Harbins...Get Lost

cincifbfan said:
Ok, well at least I'm not advocating for rewarding losses. As I stated by your logic, a team can go 0-10, but have ten "quality losses" and still be awarded with a trip to the playoffs. Let's say a team loses all 10 games by 1 point, but lost to: Elder, X, Ed's, Iggy, Pick Central, Mentor, LaSalle, Lakota West, Hoban, & WW, they should be in the playoffs because "of their impossible schedule." Come on man, and I'm the one with dumb comments?

Get a grip and realize your argument doesn't hold up under any scenario. Just win baby!!
Click to expand...

You've proven to be an idiot over and over again. You still don't understand subjective vs. objective measures. And yes quality of wins and losses definitely matter. As for rewarding only winning and losing, the Harbin doesn't reward you for beating a good team - it rewards you for beating a team that has lots of wins - big difference. And they kill you for losing, regardless of how good or bad the team you lost to was, and regardless of how close that game was.

As for your stupid, ridiculous, and just plain wrong "just win baby" mantra - the Harbin has proven over and over again that it doesn't necessarily reward you for "just winning" as lots of teams with good records have been passed over for teams with more losses. This isn't necessarily a bad thing - in many cases that was the correct result. The problem is that there is no way to know using the Harbin because it doesn't measure the right metrics in the right ways.
 
L Hand said:
You've proven to be an idiot over and over again. You still don't understand subjective vs. objective measures. And yes quality of wins and losses definitely matter. As for rewarding only winning and losing, the Harbin doesn't reward you for beating a good team - it rewards you for beating a team that has lots of wins - big difference. And they kill you for losing, regardless of how good or bad the team you lost to was, and regardless of how close that game was.

As for your stupid, ridiculous, and just plain wrong "just win baby" mantra - the Harbin has proven over and over again that it doesn't necessarily reward you for "just winning" as lots of teams with good records have been passed over for teams with more losses. This isn't necessarily a bad thing - in many cases that was the correct result. The problem is that there is no way to know using the Harbin because it doesn't measure the right metrics in the right ways.
Click to expand...
So I'm an idiot because you keep saying I am using words incorrectly when you think anything a computer does is objective. Got it:rolleyes:. I'm also aware how Harbins work with awarding points for beating teams with wins. It's a simple formula and in football simple = success. A loss is a loss regardless. We shouldn't under any circumstances award losing, period.... even if people like you have the SUBJECTIVE notion that it was a "good loss." So I will stick by my stupid, ridiculous and somehow wrong "just win baby" mantra.

I'm still waiting for anyone on here who is arguing about metrics to come up with ANY formula that can more accurately do what the Harbins do without including SUBJECTIVE criteria, such as strength of schedule or margin of victory. Those are definitely subjective, used in the absolute correct way, regardless if you like it or agree with it or not.

Our playoff system in Ohio is better than any out there, by a long shot! Do some teams feel left out, sure, but hey that's life and football is supposed to teach our young men about life lessons. The Harbin system works. Period.
 
Makes no sense to argue about a formula picking playoff teams. If you do not play all the same common opponents, then no matter what you do, it is subjective. Even if you do play the same opponents, A beats B, B beats C, C beats A which is the better team?

Play your league Schedule. Make it mean something. Goal #1. Win the league. Goal #2 do the best you can in the playoffs. Simple. Best teams at the championship part of the season rise to the top. Done
 
cincifbfan said:
So I'm an idiot because you keep saying I am using words incorrectly when you think anything a computer does is objective. Got it:rolleyes:. I'm also aware how Harbins work with awarding points for beating teams with wins. It's a simple formula and in football simple = success. A loss is a loss regardless. We shouldn't under any circumstances award losing, period.... even if people like you have the SUBJECTIVE notion that it was a "good loss." So I will stick by my stupid, ridiculous and somehow wrong "just win baby" mantra.

I'm still waiting for anyone on here who is arguing about metrics to come up with ANY formula that can more accurately do what the Harbins do without including SUBJECTIVE criteria, such as strength of schedule or margin of victory. Those are definitely subjective, used in the absolute correct way, regardless if you like it or agree with it or not.

Our playoff system in Ohio is better than any out there, by a long shot! Do some teams feel left out, sure, but hey that's life and football is supposed to teach our young men about life lessons. The Harbin system works. Period.
Click to expand...

Yes - you are an idiot because you don't know the difference between subjective and objective. You are also an idiot because you don't understand that practically any model can do a better job than the Harbin for more accurately judging relative strengths of football teams. You are now further an idiot of saying simple = success. So yes, you are an idiot multiple times over.

