forums

WTF ranking model

People are getting waaaaaay too worked up about all these new ranking models and personal rankings media is putting out for clicks.

Just keep winning baby. It'll all work itself out in the end.
 
Advertisement
This guy does several sets of power ratings. I thought most deserving was supposed to take away preseason bias & just look at who you beat (& how you played during those games), but apparently not bc fsu wouldn’t be 1.
 
Imukat said:
This has to be the worst ranking model I’ve seen that’s not coming from a troll account.

Click to expand...



FSU at 1 and their signature win isn’t even in the top 30?

Miami drops 1 after blowing out their #3 from last week?

Does not compute.
 
ESPN’s FPI is worse than all of these.

Metric based ranking systems make me LOL because the justification for metric ranking systems is that you can’t trust people… problem is that people choose the metrics that metric ranking systems use to make the rankings. Which make them just as stupid as people.
 
HTXCane said:
This ones really bad. Pate says this is a computer rating and not his ranking but how can ND still be 7th?

Click to expand...

Pate’s is based on neutral field betting line… the teams in front of you would be favored on a neutral field, the teams behind you would be underdogs on a neutral field.
 
Pate has said he has 2.5 points per turnover adjustment in their model and assumes TOs are random, so removes them in future prediction. ND losing by 3 points to Miami on the road (3 points for home field) and -2 TOs (5 more points) is likely viewed as a win in the model.

Here is a crazier idea. Florida at LSU, loses by 10. Road team (+3) five TOs (10 points), so his model would view that Florida game as a win over LSU from a power ratings perspective. Probably why LSU is all the way down at 12.

1mg of Epi said:
Pate’s is based on neutral field betting line… the teams in front of you would be favored on a neutral field, the teams behind you would be underdogs on a neutral field.
Click to expand...
 
Advertisement
MikeyCanez said:
CFB fans getting angry about a ranking based on methodology they don't understand and won't look up. A tradition unlike any other.
Click to expand...
lol why would I look up something that's clearly ******* stupid? Any kind of reasoning that has fsu at 1 is laughably silly. Besides, it says it right on the graphic.
 
Kalikgod said:
Pate has said he has 2.5 points per turnover adjustment in their model and assumes TOs are random, so removes them in future prediction. ND losing by 3 points to Miami on the road (3 points for home field) and -2 TOs (5 more points) is likely viewed as a win in the model.

Here is a crazier idea. Florida at LSU, loses by 10. Road team (+3) five TOs (10 points), so his model would view that Florida game as a win over LSU from a power ratings perspective. Probably why LSU is all the way down at 12.
Click to expand...
Except turnovers are not all random. Some players are more prone to them than others, hence some teams are as well.
Lagway did not randomly throw 5 INT’s. Maybe 1-2, but the pass in the flat that was taken back for a TD was not random by any definition of the word.

Of Beck’s 2 INT’s, I would label one as random (Toney deflecting the pass), and one as not random.

This is why I hate any model that is programmed without a human watching the play/game. Too many nuances are lost in these models such that they are basically useless.
 
Last edited:
Agree with you on the turnovers, just know Pate has mentioned it several times on his show.

Thought it would help explain why we are down at 6 and ND is 7. No idea how Penn St is still where they are other than the preseason roster values being over valued.

AtlAtty said:
Except turnovers are not all random. Some players are more prone to them than others, hence some teams are as well.
Lagway did. Not randomly throw 5 INT’s. Maybe 1-2, but the pass in the flat that was taken back for a tD was not random by any definition of the word.

Of Beck’s 2 INT’s, I would label one as random (Toney deflecting the pass), and one as not random.

This is why I hate any model that is programmed without a human watching the play/game. Too many nuances are lost in these models such that they are basically useless.
Click to expand...
 
AtlAtty said:
Except turnovers are not all random. Some players are more prone to them than others, hence some teams are as well.
Lagway did. Not randomly throw 5 INT’s. Maybe 1-2, but the pass in the flat that was taken back for a tD was not random by any definition of the word.

Of Beck’s 2 INT’s, I would label one as random (Toney deflecting the pass), and one as not random.

This is why I hate any model that is programmed without a human watching the play/game. Too many nuances are lost in these models such that they are basically useless.
Click to expand...
I agree 100%, except Toney deflecting the pass wasn’t really random either. That was a freshman being a freshman. When you are playing young players, they will make “freshman mistakes,” and that’s not random - and therefore anything not random has to be accounted for in the model.

People overthink these to the point of them being useless, or they don’t really have the knowledge to know how to do them correctly. I’m so sick of all these stupid accounts putting out bull**** rankings for internet clout. But happy to ***** about it here to make DMoney more money, I suppose. 😂
 
Kalikgod said:
Pate has said he has 2.5 points per turnover adjustment in their model and assumes TOs are random, so removes them in future prediction. ND losing by 3 points to Miami on the road (3 points for home field) and -2 TOs (5 more points) is likely viewed as a win in the model.

Here is a crazier idea. Florida at LSU, loses by 10. Road team (+3) five TOs (10 points), so his model would view that Florida game as a win over LSU from a power ratings perspective. Probably why LSU is all the way down at 12.
Click to expand...
Interesting…. That would explain it.
 
MikeyCanez said:
CFB fans getting angry about a ranking based on methodology they don't understand and won't look up. A tradition unlike any other.
Click to expand...
The playoff committee is using a new methodology, and won’t tell us what it is. Which is exactly what we’d expect them to do.
 
Advertisement
Back
Top