|
|
Post by Bevo on Oct 28, 2016 14:10:25 GMT
Wow... His system has Western Michigan at #3!!
Texas is at #92 just ahead of Ooh La La.... sheesh...
1 -Alabama 34
2 -Clemson 29
3 -Western Michigan 28
4 -Michigan 28
5 -Boise State 25
6 -Tennessee 23
7 -Texas A&M 22
8 -Louisville 21
9 -Penn State 21
10 -Nebraska 20
11 -Ohio State 19
12 -West Virginia 18
13 -Wisconsin 18
14 -Washington 17
15 -Troy 16
|
|
|
|
Post by bluehen on Oct 28, 2016 14:17:52 GMT
Why not ? I believe in CJH. Where did you find this ?
|
|
|
|
Post by Bevo on Oct 28, 2016 15:06:49 GMT
Why not ? I believe in CJH. Where did you find this ? Oh... Sorry... I meant to post the link...
objectiverules.blogspot.com.br/
I think his system is pretty darn accurate as well..... from around this part of the season forward.
I mean, he had Texas at #1 after the first week! BAHAHAHA!
It takes a little time for the data accumulation to become meaningful.
|
|
|
|
Post by bluehen on Oct 28, 2016 22:16:56 GMT
Why not ? I believe in CJH. Where did you find this ? Oh... Sorry... I meant to post the link...
objectiverules.blogspot.com.br/
I think his system is pretty darn accurate as well..... from around this part of the season forward.
I mean, he had Texas at #1 after the first week! BAHAHAHA!
It takes a little time for the data accumulation to become meaningful.
It's a true 'standings' system based on a set of rules. Standings are better than rankings...just like playing is better than voting. |
|
|
|
Post by Bevo on Oct 29, 2016 5:05:58 GMT
Oh... Sorry... I meant to post the link...
objectiverules.blogspot.com.br/
I think his system is pretty darn accurate as well..... from around this part of the season forward.
I mean, he had Texas at #1 after the first week! BAHAHAHA!
It takes a little time for the data accumulation to become meaningful.
It's a true 'standings' system based on a set of rules. Standings are better than rankings...just like playing is better than voting. I aqree... it's the best NON-SUBJECTIVE ranking method I've seen. Should be used to choose the 4, or the 8 if we ever get there. |
|
|
|
Post by Hero on Oct 29, 2016 9:24:25 GMT
I like his system.
You could never use it during the first few weeks of the season.
At that point it makes no sense at all.
|
|
|
|
Post by tigercpa on Oct 29, 2016 11:31:11 GMT
I like his system. You could never use it during the first few weeks of the season. At that point it makes no sense at all. Sagarin's is the same...until you get commonality of opponents and a larger population of games, it's tough to really gauge things.
|
|
|
|
Post by bluehen on Oct 30, 2016 22:49:05 GMT
I like his system. You could never use it during the first few weeks of the season. At that point it makes no sense at all. Sagarin's is the same...until you get commonality of opponents and a larger population of games, it's tough to really gauge things.
No, not really. Sagarin offers a pre-season ranking....a starting reference point. In 'standings' based systems there are no pre season rankings and no free head starts. Every single team's 0-0 record has absolute equal value...like it should. |
|
|
|
Post by tigercpa on Oct 31, 2016 0:54:09 GMT
Sagarin's is the same...until you get commonality of opponents and a larger population of games, it's tough to really gauge things.
No, not really. Sagarin offers a pre-season ranking....a starting reference point. In 'standings' based systems there are no pre season rankings and no free head starts. Every single team's 0-0 record has absolute equal value...like it should. Does he really print a preseason rank? |
|
|
|
Post by GatorGrad on Oct 31, 2016 14:27:06 GMT
No, not really. Sagarin offers a pre-season ranking....a starting reference point. In 'standings' based systems there are no pre season rankings and no free head starts. Every single team's 0-0 record has absolute equal value...like it should. Does he really print a preseason rank? Yes, but my understanding is that the pre-season rankings are just based on prior performance data...as the current season moves along the current season data phases out the pre-season data from prior years. So the pre-season ranking is just for fun to be able to list something but unlike the human polls, it has no impact on the final ratings at the end of the current season. |
|
|
|
Post by bluehen on Nov 1, 2016 11:44:11 GMT
Does he really print a preseason rank? Yes, but my understanding is that the pre-season rankings are just based on prior performance data...as the current season moves along the current season data phases out the pre-season data from prior years. So the pre-season ranking is just for fun to be able to list something but unlike the human polls, it has no impact on the final ratings at the end of the current season. I think it's more than 'fun'. Wins and losses in week one are credited vs the value of those opposing teams' pre-season starting reference numbers.For example If Alabama was assigned Sagarin's # 127 'for fun' ranking and clobbered Sagarin's # 107 'for fun' ranked team would Alabama automatically shoot up to Sagarin's # 1 'for fun' ranking ? I don't think so. Free head starts are huge in these mythical systems. |
|
|
|
Post by GatorGrad on Nov 1, 2016 14:36:58 GMT
Yes, but my understanding is that the pre-season rankings are just based on prior performance data...as the current season moves along the current season data phases out the pre-season data from prior years. So the pre-season ranking is just for fun to be able to list something but unlike the human polls, it has no impact on the final ratings at the end of the current season. I think it's more than 'fun'. Wins and losses in week one are credited vs the value of those opposing teams' pre-season starting reference numbers.For example If Alabama was assigned Sagarin's # 127 'for fun' ranking and clobbered Sagarin's # 107 'for fun' ranked team would Alabama automatically shoot up to Sagarin's # 1 'for fun' ranking ? I don't think so. Free head starts are huge in these mythical systems. I'm not sure why that is relevant. Will any data from 2015 be included in the final 2016 Sagarin ratings? |
|
|
|
Post by tigercpa on Nov 1, 2016 15:27:12 GMT
Does he really print a preseason rank? Yes, but my understanding is that the pre-season rankings are just based on prior performance data...as the current season moves along the current season data phases out the pre-season data from prior years. So the pre-season ranking is just for fun to be able to list something but unlike the human polls, it has no impact on the final ratings at the end of the current season. I've googled and can't find one I know that his early rankings are based on prior performance and he uses an algorithm (Bayesian - the grass is wet, so it either rained or the sprinkler ran) to help connect the teams, 3which creates a bias, but like my original comment, once there is commonality and sufficient population, he removes the prior performance and only uses current wins and losses. |
|
|
|
Post by tigercpa on Nov 1, 2016 18:50:58 GMT
Why focus on W-L records ?
Margin of victory is better indicator - at least according to ESPN and Louisville fans.
|
|
|
|
Post by bluehen on Nov 1, 2016 19:32:35 GMT
|
|
