New IHSAA State Wrestling Proposal

[Ch1StyleZ84]

1 hour ago, Coach Peck said:

I feel like Sumo is the best option.  The matches are short and therefore we can get the Tournament down quickly which is what the IHSAA will like.

Having been to few Sumo matches since I've been living in Japan I have to agree. Sumo is far and away more exciting than those other archaic wrestling styles. I mean, Folkstyle is one big snooze fest that should have been replaced by Freestlye long ago.

[FCFIGHTER170]

Btw I know this is a troll, Scott Kelly was a stud at Folkstyle!

[08grad]

14 hours ago, Scott Kelly said:

This thread is intended to propose a new format of the IHSAA state wrestling tournament.

 

1. We ditch folk style wrestling, and we wrestle Greco Roman instead.  Why: 

-Folkstyle has many more ways to stall relative to the other styles 

-Greco Roman is more entertaining for the kids with the throws etc., & it just forces action better

-A college coach could easily teach what isn’t known about folkstyle to a great Indiana wrestler, and it would be a fresh change for the kids when they wrestle in college in a new style that they could use many techniques they already know.

-I argue more interest would get generated in Indiana wrestling because multi sport athletes could more quickly pick up Greco Roman and have more fun with it

 

2.  We ditch double elimination bracket style for a single league style with the winner of state having been the wrestler who accumulated the highest “score” similar to a chess latter system.  Score would not start at zero at state, starting score should be developed by your wins and losses over ranked opponents in the season and by how great a margin your victory was.  This rewards you for competing throughout the year, does not penalize you for switching weight classes-but rewards you for switching weight classes to try to beat higher score opponents near your weight and all the while allowing it possible for you to win at the end of the year even if you did not have a perfect year.  A computer program gets built at state to assign matches every round based on score where if you win a round, you wrestle a higher score ranked opponent than before (in general).  Why:

-This score method helps eliminate biases present in our current system such as school class and location.

-I argue This would encourage more risk taking, I’m not a fan of “you lose once you can’t win it all” at state, and not a fan of you lose twice you’re done.  I’m not a fan of whatever happened in season stops mattering when we start the state tournament.  As we know, so many factors make this unfair in current system (location, class, etc.)  Now you could potentially lose, but still win it all by score, because you went potentially 20-4 in season because you moved up a weight to try to take out big names at some duel meets and lost 4 times at state but you took risks and hit a couple of sick lateral drops and won a lot of matches and your score got boosted enough to win it all.

 

3.  The entire state tournament would need ran inside a Friday-Sunday period somewhere.  Why:

-The current 4 week system feels long

-would need everyone together to do proposed method 

It’s a no for me dawg

[busstogate]

I should be admin.

[Scott Kelly]

Attached is a small sample of how this could work for a weight class.

 

I coded up in VBA how a single round can work.  Sorry I didn't leave a comments in the code yet.

 

This score method would be based off of the Glicko system similar to how Chess.com ratings work, I've input his equations he uses.

 

The idea would be your score would increase or decrease relative to several factors including:

 

• How good you are
• How good your opponent is
• How certain we are if your current score, using a variable called "Rating Deviation" (which you could think of as the standard deviation of your skill level as a wrestler..sort of)

In the workbook, in [sheet1], I've input a sample size of wrestlers with a hypothetical score and rating deviation, what rank they would be based on that score, and a column i'm calling rand for now but should clean that up later it doesn't matter.

 

Click enable macros at the top, then you click the button "create round" and it would run the macro to generate round one.  This will create sheet [Round 1], in which has automated the bout matching for round one (which is just random for now, but i can develop an algorithm that is intelligent and match you however we want to later).  Here, you can input 1 or 0 in the red column depending on if the wrestler won or lost, and you can see it changes their new score in column L.  If you win it goes up, if you win against a stronger opponent it goes way up.  Vice versa is true for losing, if you lose against a strong player it goes down a little, if you lose against a lot weaker your score goes down alot.  Their rating deviation should also change, the closer to zero the rating deviation is, the less of an impact each win or loss should have on your score because we are more certain of your score.  Dr. Glicko notes that we should be careful to have a threshold that RD should not drop below so that the match results can still meaningfully update your score, so i've got to adjust that still.  I plan on making modifiers for how great a margin of victory/loss you had (minor, major, techfall/fall) and maybe some other things.

 

Please have fun with this and suggest how the score method can be improved, i will continue to code up this workbook to show how we should be doing the state tournament for wrestling in Indiana.

