Which regional is better?

[Westforkwhite]

1,2,3 and 4 have to  wrestle 5-8 to get to SS.   Regional points are given only for qualifying to SS.      So If A has a measured tougher 5-8 than Regional B,  then its more difficult to get to semi-state.   That's what they're measuring,  they're not measuring how tough 1234 are.      The top 4 kids demonstrate higher skill level by making the ticket round or by qualifying for state, and accordingly score their points then.   Is it that complicated?

 

The system is attempting to evaluate how difficult it is to advance in that given regional. Only 4 wrestlers advance to SS, so it's measuring how difficult it is to be one of those 4. The quality of 5-8 plays a factor for sure, but you would ostensibly have to beat someone from that top 4 to advance. A & B are in the same SS. So to then evaluate the performance of the 3/4 finishers in 1st round of SS is a fair way to compare quality of the 3/4s for A&B, and in turn evaluate the difficulty to advance.

 

Another way to look at is, if you swap A1-4 with B1-4, how would those outcomes look. The statistics above reflect that the we would see a similar number advance from either, suggesting the difficulty to advance is comparable in both A & B.

 

Or finally: I take the 5th from A and put him in B, despite having lesser 6-8 wrestlers in B, A5 has 2/3 greater chance of running into a state qualifier first round of regional B. So wouldn't be easier to advance out of A? The same would hold true for all 5-8 finishers in A.

[GrecoCoach]

Semistate results happen weeks after the regional categories have been finalized. They can't be used as you keep arguing.

 

You also can't objectively evaluate them on hypothetical matches that don't happen with 5th place finishers and swapping regionals. Again, you are judging based on perceived notions of a regional's strength or weakness.

 

I doubt I'll change your mind. Until the IHSAA takes it back over with an all-in format, this is what we have. And it does very well at predicting the top teams for the following year.

[awood2]

As is true for every tournament - hence the meaning of tournament!  There are only 14 medals the IHSAA (or any of us) care about...  The 14 hanging from a blue ribbon!  If you're not first...

[Westforkwhite]

Semistate results happen weeks after the regional categories have been finalized. They can't be used as you keep arguing.

You also can't objectively evaluate them on hypothetical matches that don't happen with 5th place finishers and swapping regionals. Again, you are judging based on perceived notions of a regional's strength or weakness.

I doubt I'll change your mind. Until the IHSAA takes it back over with an all-in format, this is what we have. And it does very well at predicting the top teams for the following year.

You are not following my point. I'm not saying the SS data I have provided is a part of the input criteria, but rather a means in which to judge the output. I think most would agree that the 1st round SS results is a fair and objective way to evaluate the 1-4 finishers from respective regionals in the same SS. Obviously you cant evaluate wrestlers 5-8 at SS, but you can easily see they are the finishing group the output is skewed towards. (I'm not implying the input has been intentionally weighted towards 5-8, this has just been a result in this specific instance)

 

The examples deal with simple probability, not perceived notions.

 

I understand this is a small part of the total process, and this could be a statistical anomaly. I think the process as a whole is very good, but even good processes needed unbiased metrics in which to evaluate the output (output being the values assigned to regionals for advancement scoring purposes)

[Wrestling Scholar]

You are not following my point. I'm not saying the SS data I have provided is a part of the input criteria, but rather a means in which to judge the output. I think most would agree that the 1st round SS results is a fair and objective way to evaluate the 1-4 finishers from respective regionals in the same SS. Obviously you cant evaluate wrestlers 5-8 at SS, but you can easily see they are the finishing group the output is skewed towards. (I'm not implying the input has been intentionally weighted towards 5-8, this has just been a result in this specific instance)

 

The examples deal with simple probability, not perceived notions.

 

I understand this is a small part of the total process, and this could be a statistical anomaly. I think the process as a whole is very good, but even good processes needed unbiased metrics in which to evaluate the output (output being the values assigned to regionals for advancement scoring purposes)

1st I don't think you've demonstrated simple probability.   You're argument is based on the perceived notion that if Regional B does better in the 1st round of SS then they must have wrestled at a better level to get there.    I think you're missing the whole objective  of the process.    Adrian has took the past history of the season using his genius ratings to measure the difficulty of qualifying to the next level and is evaluating the total depth of the 8 wrestlers in the regional.    You're point is looking at future results of the top one or two wrestlers at the next level, then retroactively evaluating the regional based on the results.   

[warsawwrestling]

So you are saying Rochester Regional is best. See how easy that was!

[Wrestling Scholar]

So you are saying Rochester Regional is best. See how easy that was!

Actually,  he's saying the Jay county regional is a lot better than Carroll Regional even though Jay County scored a 3   and Carroll county got a 4 score.  In other words, he's   saying   the Carroll Regional is like the Logansport Regional of the Fort Wayne semi-state with one state placer like Logansport.

