Harbins...Get Lost

If you want to switch to a poll, make it ONLY the teams your school played.

For instance, you rank each school you play, as you play them. You can't change them other than slotting the newest team in your list.

For instance, if you play Week 1, that team is #1. You play Week 2, you put that team either #1 or #2, moving them other team down or leaving them at #1. After week 3, you place them 1st, second, or third. The other two can't swap order. Then fourth week, etc.

After ten weeks, you have ten teams ranked #1 through #10 (if played less than 10, you have #1 through #9, etc).

10 points for your opponent ranking you #1, 9 for #2, 8 for #3, etc. Whether you won or lost, you get ranked by all ten possible OHSAA opponents (if not an OHSAA opponent then you only get ranked on the games you played in state).

Then, adjust for division. For every division up you played, you gain 0.5 to the multiplier, for every division down is -0.5. So, A D1 playing D1 would be 1.0. A D1 playing D7 would be 0.7. A D7 playing D7 would be 1.0. A D7 playing D1 would be 1.3.

So, if a D4 played a D6, and the D6 school said the D4 team was the 4th best team they played all year, they'd get 7 (ranking score) * 0.9 (division adjustment) = 6.3 for your points for that game.

Take all ten of your eligible opponent's rankings and average them by the number of eligible games you played (9 games would be totaled and divided by 9, for example).

Take everyone's score in your region and that's your standings

This prevents teams from being graded by how they finished, rather than how they actually played in that particular game. if a team improved or declined as the year went on, the more recent opponents would reflect that accordingly.
 
At the end of the day the OHSAA could put together a system that took millions of data points on ever team in the state and uses those to put teams in the playoffs and someone is still going to be mad that the system didn't favor their favorite team.
 
At the end of the day the OHSAA could put together a system that took millions of data points on ever team in the state and uses those to put teams in the playoffs and someone is still going to be mad that the system didn't favor their favorite team.
I don't have a dog in the fight.

It's just a matter that but only awarding points for wins teams only want you schedule games with a reasonable chance at winning. The constant league hopping and avoiding open teams looking for opponents is proof enough the system is broken.

A is better than B, B is better than C. If B played C and wins, they get points. Play A, and they get no points.

If Massillon beating Vanlue is a better statistical indicator of playoff worthiness than a loss to Pickerington Central, there's a problem.
 
Vanlue can lose to Pick Central every bit as good as 5-1 Groveport Madison or Pick North. If someone is advocating including losses as positives, the formula needs include quality of loss, right? I can imagine some interesting end-game scenarios that will create.
 
If you want to switch to a poll, make it ONLY the teams your school played.

For instance, you rank each school you play, as you play them. You can't change them other than slotting the newest team in your list.

For instance, if you play Week 1, that team is #1. You play Week 2, you put that team either #1 or #2, moving them other team down or leaving them at #1. After week 3, you place them 1st, second, or third. The other two can't swap order. Then fourth week, etc.

After ten weeks, you have ten teams ranked #1 through #10 (if played less than 10, you have #1 through #9, etc).

10 points for your opponent ranking you #1, 9 for #2, 8 for #3, etc. Whether you won or lost, you get ranked by all ten possible OHSAA opponents (if not an OHSAA opponent then you only get ranked on the games you played in state).

Then, adjust for division. For every division up you played, you gain 0.5 to the multiplier, for every division down is -0.5. So, A D1 playing D1 would be 1.0. A D1 playing D7 would be 0.7. A D7 playing D7 would be 1.0. A D7 playing D1 would be 1.3.

So, if a D4 played a D6, and the D6 school said the D4 team was the 4th best team they played all year, they'd get 7 (ranking score) * 0.9 (division adjustment) = 6.3 for your points for that game.

Take all ten of your eligible opponent's rankings and average them by the number of eligible games you played (9 games would be totaled and divided by 9, for example).

Take everyone's score in your region and that's your standings

This prevents teams from being graded by how they finished, rather than how they actually played in that particular game. if a team improved or declined as the year went on, the more recent opponents would reflect that accordingly.

I've got a better idea - don't even THINK about switching to a subjective poll. As bad as the Harbins are, they are 100 times better than subjective polls.
 
At the end of the day the OHSAA could put together a system that took millions of data points on ever team in the state and uses those to put teams in the playoffs and someone is still going to be mad that the system didn't favor their favorite team.

Not me. I cheer for a team that I think, by and large, is helped by the flaws in the Harbin system. Yet here I am cheering for it to be scrapped in favor of something better. And it is not rocket science - I am not advocating a system that includes stats like game control or even offensive or defensive stats (even though that's what gamblers use - and believe me, more than anything else - they want an ACCURATE indication of the respective strength of the teams they are betting on). You could put together a greatly improved system with nothing more than game results and margin of victory.
 
