Posted by arssanguinus on 5/17/2012 11:56:00 AM (view original):
So gil, what would you say is the solution - have the first few years at a job be a "mulligan" - only counting at a greatly discounted rate?
i would suggest a heuristic to measure the state of the school, and use that to figure out how well someone did. basically, i think it should work like this.
1) write a heuristic to measure the "state" of the school, lets call it the current state ranking. there is already a program that ranks programs pre-season, you can use that, in conjunction with recent success and prestige, to get a pretty good picture. its tough to go through and say, penalize a school with more bad sophmores, which is realistically a tougher situation and reflects less in the success of the school and pre-season rankings than other, more preferable situations. so even though it doesn't necessarily reflect on how hard this program is to coach a year, two years from now (how well is the ground work laid), at least you get a decent snap shot of how it looks now.
2) run that heuristic against every team, every season. then you can give coaches credit for the change in that figure, from the time they started, until they left. quickly turning around a bad program, and getting to the NT regularly, would then give more credit than simply maintaining a school at that level - or worse - taking a great program, and turning it into a perennial NT qualifier (low end qualifier i mean).
im not sure exactly how to weight the change in current state against success each season, but it should basically just be calibrated i think, so that a coach who is at a school for say 6 seasons, with 3 bad seasons (starting with a bad state) and then 3 early NT exits, would get more credit than a coach who took over a school making the NT every year, and have 6 seasons of early NT exits. over time, youd have to diminish this - like, the state of the school when you took over simply doesn't matter any more when you've been there 20 seasons. does it at 10? yeah, to an extent, but it is more about the early seasons.
one issue i see potentially is if job status is just "accumulated" instead of looking backwards, it might be tough to factor in. in that case, you should just not really penalize a coach for having an equally bad or slightly better season than their school has had recently, and give them extra credit for those first few NT appearances. so maybe, in that 3rd season when you make the PIT, you basically are getting like NT rd 2 credit, or when you make the NT1, might be like a sweet 16 for a coach at a school that regularly had NT1/NT2 type seasons. in real life, thats how people do it, you go wow, billy gillispie (had to use it) took texas a&m from 0-16 to the sweet 16 in such a short time, thats pretty impressive. where as at the same time, tubby smith might have made a sweet 16, and people went man when is he ever gonna make a final 4 again?
if seble wanted, he could definitely enhance #1 to look at how the potential for success, fundamentally, which would help to offset the biggest hole in the simple heuristic mentioned in #1, which is that a coach often takes a program, recruits decently, but just doesn't play it well, making the school look bad on paper but really a good coach can come in and be pretty successful. you could look for at the quality of players on the team, you basically have better groundwork if you have a couple stars you can lean on and if you have some usable players. also a team with more openings instead of more unusable players is better off. working in something like should give you a good measurement of the current state of the team.
i don't know if you have to go that far, i think this idea would be largely helpful using the simple, readily available factors in #1. although, i guess pre season ranking includes your recruits, so it would have to be tweaked a little, but thats the general idea i have, i guess.