F/+ Top 10 (Week 12)

[chaseball]

Just a quick note before I get to the ranking. F&P+ was the old way to refer to the system. The new abbreviation is "F/+" so I will be using that in place of F&P+ from here on out.

Latest F/+ Top 10 (updated today 11/13/2019)

1. OSU, 55.1%
2. Alabama, 54.1%
3. LSU, 53.5%
4. Clemson, 53.1%
5. Georgia, 52.0%
6. Oklahoma, 50.4%
7. Wisconsin, 49.6%
8. Utah, 49.4%
9. Penn St., 49.4%
10. Auburn, 46.3%

21. Navy, 39.0%
22. Notre Dame, 38.8%

Some notes:
Triple Option teams don't really fit the model as well as other teams that run more prototypical offenses so take Navy's rating with some additional uncertainty.

Despite Alabama's head to head loss, the system still thinks Alabama has performed better on the season as a whole vs their schedule than LSU has vs theirs. As a result, if Alabama were to face LSU in a hypothetical rematch this week, F/+ would have Alabama as the slight favorite.

Notre Dame way down in the 20s in this ranking kind of fits the mold of Brian Kelly's peaks and valleys in F/+ over the years. He'll have a 1-2 year run in and or around the top 10 and then have a head scratching season way down in the 20s-30s. At the end of the day it all averages out to somewhere around 15th over his career.

His best three year run is still 2010-2012. The season isn't over yet though, so ND can still move up the rankings with a good showing over their final three games.

 

[88ND]

I echo again..

Bama and Georgia were both ranked higher than Clemson last year AFTER the title game.

Moreover Georgia was ranked 30 spots higher than Texas after Texas literally steamrolled them in the sugar bowl.

Credibility lost I'd say. Just a tad

 

[tbonesays]

Bama and Georgia were both ranked higher than Clemson last year AFTER the title game.

they lost to better teams

 

[Catholicfan95]

Good thing this matters absolutely zero in reality.

 

[chaseball]

I echo again..

Bama and Georgia were both ranked higher than Clemson last year AFTER the title game.

Moreover Georgia was ranked 30 spots higher than Texas after Texas literally steamrolled them in the sugar bowl.

Credibility lost I'd say. Just a tad

You are looking at S&P+ .. not F/+.

Here are the top 3 in F/+ last year (F/+ is S&P+ combined with FEI).

1. Alabama 52.8%
2. Clemson 52.7%
3. Georgia 51.5%

S&P+ is looking at all kinds of granular data and not really taking into account the outcome of games. Where FEI is far more focused on final score, margin of victory, who won who lost, etc.

That's why the combined rating works so much better in this particular context.

 

[Catholicfan95]

You are looking at S&P+ .. not F/+.

Here are the top 3 in F/+ last year (F/+ is S&P+ combined with FEI).

1. Alabama 52.8%
2. Clemson 52.7%
3. Georgia 51.5%

S&P+ is looking at all kinds of granular data and not really taking into account the outcome of games. Where FEI is far more focused on final score, margin of victory, who won who lost, etc.

That's why the combined rating works so much better in this particular context.

"That's why the combined rating works so much better in this particular context." So ranking Alabama ahead of a Clemson team that beat them by 30 points makes sense, or works how? Why even play the game? Does not seem to matter to even play the game in this context.

 

[chaseball]

"That's why the combined rating works so much better in this particular context." So ranking Alabama ahead of a Clemson team that beat them by 30 points makes sense, or works how? Why even play the game? Does not seem to matter to even play the game in this context.

Alabama and Clemson played in 14 games last year, and only one of those games were head to head.

The system is rating the teams based on their performance throughout the season holistically and not overreacting to the results of any one game.

With all of that said, before the game Alabama had a large lead over Clemson in F/+ and after Alabama's head to head convincing loss to Clemson that lead was shrunk all the way down to a virtual tie.

 

[88ND]

Alabama and Clemson played in 14 games last year, and only one of those games were head to head.

The system is rating the teams based on their performance throughout the season holistically and not overreacting to the results of any one game.

With all of that said, before the game Alabama had a large lead over Clemson in F/+ and after Alabama's head to head convincing loss to Clemson that lead was shrunk all the way down to a virtual tie.