The Harbin system works. Keep it. We had a coach in Poland vote the team who is the number 1 seed as dead last in order to attempt to get his team a higher seed.
This. Happened everywhere. And as mentioned, many areas of the state are isolated from one another so having coaches vote every year would be a disaster. I do believe that the Harbin's can account for SOS. Tweak.

Kramerica uses the Winton Woods example as proof but it is ONE example. They played a 9 game schedule. And Moeller did not help them at 2-8 that season. They were robbed, no doubt, but it is an outlier. On top of that, the playoffs were going to be expanded this year so problem fixed?

Last season Northwood and Gibsonburg missed the playoffs at 9-1 and 10-0 respectively. They played nothing but tomato cans. The Harbins got it right. Those two same schools would have more than likely made it under the crazy Kramerica proposal and get rewarded for playing Little Sisters of the Poor. Plus, as mentioned, playoff expansion.

The Harbins work. Keep tweaking. If we allowed coaches to vote Kramerica would be complaining about that as well. He is THAT guy.

Cosmo is Yappi's biggest Karen troll. He never played the game and lives in his parents basement in Olentangy and he is obviously very bitter about those two facts.
 
This. Happened everywhere. And as mentioned, many areas of the state are isolated from one another so having coaches vote every year would be a disaster. I do believe that the Harbin's can account for SOS. Tweak.

Kramerica uses the Winton Woods example as proof but it is ONE example. They played a 9 game schedule. And Moeller did not help them at 2-8 that season. They were robbed, no doubt, but it is an outlier. On top of that, the playoffs were going to be expanded this year so problem fixed?

Last season Northwood and Gibsonburg missed the playoffs at 9-1 and 10-0 respectively. They played nothing but tomato cans. The Harbins got it right. Those two same schools would have more than likely made it under the crazy Kramerica proposal and get rewarded for playing Little Sisters of the Poor. Plus, as mentioned, playoff expansion.

The Harbins work. Keep tweaking. If we allowed coaches to vote Kramerica would be complaining about that as well. He is THAT guy.

Cosmo is Yappi's biggest Karen troll. He never played the game and lives in his parents basement in Olentangy and he is obviously very bitter about those two facts.
Very few on here that are seriously advocating for a poll system - Kramerica certainly isn't.

For most people, the issue is not with a computer model generally, but with the one we have because it rewards things it shouldn't and punishes things it shouldn't and just plain isn't really fair.

I guess to the extent your "keep tweaking" statement is let's scrap it and go with something much better, I agree with you. As I said, lots of ways to make a fairer, more accurate system with nothing more than wins, losses, and margin of victory. However I do believe that margin of victory needs to be a component to any system that is seriously trying to provide an accurate representation.
 
Very few on here that are seriously advocating for a poll system - Kramerica certainly isn't.

For most people, the issue is not with a computer model generally, but with the one we have because it rewards things it shouldn't and punishes things it shouldn't and just plain isn't really fair.

I guess to the extent your "keep tweaking" statement is let's scrap it and go with something much better, I agree with you. As I said, lots of ways to make a fairer, more accurate system with nothing more than wins, losses, and margin of victory. However I do believe that margin of victory needs to be a component to any system that is seriously trying to provide an accurate representation.
Do not put words in my mouth. Harbins works. Unsure how you get "scrap it" from "keep tweaking?" The system works and without covid, was to be expanded this season so no more hurt feelings.
 
I think it is a bit deceiving to refer to Harbin's as a computer model. Yes I know the OHSAA says its a computer poll, but it really is just an easy formula that the average person can do. BTW I think Harbins are fine and lets get real here, no matter the hand wringing that goes on yearly here it is not going to be changed, otherwise they would have changed it when they added the 4 extra spots.
 
Letting everyone in is a disgrace. You should have to qualify into the playoffs. Do we really need 1-9 teams in the playoffs?

Oh I'm perfectly fine with the system that we've been using. I'm just saying if you want to change it and the OHSAA needs more revenue there is a path for an expanded playoff system.
 
Do not put words in my mouth. Harbins works. Unsure how you get "scrap it" from "keep tweaking?" The system works and without covid, was to be expanded this season so no more hurt feelings.

What doesn't "work?" Polls work. Letting everyone in works. Not having a playoff and voting a poll champion works. "Works" isn't exactly a high bar - heck it's no bar at all.

I agree that the Harbins work better than a poll (also not a high bar). Why are people so opposed to making it better? It clearly can be done (easily) and it doesn't cost anything to do it.
 
I think it is a bit deceiving to refer to Harbin's as a computer model. Yes I know the OHSAA says its a computer poll, but it really is just an easy formula that the average person can do. BTW I think Harbins are fine and lets get real here, no matter the hand wringing that goes on yearly here it is not going to be changed, otherwise they would have changed it when they added the 4 extra spots.

You can call it what you want - if you want to call it a relatively simple algorithm, that's fine with me.
 
There are already lots of blow outs in the first, and even the second rounds. Why add additional rounds of blow outs? Teams would be better off just playing a week 10 game that they have a chance to compete in. Bye weeks suck, just sayin'.

