Most Current FHSAA Rankings

[anonimis]

1 hour ago, PinellasFB said:

Great data.  I think it would tell us a more complete story if you also added the median rating for each class.  I have a feeling there are a handful of schools really skewing the data at the top.

In the box plots, the line inside the boxes are the medians.  The red dotted line is the overall mean.  There are a handful of schools at the bottom as well.
The overall skewness of 'rating': 0.04703871518510849.

My biggest question right now is why the scatterplot has rows?

CLASS

 

SIZE

 

DIVISION

 

[FSULOVER]

For those who suggested separating publics and privates, consider this.  How you gonna break up the teams below into classes?  Its a problem.  I included estimated students each.

With 3 classes you might get STA and the 2 AH schools in the largest class with very few others.  Are the 2 AH schools gonna play each other in the semis?  I wouldn't think they would want that.  And how is STA gonna feel facing off with AHP every year for a title.  Those two schools like claiming state titles and with 3 classes one is gonna win and one is not.  Neither of these 2 schools has EVER played another private school in a state final and I can see them putting up a fuss.  

With 2 classes, you could have Gibbons playing AHP or STA, which are MUCH larger.  Its gonna be AHP v STA in the finals or semis EVERY year.  Like above, they aint gonna like that. Unless there is a change, just give the smaller class title to Nade every year.

Throw all privates everyone into one class and it's Broward v the world.  How would it look having 4 Broward teams STA, AHP, Nade and Gibbons playing each other every year in the semis?  If you break them up in the brackets, its still gonna be a Broward v Broward final most years.   You could have teams playing other teams that are 4 times their size in the playoffs.

We got a couple monsters like AHP and STA that are kinda kept in check being in with public schools.  Put them alone with other privates and it could really be ugly.  Nade can beat them this year but they will have a downturn at some point like they have before.  And it could take away from a state final victory if people are saying STA beat a school less than one third its size. 

AMH plantation 2800

STA 2000

AMH Delray 1800

Columbus 1700

Trinity Ch 1400

Bishop moors 1400

Cardinal gibbons 1200

Bolles 1000

Mon Pace 900

Jesuit 800

CCC 600

Chaminade 600

Card Newman 500

[DarterBlue2]

We got a couple monsters like AHP and STA that are kinda kept in check being in with public schools.  Put them alone with other privates and it could really be ugly.  Nade can beat them this year but they will have a downturn at some point like they have before.  And it could take away from a state final victory if people are saying STA beat a school less than one third its size. 

AMH plantation 2800

STA 2000

AMH Delray 1800

Columbus 1700

Trinity Ch 1400

Bishop moors 1400

Cardinal gibbons 1200

Bolles 1000

Mon Pace 900

Jesuit 800

CCC 600

Chaminade 600

Card Newman 500

I get what you are saying and I mostly agree. It would not be practical in the state of Florida to breakout the privates separately. This in not Ohio or New Jersey.

I do have one question, though: If American Heritage Plantation's enrollment stats are correct, why are they in 2M? 2,800 is either 4M enrolment or at the very top of 3M. I believe its the former. 

[Dan in Daytona]

2 hours ago, anonimis said:

In the box plots, the line inside the boxes are the medians.  The red dotted line is the overall mean.  There are a handful of schools at the bottom as well.
The overall skewness of 'rating': 0.04703871518510849.

My biggest question right now is why the scatterplot has rows?

CLASS

 

SIZE

 

DIVISION

 

Don't want to be the class Dummy here, but what exactly is statistically significant  with your graphs and charts ? Please in layman terms. Are all classes in a sense similar top to bottom ? 

[anonimis]

30 minutes ago, Dan in Daytona said:

Don't want to be the class Dummy here, but what exactly is statistically significant  with your graphs and charts ? Please in layman terms. Are all classes in a sense similar top to bottom ? 

I would conclude that there isn't a significant difference between classes, school population, or rural/suburban /metro split.  Splitting teams by the current classes isn't making the class more competitive, since there is more variance within classes than between classes.

[Dr. D]

Your conclusion about size/population is suspect, since the Metro and Suburban cut-offs for population are different.  Your conclusion may be valid for these "groups" of schools, but I feel like you would have to run the rating against the individual school's respective enrollment to get a valid assessment of the impact of enrollment?   (But I wouldn't be surprised if your conclusion still held true.)