This is the problem or why nerds have too much time on their hands.

As with boxing styles make fights.

Team A can beat Team C by 24 points but Team B beat Team C by 7....

Team A and Team B play and Team B wins by 3.....

If both finish the year having Team A with only one loss and Team B undefeated ... Using that one example as the way their collective seasons went with Team A having a higher margin of victories in average ....

Your system will have Team A still ranked in front of Team B.

That's basically it in a nutshell.....and ridiculous

 

[Catholicfan95]

Alabama and Clemson played in 14 games last year, and only one of those games were head to head.

The system is rating the teams based on their performance throughout the season holistically and not overreacting to the results of any one game.

With all of that said, before the game Alabama had a large lead over Clemson in F/+ and after Alabama's head to head convincing loss to Clemson that lead was shrunk all the way down to a virtual tie.

The point is to determine which team is the best correct? So you are telling me, you believe in this system that ranks a team ahead of another team after they just got beat by said team?

I try to understand this I really do, but it get worse with every ranking. Why do you like it so much?

 

[lyonhawk]

Top 10 in Strength of Record (a metric the committee actually uses)

LSU
Ohio St
Baylor
Minnesota
Clemson
Penn St.
Auburn
Florida
Oregon
Alabama

 

[chaseball]

The point is to determine which team is the best correct? So you are telling me, you believe in this system that ranks a team ahead of another team after they just got beat by said team?

I try to understand this I really do, but it get worse with every ranking. Why do you like it so much?

I like it because despite its flaws it's the best objective data on how teams are stacking up to one another. What's your alternative in its place? What other system is able to objectively compare teams to one another who play widely different levels of competition? The other ones out there are nowhere near as robust/comprehensive as F/+ and have many of their own flaws.

The system is algorithm based with no manual intervention. The strength of this approach is that it omits any human bias from the rankings and just rates teams based entirely on their respective performance data throughout the season.

The weakness--and the strength depending on the context--is that the system can be slow to react to the outcomes of a single game.

I will maintain that just because Clemson beat Alabama convincingly in one game doesn't mean that in a rematch the outcome would be the same. That's essentially what the system is saying. Based on the history of both Clemson and Alabama last season--and even based on the performance of both teams in that single game--the system STILL thinks Alabama would be a slight favorite in a rematch.

There's a reason why there are best of 3/5/7 series in every other sport. Even when one team wins a game in a series (even convincingly) that doesn't necessarily mean that they will win the entire series.

And to 88NDs point, the system ignores matchups entirely. So maybe Clemson just matches up better to Alabama than all other teams did during the season, but every form of polling/ranking system struggles to take this into account.

The problem here is that in college football everything comes down to the outcome of one game where one game simply isn't enough of a sample size to determine who really is the better team. It's a limitation more with the college football format than anything else though.

 

[nemeth#5]

Surprised Missouri still isnt in the top 10..... after they lost to Wyoming they were still top 10 as the algorithm said their loss to Wyoming shouldnt of happened.... they than lost to Vanderbilt....... I guess that one must of mattered.

 

[twalsh]

I like it because despite its flaws it's the best objective data on how teams are stacking up to one another. What's your alternative in its place?

How about results, and watching games?

 

[chaseball]

Surprised Missouri still isnt in the top 10..... after they lost to Wyoming they were still top 10 as the algorithm said their loss to Wyoming shouldnt of happened.... they than lost to Vanderbilt....... I guess that one must of mattered.

Yea I remember that conversation earlier in the season.

In order to get initial rankings/ratings the system uses: recruiting class/talent rankings, last 5 years of weighted performance, and returning production from the previous season up until week 7 of the new season. As a result, Missouri was ranked pretty aggressively initially. But as more data came in during the new season (as we got further into the season), Missouri was ranked down more appropriately based on the results of their performance (they are now ranked 38th).

 

[chaseball]

How about results, and watching games?

I don't know any human being (or even a large group of human beings) who can watch all 60+ football games that take place every week and have no emotion/bias towards what they are seeing and then accurately rank all 130+ teams using the same exact criteria.