SCOTT_KELLYS_IHSAA_PROPOSED_STATE_WRESTLING_TOURNAMENT.xlsm

[Silence Dogood]

 

11 minutes ago, Scott Kelly said:

Attached is a small sample of how this could work for a weight class.

 

I coded up in VBA how a single round can work.  Sorry I didn't leave a comments in the code yet.

 

This score method would be based off of the Glicko system similar to how Chess.com ratings work, I've input his equations he uses.

 

The idea would be your score would increase or decrease relative to several factors including:

 

• How good you are
• How good your opponent is
• How certain we are if your current score, using a variable called "Rating Deviation" (which you could think of as the standard deviation of your skill level as a wrestler..sort of)

In the workbook, in [sheet1], I've input a sample size of wrestlers with a hypothetical score and rating deviation, what rank they would be based on that score, and a column i'm calling rand for now but should clean that up later it doesn't matter.

 

Click enable macros at the top, then you click the button "create round" and it would run the macro to generate round one.  This will create sheet [Round 1], in which has automated the bout matching for round one (which is just random for now, but i can develop an algorithm that is intelligent and match you however we want to later).  Here, you can input 1 or 0 in the red column depending on if the wrestler won or lost, and you can see it changes their new score in column L.  If you win it goes up, if you win against a stronger opponent it goes way up.  Vice versa is true for losing, if you lose against a strong player it goes down a little, if you lose against a lot weaker your score goes down alot.  Their rating deviation should also change, the closer to zero the rating deviation is, the less of an impact each win or loss should have on your score because we are more certain of your score.  Dr. Glicko notes that we should be careful to have a threshold that RD should not drop below so that the match results can still meaningfully update your score, so i've got to adjust that still.  I plan on making modifiers for how great a margin of victory/loss you had (minor, major, techfall/fall) and maybe some other things.

 

Please have fun with this and suggest how the score method can be improved, i will continue to code up this workbook to show how we should be doing the state tournament for wrestling in Indiana.

SCOTT_KELLYS_IHSAA_PROPOSED_STATE_WRESTLING_TOURNAMENT.xlsm 35.41 kB · 4 downloads

If this is a troll, he's going all in.

 

[TysonNisley]

2 hours ago, Scott Kelly said:

Attached is a small sample of how this could work for a weight class.

 

I coded up in VBA how a single round can work.  Sorry I didn't leave a comments in the code yet.

 

This score method would be based off of the Glicko system similar to how Chess.com ratings work, I've input his equations he uses.

 

The idea would be your score would increase or decrease relative to several factors including:

 

• How good you are
• How good your opponent is
• How certain we are if your current score, using a variable called "Rating Deviation" (which you could think of as the standard deviation of your skill level as a wrestler..sort of)

In the workbook, in [sheet1], I've input a sample size of wrestlers with a hypothetical score and rating deviation, what rank they would be based on that score, and a column i'm calling rand for now but should clean that up later it doesn't matter.

 

Click enable macros at the top, then you click the button "create round" and it would run the macro to generate round one.  This will create sheet [Round 1], in which has automated the bout matching for round one (which is just random for now, but i can develop an algorithm that is intelligent and match you however we want to later).  Here, you can input 1 or 0 in the red column depending on if the wrestler won or lost, and you can see it changes their new score in column L.  If you win it goes up, if you win against a stronger opponent it goes way up.  Vice versa is true for losing, if you lose against a strong player it goes down a little, if you lose against a lot weaker your score goes down alot.  Their rating deviation should also change, the closer to zero the rating deviation is, the less of an impact each win or loss should have on your score because we are more certain of your score.  Dr. Glicko notes that we should be careful to have a threshold that RD should not drop below so that the match results can still meaningfully update your score, so i've got to adjust that still.  I plan on making modifiers for how great a margin of victory/loss you had (minor, major, techfall/fall) and maybe some other things.

 

Please have fun with this and suggest how the score method can be improved, i will continue to code up this workbook to show how we should be doing the state tournament for wrestling in Indiana.

SCOTT_KELLYS_IHSAA_PROPOSED_STATE_WRESTLING_TOURNAMENT.xlsm 35.41 kB · 10 downloads

Oh wait... he's actually serious...

[GenHeavyHandz]

On 3/2/2021 at 10:52 AM, aoberlin said:

Thinking about just deleting this post. But the forum has been pretty boring lately.

I’m sorry: I’ve been busy

[IMrule]

I really thought when I opened his proposal, that I would see a

[Scott Kelly]

Here i've made a version 2 to make this easier to understand.