[anonimis]

2 hours ago, Dr. D said:

Your conclusion about size/population is suspect, since the Metro and Suburban cut-offs for population are different.  Your conclusion may be valid for these "groups" of schools, but I feel like you would have to run the rating against the individual school's respective enrollment to get a valid assessment of the impact of enrollment?   (But I wouldn't be surprised if your conclusion still held true.)

I took the rankings and enrollments from the FHSAA website and merged the two based on school.  I dropped any schools that I couldn't easily merge, ex: schools with duplicate names like Atlantic or spelling variation like Senior vs Sr.  With the remaining schools there is essentially no correlation between team rating and school population.

[Dr. D]

Thank you for your work and explaining your methodology.  Confirms what seemed intuitive initially within the parameters available. 

The next question would be whether the FHSAA Power Rankings as currently constructed are an accurate measure of the relative strength of teams, but I'll leave that for another day. 

 

[DarterBlue2]

30 minutes ago, Dr. D said:

Thank you for your work and explaining your methodology.  Confirms what seemed intuitive initially within the parameters available. 

The next question would be whether the FHSAA Power Rankings as currently constructed are an accurate measure of the relative strength of teams, but I'll leave that for another day. 

 

Now that's the million dollar question; for the measures of mean, median and variance are meaningless if the relative strength of the teams is not calculated correctly. 

If we are going to continue using the current methodology, which, to me seems flawed, at a minimum, the way in which it is calculated should be disclosed. That's the only way teams on the bubble can double check to see if their elimination/inclusion was even computed accurately regardless of any inherent flaws that the method may have. 

[anonimis]

13 minutes ago, DarterBlue2 said:

Now that's the million dollar question; for the measures of mean, median and variance are meaningless if the relative strength of the teams is not calculated correctly. 

If we are going to continue using the current methodology, which, to me seems flawed, at a minimum, the way in which it is calculated should be disclosed. That's the only way teams on the bubble can double check to see if their elimination/inclusion was even computed accurately regardless of any inherent flaws that the method may have. 

What do you think is causing the lines in the scatter plot?  It looks like there are 10 lines.

 

[DarterBlue2]

2 hours ago, anonimis said:

What do you think is causing the lines in the scatter plot?  It looks like there are 10 lines.

 

That's an interesting question. Normally scatter plots don't cluster that way. Normally, you would have random dots ascending from let to right, but not almost straight lines. As presented, there seems to be various linear relationships between the strength of schedule and your rating that increase your rating disproportionally when you win more games. I assume that the yellow dot in the top right hand quadrant represents undefeated Chaminade which has had a tough schedule and has won all its games. Contrast that with Cocoa green dot near top right, which has an equally tough schedule according to their criteria, but has a "loss" to St. Thomas. 

Had margin of victory been included in their calculations, then we would not have those "straight lines" in the data scatter plot. 

[Dan in Daytona]

48 minutes ago, DarterBlue2 said:

That's an interesting question. Normally scatter plots don't cluster that way. Normally, you would have random dots ascending from let to right, but not almost straight lines. As presented, there seems to be various linear relationships between the strength of schedule and your rating that increase your rating disproportionally when you win more games. I assume that the yellow dot in the top right hand quadrant represents undefeated Chaminade which has had a tough schedule and has won all its games. Contrast that with Cocoa green dot near top right, which has an equally tough schedule according to their criteria, but has a "loss" to St. Thomas. 

My favorite economist breaking it down for the masses. Preach Darter preach  

[PinellasFB]

On 10/21/2023 at 9:53 AM, anonimis said:

In the box plots, the line inside the boxes are the medians.  The red dotted line is the overall mean.  There are a handful of schools at the bottom as well.
The overall skewness of 'rating': 0.04703871518510849.

From looking at the box plots, it seems like there's not as much skewness as I suspected but 1R looks top heavy while 4S has a lot of bottom feeders dragging it down.  Interesting analysis!

[PinellasFB]

17 hours ago, anonimis said:

What do you think is causing the lines in the scatter plot?  It looks like there are 10 lines.

 

If I were to guess, this is either a quantization artifact from FHSAA secret formula or a possible bias that you've discovered in the formula.  Bias is the scourge of all statistics.  

[cribboy305]

6 hours ago, PinellasFB said:

If I were to guess, this is either a quantization artifact from FHSAA secret formula or a possible bias that you've discovered in the formula.  Bias is the scourge of all statistics.  