 

[Golson5]

Just a quick note before I get to the ranking. F&P+ was the old way to refer to the system. The new abbreviation is "F/+" so I will be using that in place of F&P+ from here on out.

Latest F/+ Top 10 (updated today 11/13/2019)

1. OSU, 55.1%
2. Alabama, 54.1%
3. LSU, 53.5%
4. Clemson, 53.1%
5. Georgia, 52.0%
6. Oklahoma, 50.4%
7. Wisconsin, 49.6%
8. Utah, 49.4%
9. Penn St., 49.4%
10. Auburn, 46.3%

21. Navy, 39.0%
22. Notre Dame, 38.8%

Some notes:
Triple Option teams don't really fit the model as well as other teams that run more prototypical offenses so take Navy's rating with some additional uncertainty.

Despite Alabama's head to head loss, the system still thinks Alabama has performed better on the season as a whole vs their schedule than LSU has vs theirs. As a result, if Alabama were to face LSU in a hypothetical rematch this week, F/+ would have Alabama as the slight favorite.

Notre Dame way down in the 20s in this ranking kind of fits the mold of Brian Kelly's peaks and valleys in F/+ over the years. He'll have a 1-2 year run in and or around the top 10 and then have a head scratching season way down in the 20s-30s. At the end of the day it all averages out to somewhere around 15th over his career.

His best three year run is still 2010-2012. The season isn't over yet though, so ND can still move up the rankings with a good showing over their final three games.

Wow that is a terrible ranking. Just a joke

 

[Golson5]

I like it because despite its flaws it's the best objective data on how teams are stacking up to one another. What's your alternative in its place? What other system is able to objectively compare teams to one another who play widely different levels of competition? The other ones out there are nowhere near as robust/comprehensive as F/+ and have many of their own flaws.

The system is algorithm based with no manual intervention. The strength of this approach is that it omits any human bias from the rankings and just rates teams based entirely on their respective performance data throughout the season.

The weakness--and the strength depending on the context--is that the system can be slow to react to the outcomes of a single game.

I will maintain that just because Clemson beat Alabama convincingly in one game doesn't mean that in a rematch the outcome would be the same. That's essentially what the system is saying. Based on the history of both Clemson and Alabama last season--and even based on the performance of both teams in that single game--the system STILL thinks Alabama would be a slight favorite in a rematch.

There's a reason why there are best of 3/5/7 series in every other sport. Even when one team wins a game in a series (even convincingly) that doesn't necessarily mean that they will win the entire series.

And to 88NDs point, the system ignores matchups entirely. So maybe Clemson just matches up better to Alabama than all other teams did during the season, but every form of polling/ranking system struggles to take this into account.

The problem here is that in college football everything comes down to the outcome of one game where one game simply isn't enough of a sample size to determine who really is the better team. It's a limitation more with the college football format than anything else though.

That system sucks

 

[Golson5]

I don't know any human being (or even a large group of human beings) who can watch all 60+ football games that take place every week and have no emotion/bias towards what they are seeing take place during the game and then accurately rank all 130+ teams every week.

Well I watched the LSU @ Alabama game. LSU won on the road and they have 4 top 25 wins. Alabama has 0 top 25 wins. Yet Bama is ranked higher? What a joke

 

[B_Agate]

The problem is football just doesn't give enough data for good statistical analysis. This isn't college basketball where you have 30+ games, or the NBA with 80+ or baseball with 160+. This is 14, with maybe 60-80 snaps per team per game, with each play having 11 moving parts on each side. Throw in the greater number of injuries and the data just isn't reliable. Let alone trying to do mid-season rankings.

 

[nemeth#5]

Yea I remember that conversation earlier in the season.

In order to get initial rankings/ratings the system uses: recruiting class/talent rankings, last 5 years of weighted performance, and returning production from the previous season up until week 7 of the new season. As a result, Missouri was ranked pretty aggressively initially. But as more data came in during the new season (as we got further into the season), Missouri was ranked down more appropriately based on the results of their performance (they are now ranked 38th).

So Missouri has lost to Wyoming, Vandy, Kentucky, and Georgia and they are still 38th..... seems high to me. But the fact that it had them in the top ten at any point should prove that it is completely a fraud? So one day it says they are damn good... then they lose to some bad teams and now it says they are average..... so what good does it even do? It obviously shows no fore-site for future outcomes.... and it doesnt even reflect correctly after the outcomes is established?