 

In sheet1, type in cell J2 how many rounds you want to simulate, and it will simulate a whole tournament of rounds and output a summary of every wrestler's score round by round and a line graph showing their score round by round of the top 10 wrestlers by the ending round.

 

This can help you play with the RD and score parameters in the beginning in sheet1 to see if the changes in score are great enough when people win or lose, as well as observe what this could look like for a tournament.

 

So far this is purely based on win or lose in chess.  Next I plan on putting in some kind of modifiers for wrestling after this, such as the severity of the win/loss should impact the ending score for both wrestlers.  I want to develop intelligence of pairing up wrestlers by their scores and have it be dynamic from round to round so the match you wrestle in will impact who you wrestle next!  Enjoy!

 

Remember you have to enable macros, or it wont work when you click the simulate tournament button.  This is really addicting to work with, please have fun!

SCOTT_KELLYS_IHSAA_PROPOSED_STATE_WRESTLING_TOURNAMENT_V2.xlsm

[GenHeavyHandz]

Greco is the Sport of

the Gods.  High School doesn’t deserve it. 

[AdamsCoBuschhh]

On 3/2/2021 at 7:40 PM, Ch1StyleZ84 said:

Having been to few Sumo matches since I've been living in Japan I have to agree. Sumo is far and away more exciting than those other archaic wrestling styles. I mean, Folkstyle is one big snooze fest that should have been replaced by Freestlye long ago.

 

Having spent a few years in Japan in the Marines and watching amateur Sumo events.....sober or not they're just a tad bit more entertaining than Greco.  Even the smaller guys are as stout as a German tank.  

[Scott Kelly]

I got some more time to work on it tonight.

 

Okay i fixed the bug where it was really only configured to sort for 3 rounds in the summary sheet, now it will sort dynamically for as many rounds as you want.  fixed some other things after testing some more.

 

Added a control worksheet, and relocated the run simulation button there as well as the "how many rounds" cell you input for how many rounds of simulation we want to do (remember to turn macros on, or the button to simulate won't work).  Added an inventory worksheet, and that's where the sample inventory of wrestlers are in our hypothetical season with their scores and rating deviation now.

 

I added a "tier" parameter, in which each wrestler gets placed in a tier by their "score".  Then, i added intelligence to the matching maker in which it first should sort you by tier then by a random number to pair you up within your tier or the next tier above/below you.  I configured the tiers' bands using a target distribution I set up, I think I/we need to play around with this now and run a bunch of different tournament simulations to get the right setting for the rating deviation so the amount you appreciate/depreciate in "score" isn't too fast or too small over time.  I'm thinking right now probably need to set it so that a target distribution still gets met within some kind of tolerance after you run like 30 rounds, because that's about how many matches i think people have normally in a season.  Might be a while before i post again.

 

SCOTT_KELLYS_IHSAA_PROPOSED_STATE_WRESTLING_TOURNAMENT_V3.xlsm

Edited March 5, 2021 by Scott Kelly

[GrecoCoach]

Scott, you really need a new hobby.  You're too mathematically smart for 99% of the wrestling community.  Most parents couldn't figure out vertical pairing, and you're trying to incorporate chess rankings 

[Scott Kelly]

51 minutes ago, GrecoCoach said:

Scott, you really need a new hobby.  You're too mathematically smart for 99% of the wrestling community.  Most parents couldn't figure out vertical pairing, and you're trying to incorporate chess rankings 

I see where you’re coming from but hear me out, here’s some new things that are great about this approach:

 

We don’t have to know anything about math for this to totally work.  I’m not good at math.  All we need to care about:

 

-back in the 90s, some nerd built a system of equations in order to make it possible to rate everyone’s skill at something for a season (sure his idea was specifically for chess, but we can steal it) and make it sensibly rate your victories and losses.

 

-in a non bracket state tournament, now mediocre/casual wrestlers can have something to work toward:  building up their score to get to the next tier.  Maybe they can’t hope to get in the top 8, but they can try to battle up the next tier.  It’s like a video game now.  Beat someone rated high = big score boost.  How high?  Who cares, but it will become an element of fun to look it up at the end of a hard day on the mat.  Now, when we see those big upsets in a dual meet mid year (where a kid really does show he/she improved) its remembered by the system and it counts for that kid’s score for the season.

 

-in a non bracket tournament, we won’t just see where the top 8 stand.  We will know exactly what place you are by score.  You could shoot for top 50 your novice/freshman year if you’re new but kinda good, etc.