I don't know what you do for a living, but I hope to god it doesn't have anything to do with using numbers and/or math.
 
psycho_dad said:
Makes no sense to argue about a formula picking playoff teams. If you do not play all the same common opponents, then no matter what you do, it is subjective. Even if you do play the same opponents, A beats B, B beats C, C beats A which is the better team?

Play your league Schedule. Make it mean something. Goal #1. Win the league. Goal #2 do the best you can in the playoffs. Simple. Best teams at the championship part of the season rise to the top. Done
Click to expand...

I understand your "let everyone in" preference. But sorry, just saying you can't correctly model relative team strength if everyone in the region doesn't play everyone else is just plain wrong. Yes - the greater number of cross-over games between teams that are being compared to one another increases the accuracy of the ranking systems. No - you do not need complete cross-over in order to provide a system with relatively good accuracy.
 
L Hand said:
Yes - you are an idiot because you don't know the difference between subjective and objective. You are also an idiot because you don't understand that practically any model can do a better job than the Harbin for more accurately judging relative strengths of football teams. You are now further an idiot of saying simple = success. So yes, you are an idiot multiple times over.

I don't know what you do for a living, but I hope to god it doesn't have anything to do with using numbers and/or math.
Click to expand...

Here you are again throwing around more and more insults instead of actually contributing to a conversation. You called me Trumpian earlier, yet calling someone an idiot 4 times is from page 1 of Trump debate manual.

Throughout this thread, you have continuously stated that practically any computer model will do a better job than Harbins; yet, your only suggestion is to include margin of victory. You keep disparaging people's intelligence time and time again, yet you don't understand why margin of victory has no real statistical value for determining playoff caliber teams and how it would only cause harm to high school football in Ohio.

I want you, the clearly intellectual superior and statistical guru, to give all of the simpletons like myself a lesson on the ideal computer model. Since you continue to claim there are so many alternatives, please give us a multitude of options to civilly discuss. Perhaps you could include pros and cons of each model with its intricacies and break them down for the ignoramuses on here such as myself.
 
L Hand said:
I understand your "let everyone in" preference. But sorry, just saying you can't correctly model relative team strength if everyone in the region doesn't play everyone else is just plain wrong. Yes - the greater number of cross-over games between teams that are being compared to one another increases the accuracy of the ranking systems. No - you do not need complete cross-over in order to provide a system with relatively good accuracy.
Click to expand...
How can you take into account a coach playing his les experienced but more talented inside linebacker in the first 3 games to get him experience and maybe taking a loss or two that keeps them out of the playoffs, but the team is much better the last 3 games of the season as a result of the move?

Has there been any critical analysis of the accuracy of the computer based selections? Can we go back and say that the number 1 ranked teams in the Region end up winning the region 98% of the time?

I am very good at ranking HS track and cross country teams. However, I'm not perfect and after the top few teams, it is much less accurate. How accurate is Harbin or any other system after the top 2 teams in a Region? What is the overall accuracy for the complete Region.
 
Ok so even if we were to go with one of these said systems, what do you do about out of state/non OHSAA opponents? Rank the entire country? Not count them?
 
psycho_dad said:
How can you take into account a coach playing his les experienced but more talented inside linebacker in the first 3 games to get him experience and maybe taking a loss or two that keeps them out of the playoffs, but the team is much better the last 3 games of the season as a result of the move?

Has there been any critical analysis of the accuracy of the computer based selections? Can we go back and say that the number 1 ranked teams in the Region end up winning the region 98% of the time?

I am very good at ranking HS track and cross country teams. However, I'm not perfect and after the top few teams, it is much less accurate. How accurate is Harbin or any other system after the top 2 teams in a Region? What is the overall accuracy for the complete Region.
Click to expand...

All of those arguments are arguments for letting everyone in, not arguments for why a computer model isn't better than a coaches poll, and/or why other computer models aren't better than the Harbin.

You want to let everyone in - fine. But assuming you DON'T let everyone in, you have to decide who gets in. While the Harbin is better than a coach vote, it is far from the best possible system. Why not have a system that is better?
 
kingpin2010 said:
Ok so even if we were to go with one of these said systems, what do you do about out of state/non OHSAA opponents? Rank the entire country? Not count them?
Click to expand...

Do it exactly the way the Harbin does - include everyone in the system that plays an OHSAA team.
 
L Hand said:
Sagarin pure points or ELO chess; Massey Ratings, ESPN FPI, Sagarin Golden Mean, ...