 

[chaseball]

The problem is football just doesn't give enough data for good statistical analysis. This isn't college basketball where you have 30+ games, or the NBA with 80+ or baseball with 160+. This is 14, with maybe 60-80 snaps per team per game, with each play having 11 moving parts on each side. Throw in the greater number of injuries and the data just isn't reliable. Let alone trying to do mid-season rankings.

I agree, sample sizes are tiny in college football, which is exactly what makes this type of system so essential. That is why the system is evaluating teams/performance at the play by play and drive by drive level. It isn't enough to look at the final score of a game and determine which teams are better. You need more granularity which this system provides.

As for injuries and such, you are correct. From what I understand the system doesn't have a way to account for this (maybe they weight recent performance more than performance during earlier games in the season? im not sure), but with 85 man rosters, the teams that have deeper/better rosters will be better equiped to prevail bad injury luck, and still perform optimally despite whatever roster attrition they are dealing with.

 

[lyonhawk]

The answer to which system is best is there isn’t one. Like all statistical analysis, you should take all the systems and use them as a data point and come to your own conclusions. Not just pick one and say I think this one is best so whatever it says must be right.

 

[chaseball]

So Missouri has lost to Wyoming, Vandy, Kentucky, and Georgia and they are still 38th..... seems high to me. But the fact that it had them in the top ten at any point should prove that it is completely a fraud? So one day it says they are damn good... then they lose to some bad teams and now it says they are average..... so what good does it even do? It obviously shows no fore-site for future outcomes.... and it doesnt even reflect correctly after the outcomes is established?

Earlier in the season there is far more volatility in the rankings, but as more data comes in from the current season, the system gets better and better at evaluating teams and ranking them accordingly.

At this point in the year, after 9 games or so Missouri is much more likely a team ranked in the 30s than they are a top 15 type of team.

 

[Catholicfan95]

I like it because despite its flaws it's the best objective data on how teams are stacking up to one another. What's your alternative in its place? What other system is able to objectively compare teams to one another who play widely different levels of competition? The other ones out there are nowhere near as robust/comprehensive as F/+ and have many of their own flaws.

The system is algorithm based with no manual intervention. The strength of this approach is that it omits any human bias from the rankings and just rates teams based entirely on their respective performance data throughout the season.

The weakness--and the strength depending on the context--is that the system can be slow to react to the outcomes of a single game.

I will maintain that just because Clemson beat Alabama convincingly in one game doesn't mean that in a rematch the outcome would be the same. That's essentially what the system is saying. Based on the history of both Clemson and Alabama last season--and even based on the performance of both teams in that single game--the system STILL thinks Alabama would be a slight favorite in a rematch.

There's a reason why there are best of 3/5/7 series in every other sport. Even when one team wins a game in a series (even convincingly) that doesn't necessarily mean that they will win the entire series.

And to 88NDs point, the system ignores matchups entirely. So maybe Clemson just matches up better to Alabama than all other teams did during the season, but every form of polling/ranking system struggles to take this into account.

The problem here is that in college football everything comes down to the outcome of one game where one game simply isn't enough of a sample size to determine who really is the better team. It's a limitation more with the college football format than anything else though.

It's a lot of explaining for a metric that is suppose to simplify ranking teams, seems to not be doing a good job to me. It's football, not baseball or basketball where best of 3/5/7 is a thing, the metric should take that into account, if it doesn't than it's not a good football metric.

 

[chaseball]

It's a lot of explaining for a metric that is suppose to simplify ranking teams, seems to not be doing a good job to me. It's football, not baseball or basketball where best of 3/5/7 is a thing, the metric should take that into account, if it doesn't than it's not a good football metric.

It's very simple on the surface and I posted the results in the OP.

There are posters like yourself who want more info/context so that they can trust the data they see. I'm just trying to help provide that extra clarity/information and in the process I'm learning more about how it works as well.

 

[88ND]

It's very simple on the surface and I posted the results in the OP.

