I'm going to hop in here since I think this is a very worthwhile discussion topic...
But first I'd like to bring up the
Dunning-Kruger effect, which is basically scientific backing to Socrates' "
The only thing that I know is that I know nothing." mantra.
Basically, the smarter you are, the smarter you perceive everyone else compared to you; the dumber you are, the smarter you perceive yourself to everyone else. The poorest performers are the least aware of their own incompetence, etc. So before we claim we're smarter than everyone else, consider that there's probably a lot that goes on beyond what we're aware of when it comes to how "other" seeding systems are implemented. [Or, at least that's what more intelligent people are likely to do]
At the same time, I'm not saying that existing utilized systems are the "best" systems. I'm a firm believer that additional knowledge does not necessarily point you towards to best answer. If you asked a random person which city has more people, Detroit or El Paso, they'll probably say Detroit. If you asked a geography enthusiast which city has more people, Detroit or El Paso, he might say El Paso since southern cities like that are growing and Detroit has been falling apart for years. Well, the answer is Detroit. (This is the argument I use to get casual sports fans to play in company Fantasy Sports leagues, btw).
Anyway, on the topic of the purpose of the ranking system...I find there's a lot of confusion between the two notions:
1. The purpose of a championship is to reward the best team (i.e. the purpose of playing the games is to find out who's the best)
2. The purpose of a championship is to give something to play for (i.e. the purpose of building a great team is to increase the chances of winning games)
I will fight to the grave that #2 is the way it should be looked at, primarily due to the lack of certainty in games. Moneyball/sabremetric afficionados will probably agree with me. This is why I agree with Colonels' stance that the idea of a ranking system is to see whom has had the "best season" not to see who is the "best team". However, I think that over an infinite number of games with perfectly balances schedules and no variance, the best team should always end up #1. Testing this, of course is impossible.
So we're back to how picking seeds (in a world not concerned with how "entertaining" the tournament is) would then ideally be based on a system that at least maintains that it behooves all teams to do their best in every game...you can't ask any more of the system because you can't prove that any system does more than that. Basically, I feel the whole quality-of-ranking-system debate is really just for entertainment as you can never guarantee the "best" team has had the "best" season by any stretch.
On top of that, my pet peeve in all of this is when a team is supposedly reseeded when a major player gets hurt, as it flies right in the face of my Option #2 above.
6/5/2013 1:45 PM (edited)