always existed - do the metrics adequately take into account the disparity between teams that have not had comparable schedules?