There are posters like yourself who want more info/context so that they can trust the data they see. I'm just trying to help provide that extra clarity/information and in the process I'm learning more about how it works as well.

Chase

My question to you is this...

When I point out to you just how foolish a ranking is that has teams ranked lower than their defeated opponent your response is well...
F plus p w23 +-*/"+ t takes all tings into account like how many cigarettes the coach smokes to how many tattoos the players have.

I'm being friendly sarcastic to you but....

You do resoond with...

"yeah but"...


When you look at a list that has Texas ...after just dominating it's Georgia opponent in the Sugar Bowl ranked approximately 30 spots lower.... The two are only separated by one loss mind you and when playing one another the score wasn't even close to the actual game played. It was all Texas and they bullied the bully. Yet after the actual game was played your ranking system had Texas 30 spots lower.

Without trying to pull out all these nerdy formulas that you're defending...

I'd really like to know just how the above example makes you feel.

Or how did you like Bama being ranked ahead of Clemson after just getting their ass whipped with Georgia literally almost as high as Clemson.


You really think what of this all?

 

[chaseball]

Chase

My question to you is this...

When I point out to you just how foolish a ranking is that has teams ranked lower than their defeated opponent your response is well...
F plus p w23 +-*/"+ t takes all tings into account like how many cigarettes the coach smokes to how many tattoos the players have.

I'm being friendly sarcastic to you but....

You do resoond with...

"yeah but"...


When you look at a list that has Texas ...after just dominating it's Georgia opponent in the Sugar Bowl ranked approximately 30 spots lower.... The two are only separated by one loss mind you and when playing one another the score wasn't even close to the actual game played. It was all Texas and they bullied the bully. Yet after the actual game was played your ranking system had Texas 30 spots lower.

Without trying to pull out all these nerdy formulas that you're defending...

I'd really like to know just how the above example makes you feel.

Or how did you like Bama being ranked ahead of Clemson after just getting their ass whipped with Georgia literally almost as high as Clemson.


You really think what of this all?

I think we are at a stalemate. You can't look past the fact that teams can be ranked higher than other teams despite a head to head loss no matter how much information/context I think I'm providing to describe why that is happening. Everything that you just said in your post I've responded to in quite a bit of detail in this very thread.

I don't have an explanation for you that seems sufficient. If you look at my posts I'm not using any math whatsoever. I'm speaking in plain english.

I take your inquisitiveness on this issue as a major sign of respect for my opinion, otherwise you wouldn't be trying to make sense of something that runs counter to how you see things and I appreciate that which is why I've made an effort to elaborate.

I think this is just one of those things where we'll just have to leave it there.

 

[Golson5]

I think we are at a stalemate. You can't look past the fact that teams can be ranked higher than other teams despite a head to head loss no matter how much information/context I think I'm providing to describe why that is happening. Everything that you just said in your post I've responded to in quite a bit of detail in this very thread.

I don't have an explanation for you that seems sufficient. If you look at my posts I'm not using any math whatsoever. I'm speaking in plain english.

I take your inquisitiveness on this issue as a major sign of respect for my opinion, otherwise you wouldn't be trying so hard to make sense of something that just runs counter to how you see things and I appreciate that which is why I've made an effort to elaborate.

I think this is just one of those things where we'll just have to leave it there.

The correct answer is the system is BS.

 

[88ND]

I think we are at a stalemate. You can't look past the fact that teams can be ranked higher than other teams despite a head to head loss no matter how much information/context I think I'm providing to describe why that is happening. Everything that you just said in your post I've responded to in quite a bit of detail in this very thread.

I don't have an explanation for you that seems sufficient. If you look at my posts I'm not using any math whatsoever. I'm speaking in plain english.

I take your inquisitiveness on this issue as a major sign of respect for my opinion, otherwise you wouldn't be trying so hard to make sense of something that just runs counter to how you see things and I appreciate that which is why I've made an effort to elaborate.

I think this is just one of those things where we'll just have to leave it there.

The point being..
I'm a person of reality. Not fantasy. It might piss off some of the BK apologists but oh well... I'm as black and white as it gets.

What I don't understand is if you enjoy watching the game...and the intricate game inside the game that goes on how can you possibly put any stock into a system with such bizarre variances.