I don't disagree. I wouldn't change anything but if people want to get rid of Harbins here is an option.
 
So why not do what we are doing this year but just with some adjustments?

1. Play a 9-game regular season.
2. Teams can still opt-out of playoffs just like this year and schedule a 10th game with another opt out.
3. The teams & the OHSAA will split the revenue of first round & second round (week 10 & 11) playoff games. This will make up for lost home game for schools and give the OHSAA revenue they currently do not have.
4. The top seeds still get byes for week 10 which is good because they are the most likely to play 6 playoff games.

What the downside to this?
I agree. It's HS. It makes your league mean more. It stabilizes leagues. It allows for teams to get better during the season without being penalized for their QB being hurt games 1 and 2. Might cut down a little on transfers. Allows teams to take chances. Again, it's HS. The regular season is one part and then you go into the championship part of the season. All the other sports do it and it works just fine.

Don't have situations like has happened in our league where a team got into the playoffs with the same record as another team in the league that didn't get in but with one of the losses being to the team that didn't get in.
 
Give us your optimal model, please.

No matter what model people prefer there are easy holes to pick at. We see now how the coaches vote and the politics to that. If you go with a model like Fantastic 50 you’re encouraging running up the score which is a no go.
 
You don't have to add "quality losses" (I.E. a reward for "good" losses) to the Harbin System. What is needed (as I have posted many times over a period of many many years) is treating losses with an inverse Harbin penalty. Lose to a team, have points deducted for each loss the team that beats you sustains. Lose to a 10-0 Lakewood St Edward team have first level points deducted for the loss and second level points deducted for each game they lost. (In this fictitious example that would be none.) Lose to a 2-8 "Cleveland East" (now defunct school, used strictly for example purposes) have first level points deducted for the loss and second level points deducted for each game they ("East") lost. Which would be X number of points times 8 in this hypothetical example. IOW a loss to St Ed in this case would cause a little damage and a loss to "East" would be catastrophic.

Again for the posters that ALWAYS fail to understand this point, this is NOT "quality losses" it is a PENALTY for each loss you suffer. Lose to a very good team, the penalty is slight. Lose to a bad team and the penalty really stings.
This is an underrated suggestion. I would very much love to see some rankings over the years using this formula. It would be unfortunate to see Stebbins end up with negative points so often, but would also make those wins much more sweet as they further damage the opponent!
 
No matter what model people prefer there are easy holes to pick at. We see now how the coaches vote and the politics to that. If you go with a model like Fantastic 50 you’re encouraging running up the score which is a no go.

This is false. His model contains diminishing returns as the margin increases, as any good model should.
 
This is false. His model contains diminishing returns as the margin increases, as any good model should.

That doesn’t matter though, if point margin counts then running up the score is inherently incentivized. Even if there is diminished returns, those small returns add up. Winning by 14 is better than 7. Winning by 35 is better than 28. Running up the score isn’t just dropping 60 on a team. Being up 27 with 4 minutes to go and going for a td with the varsity instead of putting in the JV and running clock is running up the score.
 
Last edited:
Why are people so opposed to making it better? It clearly can be done (easily) and it doesn't cost anything to do it.
I am not opposed to making it better. I was just pointing out that this is a slippery slope. The pessimist in me can just see how many arguments will come out of teams running it up for margin of victory or arguments over how a loss to team A by 7 should matter more than a win over team B by 3.
 
I am not opposed to making it better. I was just pointing out that this is a slippery slope. The pessimist in me can just see how many arguments will come out of teams running it up for margin of victory or arguments over how a loss to team A by 7 should matter more than a win over team B by 3.
But what's the slippery slope?

Right now, the harbins only counts victories, but divides by total games played. If two victories can be worth different points, then so should two losses. Saying a loss to a D7 team that goes 1-9 is the same as a D1 that went 10-0 is foolish at best. You would never see anyone argue victories be counted in such a way.
 
I am not opposed to making it better. I was just pointing out that this is a slippery slope. The pessimist in me can just see how many arguments will come out of teams running it up for margin of victory or arguments over how a loss to team A by 7 should matter more than a win over team B by 3.

If you care about accuracy, margin of victory has to be included.
 
A lot of Karen-ing going on about how various coaches voted, and some very impressive troll jobs by many of the coaches, but the really interesting thing is despite how incredibly bad some of that voting was, the end result was still MILES ahead and FAR more equitable than the brackets the archaic 1940's Harbin computer produces.

Harbins are COMPLETE GARBAGE and for as outraged as so many of you homers are capable of getting on other subjects, I find it amazing that we can't channel this outrage for the common cause of modernizing this hopelessly outdated and incredibly unfair and illogical rating system.

I bet if someone broke one of the vacuum tubes on the Harbin computer, it would be so old that no one is making replacement parts for it anymore
 
Top