[M109R]

On 3/2/2021 at 12:20 PM, UncleJimmy said:

I feel like this is a @ClownBabyburner account for some reason. Sounds like him. Where is that dude? Where's a lot of dudes, come to think of it? Maybe an offseason thread.."Where have they gone.."

I want Caliboy back .

[TysonNisley]

3 hours ago, Scott Kelly said:

I see where you’re coming from but hear me out, here’s some new things that are great about this approach:

 

We don’t have to know anything about math for this to totally work.  I’m not good at math.  All we need to care about:

 

-back in the 90s, some nerd built a system of equations in order to make it possible to rate everyone’s skill at something for a season (sure his idea was specifically for chess, but we can steal it) and make it sensibly rate your victories and losses.

 

-in a non bracket state tournament, now mediocre/casual wrestlers can have something to work toward:  building up their score to get to the next tier.  Maybe they can’t hope to get in the top 8, but they can try to battle up the next tier.  It’s like a video game now.  Beat someone rated high = big score boost.  How high?  Who cares, but it will become an element of fun to look it up at the end of a hard day on the mat.  Now, when we see those big upsets in a dual meet mid year (where a kid really does show he/she improved) its remembered by the system and it counts for that kid’s score for the season.

 

-in a non bracket tournament, we won’t just see where the top 8 stand.  We will know exactly what place you are by score.  You could shoot for top 50 your novice/freshman year if you’re new but kinda good, etc.

What if the #1 kid in the state gets injured and can't build up his score for most of the season? This system wouldn't work cuz different schools are wrestling different people and it makes it even more unfair than our already somewhat unfair state tournament.

[Scott Kelly]

Here I think I got the calibrations all figured out with this one attached.

 

In the control worksheet i added a histogram and some summaries to help us see what happens to score overall and in the round.

 

I was able to get things calibrated pretty good i think by setting it so the score bands are dynamic and are reset every round.

 

This should help you test this out to see if the distributions and scores and everything work the way you think they should.

SCOTT_KELLYS_IHSAA_PROPOSED_STATE_WRESTLING_TOURNAMENT_V4.xlsm

[Scott Kelly]

2 hours ago, TysonNisley said:

What if the #1 kid in the state gets injured and can't build up his score for most of the season? This system wouldn't work cuz different schools are wrestling different people and it makes it even more unfair than our already somewhat unfair state tournament.

This is a great question.

 

im thinking 2 things:  1) what we should do is calibrate it to award bigger point swings at the actual state tournament.  I’m thinking in a Friday night to Sunday window we should have 8-10 rounds of wrestling matches for Everyone.  And we should calibrate it so a big enough swing could happen to take a lowest rank F to champ or state place winner in top 8 level if you go undefeated.  I think actually this might be possible already with how it’s setup but I’ll have to test it.  2) BUT-I don’t like placing TOO much emphasis on the final tournament.  This hypothetical injured best-in-the-end would not have been the best over-the-season and I think it’s more fair to award championship title to the best over the season.  I feel this way because I wrestled in d1 and I watched schools (including my own) rest out their best guys until “it mattered” often.  I think that’s a smart move on coaches part but it’s dumb that the system works that way to force us to want to sit people out.

Edited March 5, 2021 by Scott Kelly

[TysonNisley]

10 hours ago, Scott Kelly said:

This is a great question.

 

im thinking 2 things:  1) what we should do is calibrate it to award bigger point swings at the actual state tournament.  I’m thinking in a Friday night to Sunday window we should have 8-10 rounds of wrestling matches for Everyone.  And we should calibrate it so a big enough swing could happen to take a lowest rank F to champ or state place winner in top 8 level if you go undefeated.  I think actually this might be possible already with how it’s setup but I’ll have to test it.  2) BUT-I don’t like placing TOO much emphasis on the final tournament.  This hypothetical injured best-in-the-end would not have been the best over-the-season and I think it’s more fair to award championship title to the best over the season.  I feel this way because I wrestled in d1 and I watched schools (including my own) rest out their best guys until “it mattered” often.  I think that’s a smart move on coaches part but it’s dumb that the system works that way to force us to want to sit people out.

Yeah but some kids can't really help if they get injured, and I think that kids that actually can't wrestle their whole season (due to injury or other reasons) will end up not wrestling the season anyways.

[md66]

bump

[md66]

this  Kelly boy has smoked entirely to much crack. a mind is a terrible thing to waste. toooo late for him

[Cmoney125]

Greco-Roman? Lol no. Freestyle? 100%. The girls wrestle Freestyle in college. There's literally no reason for Folk style to be a thing anymore.