Moreover don't you think head to head should hold a lot more credence than the list is showing?

Georgia being 30 spots in front of Texas while records only separated by one game just ruins any credibility IMO

Perhaps to those that were never really involved in the game that formula might be fascinating but having been involved with the game for a long time it's very difficult to give this rating system credibility. Popularity contest? Sure. The prettier team? Yep. The team with the Al Davis motto of just win baby...not so much.

Here's a quote from your original post...

Despite Alabama's head to head loss, the system still thinks Alabama has performed better on the season as a whole vs their schedule than LSU has vs theirs. As a result, if Alabama were to face LSU in a hypothetical rematch this week, F/+ would have Alabama as the slight favorite.

That's nothing but a hypothetical laced with style points. The pretty team.

To me the most important thing is the eye test. LSU to this point has had a significantly more difficult schedule than Bama so how can you or
s plus p123 even say such a thing. Bama didn't go on the road to Texas. Hasn't played Florida. Hasn't played Auburn (yet).
So if they're comparing schedules there's no comparison whatsoever.
The only comparable up to now between LSU and Alabama is head to head yet for some bizarre reason the formula apparently puts very little credence in that. Certainly not enough IMO.

Wisconsin ranked 7th? Seriously? They beat a Michigan team soundly but got defeated even worse agaisnt Ohio state. Wisconsin also got beat by Illinois. A team Minnesota hammered who also beat the #4 team last week. Yet Wisconsin with 2 losses sits there at #7 and Minnesota isn't even in your top 10?

Alarm bells are ringing Chase

Back to my original question that you keep avoiding.

You've been on here enough over the years that it's safe to say you're a big fan. When anyone asks you about YOUR personal feelings regarding these rankings you go into this huge defense of the formula and how they came to THEIR ranking.

That doesn't answer anything and I echo again...you don't really say how you personally feel about things other than yeah but, the formula weighs this, that and the other thing. You have us all well versed in how they come to their rankings but you keep skipping the part of how you actually think of the above things I listed or others as well.

Your defense of the system is highly noted. Your actual feeling of it being BS or not is what you don't mention. Do you really believe Bama should've been ranked higher than Clemson after they got beat down by that very Clemson? Texas being 30 spots lower than Georgia who they just dominated.?.

 

[lyonhawk]

There are a lot of problems with the system (and almost all college football systems because of the vast disparity in team strengths). At the end of the day LSU has played a very difficult schedule, beaten every team on it, and basically dominated Alabama in Tuscaloosa. It's very hard to put any relevance in a system that says Alabama (who has basically no quality wins) is currently better than an undefeated team who beat them head to head in their own stadium and has played a much more difficult schedule based solely on the fact that Alabama runs up the score against a bad schedule.

 

[Catholicfan95]

I think we are at a stalemate. You can't look past the fact that teams can be ranked higher than other teams despite a head to head loss no matter how much information/context I think I'm providing to describe why that is happening. Everything that you just said in your post I've responded to in quite a bit of detail in this very thread.

I don't have an explanation for you that seems sufficient. If you look at my posts I'm not using any math whatsoever. I'm speaking in plain english.

I take your inquisitiveness on this issue as a major sign of respect for my opinion, otherwise you wouldn't be trying to make sense of something that runs counter to how you see things and I appreciate that which is why I've made an effort to elaborate.

I think this is just one of those things where we'll just have to leave it there.

I think the theme here is probably 90% of people who watch football and 99% of the people on this board are going to discredit any kind of metric that ranks team A over team B, when team B beats team A, point blank. All of the information that is put into this metric may be a good way to gauge teams but it is missing the most important factor which is head to head. Head to head should be the very first factor, why should it not be? I do not care if a team is "better" on paper than any other team, if they cannot win then they are not better, is that crazy to think? I mean based on this metric Bama won the championship last year, do you think that is right? That among a myriad of other problems makes me wonder again, why you value this data. You seem like an intelligent guy, your into numbers and I can appreciate that, but not every number matters and not every metric is worth its salt.

 

[Pennick4]

The greatest football coach in history was asked a few weeks ago how much he relied on analytics... Belichick replied less than zero.
Nerds need to find another hobby.