Week 9 Polls

[Mayville Bison]

But you ignore the hot mess the rest of the conference is. The fundamental difference of why the MVFC is better is the top to bottom nature. The only traditional "lay ups" are Misery St. and Indiana St.; no one else. The Big Sky is riddled with lay up teams for the elite. So, if you combine your comparison of the interconference play with the horrible bottom of the Big Fluffy, there is no way to say they are better.

I'll give you that. Their crappy teams are crappier than our crappy teams and they have more of them.

To change the goalposts a little which group would you take

NDSU EWU
SDSU UM
WIU UND
UNI CP
YSU NAU
USD SUU
ISUr WSU

Just grabbed the top 7 (arguably) of each to take out the layups. BTW, great conversation!

[bisonaudit]

Where would UNI be in the Sagarin ratings if their win over Misery State had only been by Sagarin's predicted MOV? Seems like that blowout is giving them too much lift in the ratings. And yes I am not smart enough to figure it out myself.

Don't know. Can't isolate a single result without access to all of the data and the algorithm.

However, UNI's rating improved by 2.61 points following that game.

[thebootfitter]

Where would UNI be in the Sagarin ratings if their win over Misery State had only been by Sagarin's predicted MOV? Seems like that blowout is giving them too much lift in the ratings. And yes I am not smart enough to figure it out myself.

I'd have to either go look it up or have BisonAudit or Hammersmith comment, as they are probably a bit more up on how it works than I am, but I'm pretty sure Sagarin limits the impact of margin of victory after a certain threshold. For example, if you're expected to win by 10 and you win by 40, the model doesn't give you a full 30 points more for the extra margin of victory. In other words, I believe that a win by, say... 20 and a win by 40 will be fairly similar in terms of the bump you receive.

If anyone wants to clarify or correct, I won't be offended.

Edited to add: Oops, BisonAudit beat me to it, and it looks like I may be off base anyway. (I thought I remembered reading an article that mentioned Sagarin limits the margin of victory weight at some point. But I might have been thinking about some other rating system.)

Seeing Sagarin's brief explanations (which don't have any details into the actual algorithms), I wonder if I was reading about the "Golden_Mean" score, and that had a limited impact for margin of victory after a certain threshold?
----------------------------------------------------------------------------------------------------------------
The PREDICTOR, is such that the score is the only thing that matters.
PREDICTOR is also known as PURE_POINTS, BALLANTINE, RHEINGOLD, WHITE OWL
and is a very good PREDICTOR of future games.
----------------------------------------------------------------------------------------------------------------
GOLDEN_MEAN also utilizes the actual SCORES of the games in a different way but is also completely SCORE BASED
and thus should be a good match for the PURE POINTS in terms of predictive accuracy for upcoming games.
----------------------------------------------------------------------------------------------------------------
The RECENT, is score-based and weights RECENT play more heavily than earlier games. Its effect will become
more pronounced the longer a season goes if a given team happens to have an upward or downward trend.
----------------------------------------------------------------------------------------------------------------
The overall RATING is a synthesis of the three different SCORE-BASED methods, PREDICTOR(PURE_POINTS), GOLDEN_MEAN,
and RECENT and thus should be a good predictor in its own right.
----------------------------------------------------------------------------------------------------------------

[89MTBISON]

Don't know. Can't isolate a single result without access to all of the data and the algorithm.

However, UNI's rating improved by 2.61 points following that game.

Thanks. Thought it would be more than that. They obviously get plenty of help from SOS.

[thebootfitter]

Thanks. Thought it would be more than that. They obviously get plenty of help from SOS.

This is not really the right way to look at it. They help themselves by playing teams close to or better than expected. If they had a poor strength of schedule, in theory, they'd still have around the same rating because they'd be beating the worse teams by more or beating them period instead of losing close games to good teams.

Strength of schedule in computer rating systems is more a function of the results than it is an input that determines the ratings.