I hate this, When did football become a damn mathematics problem. Only florida

[DarterBlue2]

On 10/21/2023 at 7:45 PM, Dan in Daytona said:

My favorite economist breaking it down for the masses. Preach Darter preach  

@Dan in Daytona, looking at my life in the rear view mirror, I would likely have been more fulfilled being an Economics professor. Unfortunately, I found the process to get there via a doctoral degree both somewhat distasteful and quite costly timewise with a young family to support. 

Distasteful in the sense that most doctoral dissertations are actually not based on cutting edge research based on the interests of the candidate or quality of the of the analysis of the data gathered, but rather extensions of the view and theories promulgated by his/her predecessors, who may just happen to be one's "doctoral advisor." 

Costly in the sense that at a minimum, it adds on at least another two years beyond the completion of the masters degree, and often quite a bit more. All this while earning the princely wages of a teaching and/or research assistant. 

[Perspective]

While we're on the topic of FHSAA rankings, any idea why it takes 2-3 days to release updated rankings?   If the product is objective/computer-based, the updated rankings should be available after the last game is played Friday night (or Saturday, if any games continue over).   There will always be an aberration game (like Dillard playing last night), but I see no reason to hold off the release of the updated rankings based on something like that. 

[SportsFan]

1 hour ago, Perspective said:

While we're on the topic of FHSAA rankings, any idea why it takes 2-3 days to release updated rankings?   If the product is objective/computer-based, the updated rankings should be available after the last game is played Friday night (or Saturday, if any games continue over).   There will always be an aberration game (like Dillard playing last night), but I see no reason to hold off the release of the updated rankings based on something like that. 

Because nobody knows the formula so they want to make sure "certain flavored blue bloods" get their "extra boost" 

[DarterBlue2]

20 minutes ago, SportsFan said:

Because nobody knows the formula so they want to make sure "certain flavored blue bloods" get their "extra boost" 

I hope to God it's just ineptitude or lack of staff and that what you speculate above is not the case. 

[DarterBlue2]

Latest rankings updated to reflect last week's games below:

https://fhsaa.com/rankings.aspx?path=football

I have not looked at them carefully yet, so I reserve comments till later. 

[SportsFan]

3 hours ago, DarterBlue2 said:

I hope to God it's just ineptitude or lack of staff and that what you speculate above is not the case. 

In reality I hope not and what I said was more poking fun at it but I can't say with absolute certainty that there wasn't seemingly random "last week" shifts before the playoffs that defied logic 

For example I saw examples last year during baseball seasons of the top team in the region losing during district play to a lower seed than the number 2 team beat during district tournament play and despite being neck and neck the team who lost actually finished with a HIGHER gap over the #2 team who won then they entered the week 

Could it be a simple bad data point being entered that caused it or something off with the formula that seems to inconsistently reward a team for losing/winning at certain times? Probably

But when we don't know the formula we can't say for certain that it won't be rigged from time to time in "close races" when they prefer a certain team hosts instead of another

[DarterBlue2]

1 hour ago, SportsFan said:

In reality I hope not and what I said was more poking fun at it but I can't say with absolute certainty that there wasn't seemingly random "last week" shifts before the playoffs that defied logic 

For example I saw examples last year during baseball seasons of the top team in the region losing during district play to a lower seed than the number 2 team beat during district tournament play and despite being neck and neck the team who lost actually finished with a HIGHER gap over the #2 team who won then they entered the week 

Could it be a simple bad data point being entered that caused it or something off with the formula that seems to inconsistently reward a team for losing/winning at certain times? Probably

But when we don't know the formula we can't say for certain that it won't be rigged from time to time in "close races" when they prefer a certain team hosts instead of another

And that is why the current system is inappropriate. It is totally lacking in transparency. 

[Dr. D]

7 hours ago, Perspective said:

While we're on the topic of FHSAA rankings, ola idea why it takes 2-3 days to release updated rankings?   If the product is objective/computer-based, the updated rankings should be available after the last game is played Friday night (or Saturday, if any games continue over).   There will always be an aberration game (like Dillard playing last night), but I see no reason to hold off the release of the updated rankings based on something like that. 

They manage to announce the ratings and playoff pairings by Sunday afternoon after the Week 11 games, so I'm guessing the IT guy doesn't work weekends until then. 

 

 

[anonimis]

My guess on the update timing would be to wait for those Monday games to finish and any other issues that might take some time to clean up like getting the scores correct.

[SportsFan]

3 hours ago, DarterBlue2 said:

And that is why the current system is inappropriate. It is totally lacking in transparency. 

Bingo that's kinda what I was going for but with a bit of my classic "smartass attitude"