How many do you want?
Click to expand...
Wow, how original of you to copy models with many inherent flaws. Man, why didn't we think of that.

So you don't have an original thought/model. Got it!
 
cincifbfan said:
Here you are again throwing around more and more insults instead of actually contributing to a conversation. You called me Trumpian earlier, yet calling someone an idiot 4 times is from page 1 of Trump debate manual.

Throughout this thread, you have continuously stated that practically any computer model will do a better job than Harbins; yet, your only suggestion is to include margin of victory. You keep disparaging people's intelligence time and time again, yet you don't understand why margin of victory has no real statistical value for determining playoff caliber teams and how it would only cause harm to high school football in Ohio.

I want you, the clearly intellectual superior and statistical guru, to give all of the simpletons like myself a lesson on the ideal computer model. Since you continue to claim there are so many alternatives, please give us a multitude of options to civilly discuss. Perhaps you could include pros and cons of each model with its intricacies and break them down for the ignoramuses on here such as myself.
Click to expand...

Please find me a gambler or statistician that says margin of victory has no statistical value.
As for possible replacements, I've already posted about 10 times - but since you asked - Sagarin pure points or ELO chess; Massey Ratings, ESPN FPI, Sagarin Golden Mean, ...
 
L Hand said:
Do it exactly the way the Harbin does - include everyone in the system that plays an OHSAA team.
Click to expand...
But how? You’d need to literally rank the entire country to have a system like that. How would you tell if that was a quality team without knowing the quality of their opponents, and the quality of their opponents opponents etc.
All the systems you bring up all rely on every team ranked together in a ecosystem basically.
 
cincifbfan said:
Wow, how original of you to copy models with many inherent flaws. Man, why didn't we think of that.

So you don't have an original thought/model. Got it!
Click to expand...

You - you never gave us alternatives.
Me - I did many times.
You - I don't like the alternatives you provided.
 
L Hand said:
Please find me a gambler or statistician that says margin of victory has no statistical value.
As for possible replacements, I've already posted about 10 times - but since you asked - Sagarin pure points or ELO chess; Massey Ratings, ESPN FPI, Sagarin Golden Mean, ...
Click to expand...

Margin of Victory has statistical value with gambling because the leagues are exponentially smaller. You can't use that with over 700 teams, it's not viable.
 
L Hand said:
All of those arguments are arguments for letting everyone in, not arguments for why a computer model isn't better than a coaches poll, and/or why other computer models aren't better than the Harbin.

You want to let everyone in - fine. But assuming you DON'T let everyone in, you have to decide who gets in. While the Harbin is better than a coach vote, it is far from the best possible system. Why not have a system that is better?
Click to expand...
I agree, but how accurate is Harbin? How do you know the accuracy of a different system? If it's not more accurate, what is the point. Is anything highly accurate beyond 4th ranked in a Region? Do we go from 96% accuracy to crap shoot after 6th? Just asking. It seems like there would be 20+ years of data to pull from to make it the most accurate it can be.
 
L Hand said:
Seriously? You think the Harbin is the only computer that exists? Jeff Sagarin (by himself) has 3 different ones (all better than the Harbin)!! The Harbin is probably one of the world's shittiest computer models. Still, the world's shittiest computer model is better than a coaches poll.

Responding to Kramerica, you can't really use what the Harbin would have looked like this year to trash the Harbin - way too many weird instances and not enough games. As you know for any computer model, the more games and the more input, the greater the accuracy.
Click to expand...
In Ohio we use it. That’s what I meant. Not other places. There could be millions and millions of computer models based on a 10 game schedule. The Harbin system works.
 
If someone says we should use margin of victory again just think about this. there is a school in the conference that my Favorite school plays in. Every year they lose multiple games by 40+ points. Almost every game they play the starters are out by halftime. Margin of victory was a playoff metric, people would be forced to keep their starters in and run the score up. We’re talking about possibly 70-80-90 point blowouts every week. And obviously the state would be forced to get rid of the running clock because of the playoff metric. Kids would quit football around the state in droves. Who the hell wants to play for a team that loses by 70? it’s dumb . Some years a region may be better than another region, and a team or two will be left out of the playoffs because of their region. That doesn’t mean we should completely redesign our playoff system. They already moved it to 12 teams per region. That is going to let even the fringe teams into the playoffs.
 
L Hand said:
Sagarin pure points or ELO chess; Massey Ratings, ESPN FPI, Sagarin Golden Mean, ...

How many do you want?
Click to expand...

