I've been at this a few years and the comments and questions of why a certain team is ranked where never stop. (nor will they ever) The quick and easy answer is that NOTHING is perfect, nor will it ever be.
There are a million ways to create a ranking system. Heck, I have even changed mine quite a bit over the years. For me, the biggest problem is the sheer number of teams. When I first started, fewer than 100 school played six-man football. Now the number of teams in the system are over 250.
Below is how the system worked this year. It always gets better as the year progresses. That's what it is supposed to do. It goes from an 'educated' guess to having some real data behind it.
wk record % 45's
1 76-43 63.87% 55
2 84-33 71.79% 72
3 93-24 79.49% 65
4 77-32 70.64% 52
5 85-28 75.22% 62
6 96-18 84.21% 63
7 79-21 79.00% 54
8 64-15 81.01% 43
9 91-19 82.73% 69
10 90-16 84.91% 56
11 84-11 88.42% 60
overall 919-260 77.95% 651
This is not a points-based system. It is not about your one favorite team. This is not about ranking the top-10.
This system is trying to place all 254 teams in an appropriate place relative to who they played. It is an optimization problem.
An interesting fact occurred this week -- the system actually contracted. If you look at the top teams, they actually decreased. Many would argue this is a mistake, but in reality, the 'error' in the rankings is at its least due to this. As a side note, I suspect it will grow again during the playoffs, as the top teams are the ones competing.
There are a million ways to create a ranking system. Heck, I have even changed mine quite a bit over the years. For me, the biggest problem is the sheer number of teams. When I first started, fewer than 100 school played six-man football. Now the number of teams in the system are over 250.
Below is how the system worked this year. It always gets better as the year progresses. That's what it is supposed to do. It goes from an 'educated' guess to having some real data behind it.
wk record % 45's
1 76-43 63.87% 55
2 84-33 71.79% 72
3 93-24 79.49% 65
4 77-32 70.64% 52
5 85-28 75.22% 62
6 96-18 84.21% 63
7 79-21 79.00% 54
8 64-15 81.01% 43
9 91-19 82.73% 69
10 90-16 84.91% 56
11 84-11 88.42% 60
overall 919-260 77.95% 651
This is not a points-based system. It is not about your one favorite team. This is not about ranking the top-10.
This system is trying to place all 254 teams in an appropriate place relative to who they played. It is an optimization problem.
An interesting fact occurred this week -- the system actually contracted. If you look at the top teams, they actually decreased. Many would argue this is a mistake, but in reality, the 'error' in the rankings is at its least due to this. As a side note, I suspect it will grow again during the playoffs, as the top teams are the ones competing.