I have to disagree about the computer rankings (full disclosure: I do the national rankings for MaxPreps) and their accuracy.
But first, they're not perfect. For whatever reason, the algorithm heavily favors Texas and California teams, and the national rankings I produce are often significantly different than the national computer rankings. So I want to make it clear I'm not trying to claim too much ...
Still, after a full season's data, I will claim that, within California, they are a pretty accurate representation of a team's strength. Of course, if the difference between two teams is one rating point, that doesn't mean much, but in terms of those teams that are elite, and those teams that are pretty good, the system works well.
I also don't think you can game the system, but if so, I'd love to hear how. And also, what rankings do you feel are inaccurate?
All that said, the biggest thing that the rankings cannot account for are injuries and transfers, so that a team's actual strength when postseason begins may be different than its strength when the whole year is taken into account. But that flaw is true of any rating system that tries to be unbiased -- any judgement that requires a belief or feeling that a team for its game tonight is better than its overall record is much more subjective than one that takes into account multiple results.
To put it another way, let's say Team A has a better record and a better computer ranking than Team B, but the committee members vote 3-2 that Team B should be seeded higher because they feel Team A was lucky or Team B is now healthy or whatever. As a coach, I would rather have my full body of work taken into account rather than a committee's "feeling" on the day the seedings are done.