Great,

Let's use a program that gamblers rely on, ones that the author will not reveal all the details of said program, one that uses the opponent's strength during a loss, one that has a travel, rest, and injury component built into the rating, et al...,

It's High School football.... It's why the Harbins exist. Simple and to the point.

No system is perfect, but the more qualifiers you add (See Sagarin) the more unintended consequences can occur.

It's high school football.
 
Keep the Harbins. If we want to be like every other state around us and play communism ball, then you'll slowly watch Ohio football...look like those states.

This year was a wise exception by OHSAA given the circumstances.
 
thehawg27 said:
Keep the Harbins. If we want to be like every other state around us and play communism ball, then you'll slowly watch Ohio football...look like those states.

This year was a wise exception by OHSAA given the circumstances.
Click to expand...
It's worked for 48 years....

They aren't going anywhere.
 
Harbins do not rank teams, they identify playoff teams for the OHSAA. Originally, all games were played at a neutral site, the seeding was a tool to determine who played who; but was not always followed. When the OHSAA went to 8 teams in a region, the top 4 in Harbins got to host a OHSAA playoff game, because of the trouble finding so many sites on short notice.
As to how accurate the Harbins have been in forecasting playoff titles, look at how many times a #1 seed has won and how many times a #8 seed has won

FYI D1 last 20 years, winners by seed:
#1 6
#2 7
#3 4
#5 1 [2016 St X]
#7 2 [2001 St I; 2018 St E]
 
Last edited:
Spread All Day said:
If someone says we should use margin of victory again just think about this. there is a school in the conference that my Favorite school plays in. Every year they lose multiple games by 40+ points. Almost every game they play the starters are out by halftime. Margin of victory was a playoff metric, people would be forced to keep their starters in and run the score up. We’re talking about possibly 70-80-90 point blowouts every week. And obviously the state would be forced to get rid of the running clock because of the playoff metric. Kids would quit football around the state in droves. Who the hell wants to play for a team that loses by 70? it’s dumb . Some years a region may be better than another region, and a team or two will be left out of the playoffs because of their region. That doesn’t mean we should completely redesign our playoff system. They already moved it to 12 teams per region. That is going to let even the fringe teams into the playoffs.
Click to expand...

That's why you have cut-offs and diminishing returns (as I have said the entire time).
 
Termite2 said:
Harbins do not rank teams, they identify playoff teams for the OHSAA. Originally, all games were played at a neutral site, the seeding was a tool to determine who played who; but was not always followed. When the OHSAA went to 8 teams in a region, the top 4 in Harbins got to host a OHSAA playoff game, because of the trouble finding so many sites on short notice.
As to how accurate the Harbins have been in forecasting playoff titles, look at how many times a #1 seed has won and how many times a #8 seed has won

FYI D1 last 20 years, winners by seed:
#1 6
#2 7
#3 4
#5 1 [2016 St X]
#7 2 [2001 St I; 2018 St E]
Click to expand...
Term - good research. Also, like the qualifier of rank vs identify. Very well stated. Harbins are just used to ID the teams that are eligible to compete for the title. They're not awarding titles based on finishes in the rankings.
 
Termite2 said:
Harbins do not rank teams, they identify playoff teams for the OHSAA. Originally, all games were played at a neutral site, the seeding was a tool to determine who played who; but was not always followed. When the OHSAA went to 8 teams in a region, the top 4 in Harbins got to host a OHSAA playoff game, because of the trouble finding so many sites on short notice.
As to how accurate the Harbins have been in forecasting playoff titles, look at how many times a #1 seed has won and how many times a #8 seed has won

FYI D1 last 20 years, winners by seed:
#1 6
#2 7
#3 4
#5 1 [2016 St X]
#7 2 [2001 St I; 2018 St E]
Click to expand...

Look these are facts: (1) if you care about the "best" 8 teams in each region getting in, there are better systems than the Harbin for determining who those are; (2) margin of victory is a valuable metric in determining who the best 8 teams are.

If you disagree with either one of these facts, you don't have enough brainpower to lightly brown one side of a piece of toast and I really don't care what your arguments are - go fight with someone else.

If you want to talk about the fact that you can never know for sure that you are identifying the "best" teams, that some other criteria than which team is "best" should be used to determine the eights spots (i.e. most deserving), that there are potential downsides to using a system that is actually better than the one we currently have, etc., I am willing to talk to you.

But if your argument is the Harbins work, quit your whining, just win baby, other systems are not viable, et al. I refer back to my second statement.
 
You must log in or register to reply here.
Top