[89MTBISON]

Thanks to Hammersmith, I thought this was telling:
Rank Team Sagarin Predict Result Diff
83 Northern Iowa 66.27
84 Iowa State 65.71 -1.88 +5.00 +6.88
105 Montana 59.96 +8.75 -6.00 -14.75
77 Eastern Washington 66.98 -3.15 -4.00 -0.85
150 Southern Illinois 50.16 +18.55 +21.00 +2.45
143 South Dakota 52.80 +11.03 -3.00 -14.03
95 Youngstown State 62.60 +1.23 -4.00 -5.23
176 Missouri State 44.89 +23.82 +54.00 +30.18
56 North Dakota State 72.32 -3.61 -4.00 -0.39
159 Indiana State 48.48 +15.35
121 Western Illinois 56.16 +7.67
93 South Dakota State 63.11 +5.60

Total Diff: +4.26
Home Field 2.44
That one win over Misery state wiped out their two large loss differentials and took them from a large negative differential to a smallish positive differential. Does that have any bearing on their Sagarin rating? We beat em so....maybe none of this matters, just curious as to why UNI hangs around with 4 losses.

[Vet70]

Very easy for a voter to justify it.

1. Best player played half the game or so?
2. Played on the road in a very hostile environment and lost in OT.

I believe in the early 90's, Notre Dame beat Florida State. Both finished 10-1, and they gave the title to Florida State. It's not exactly unprecedented. Don't agree with it but it is what it is

This has been raised a few times and at least for me, that may not have made any difference. As I recall practically everyone Gabrud threw to in the second half caught the ball and scored.

[thebootfitter]

That one win over Misery state wiped out their two large loss differentials and took them from a large negative differential to a smallish positive differential. Does that have any bearing on their Sagarin rating? We beat em so....maybe none of this matters, just curious as to why UNI hangs around with 4 losses.

Imagine trying to come up with a rating system where that differential is as small as possible for all ~250 teams that play D-I college football. In a team sport where play of a single individual can sway the outcome. It is very unpredictable by nature, but he is accurately predicting outcomes around 75-80% of the time.

That's essentially what he is doing. Given that the system is the same for all teams, some teams may have a bigger differential at some point in the season than others. Teams that are on an upswing or downswing might see their differential increasing or decreasing week to week.

[bisonaudit]

Thanks to Hammersmith, I thought this was telling:
Rank Team Sagarin Predict Result Diff
83 Northern Iowa 66.27
84 Iowa State 65.71 -1.88 +5.00 +6.88
105 Montana 59.96 +8.75 -6.00 -14.75
77 Eastern Washington 66.98 -3.15 -4.00 -0.85
150 Southern Illinois 50.16 +18.55 +21.00 +2.45
143 South Dakota 52.80 +11.03 -3.00 -14.03
95 Youngstown State 62.60 +1.23 -4.00 -5.23
176 Missouri State 44.89 +23.82 +54.00 +30.18
56 North Dakota State 72.32 -3.61 -4.00 -0.39
159 Indiana State 48.48 +15.35
121 Western Illinois 56.16 +7.67
93 South Dakota State 63.11 +5.60

Total Diff: +4.26
Home Field 2.44
That one win over Misery state wiped out their two large loss differentials and took them from a large negative differential to a smallish positive differential. Does that have any bearing on their Sagarin rating? We beat em so....maybe none of this matters, just curious as to why UNI hangs around with 4 losses.

The way I think about the differentials (I don't know if this is right or not, it's just what I think) is if each game result is a point on graph the differential is a measure of how far above or below a line of best fit that point is. The algorithm is trying to minimize the differentials by drawing the best fitting line it can for every college football game in the data set (all 1,500 or so).

[89MTBISON]

Imagine trying to come up with a rating system where that differential is as small as possible for all ~250 teams that play D-I college football. In a team sport where play of a single individual can sway the outcome. It is very unpredictable by nature, but he is accurately predicting outcomes around 75-80% of the time.

That's essentially what he is doing. Given that the system is the same for all teams, some teams may have a bigger differential at some point in the season than others. Teams that are on an upswing or downswing might see their differential increasing or decreasing week to week.

Got it. Thanks. They have a very good team, their persistently high rating perhaps speaks more to how poor the rest of our division really is. Seems like there really are only a handfull of FCS teams who have a legit chance to win it all.