Edited March 14, 2017 by Wrestling Scholar

[Westforkwhite]

Actually,  he's saying the Jay county regional is a lot better than Carroll Regional even though Jay County scored a 3   and Carroll county got a 4 score.  In other words, he's   saying   the Carroll Regional is like the Logansport Regional of the Fort Wayne semi-state with one state placer like Logansport.

LOL. You are correct those were the two I was comparing but I think another regional in the FW SS has the "Logansport" title on lockdown.

 

The second example I provided deals strictly with the probability of running into what we can all agree (based on objective metrics) are better 1-2 seeds. Going further if the 3-4s are of comparable quality (again based on objective metrics) it is clear that in this specific instance the totality of the input data is valuing the quality of 5-8 place finishers at a higher level.

 

If you don't believe looking at the 1st round of SS results is a fair and objective way to evaluate the output, what do you suggest would be? There has to be an objective means to analyze the accuracy of the output, and I would argue that performance data outside of the input criteria, provides the best vehicle to do just that.

[Flash Attakk]

Ok. I have tried to avoid statistical, mathematical and analytical conversations on these boards because it feels like work. But, the category system does not allow for comparing regional toughness. It is a system to predict where the 64 "best" wrestlers likely participated in a regional. It does not say it is tougher to make it to semi state; it says the 4/5/6 best kid here is likely a top 64 kid or not. Trying to use it in a vacuum to compare one region to another doesn't work, using it to compare the strength of teams one year away works well.

[Westforkwhite]

Ok. I have tried to avoid statistical, mathematical and analytical conversations on these boards because it feels like work. But, the category system does not allow for comparing regional toughness. It is a system to predict where the 64 "best" wrestlers likely participated in a regional. It does not say it is tougher to make it to semi state; it says the 4/5/6 best kid here is likely a top 64 kid or not. Trying to use it in a vacuum to compare one region to another doesn't work, using it to compare the strength of teams one year away works well.

I get what you are saying, but how then do you evaluate the accuracy of the system (as it relates to regional values) after the fact? Or do we just assume it's placing the appropriate values for advancement on a given regional and not worry about ever attempting to gauge accuracy after the fact?

[MOWrestler]

LOL. You are correct those were the two I was comparing but I think another regional in the FW SS has the "Logansport" title on lockdown.

So is "Logansport Regional" an officially accepted term.  And how does IndianaMudflap feel about this?

[Wrestling Scholar]

So is "Logansport Regional" an officially accepted term.  And how does IndianaMudflap feel about this?

I'm just glad to hear that we don't have to be the only kid sitting in the last row of the class room anymore.  We have some company.

[Westforkwhite]

I obviously haven't done a good job making my point. I understand that I'm talking about a small piece of the whole puzzle, and viewing things in isolation (without context of goal) can sometimes result in a distorted view. The process as a whole is designed to pick the best dual teams for the IHSWCA state duals, and I think it's doing a very good job. That said I still feel there is value in investigating the accuracy of the results the system is giving us, and there are only so many ways in which to accomplish this. Looking at the 1-4 finishers doesn't give us a look at the whole regional field but it best evaluates the ones who advance. If B is better by every objective metric in places 1-4, then it stands to reason that the aspect of the system (all various elements combined) designating regional values is placing more emphasis on the 5-8 finishers. Could this be an anomaly, absolutely.

[warsawwrestling]

Statistics and sarcasm nothing more perfect for the day of Pi

[jmills]

What would be the result if only the top 3 from regional moved on to SS as in years past?

Should we go back to top 3?

Does taking top 4 make it exceptionally easier to go to SS?

 

Also

Has moving from 103 to 106 lowered the number of forfeits? 

[UncleJimmy]

I obviously haven't done a good job making my point. I understand that I'm talking about a small piece of the whole puzzle, and viewing things in isolation (without context of goal) can sometimes result in a distorted view. The process as a whole is designed to pick the best dual teams for the IHSWCA state duals, and I think it's doing a very good job. That said I still feel there is value in investigating the accuracy of the results the system is giving us, and there are only so many ways in which to accomplish this. Looking at the 1-4 finishers doesn't give us a look at the whole regional field but it best evaluates the ones who advance. If B is better by every objective metric in places 1-4, then it stands to reason that the aspect of the system (all various elements combined) designating regional values is placing more emphasis on the 5-8 finishers. Could this be an anomaly, absolutely.

 

That would be an understatement....keep trying tho! Almost there...

[Westforkwhite]

That would be an understatement....keep trying tho! Almost there...

Sometimes you can lead a horse to water....

[warsawwrestling]

Wrestlers and statistics go together like oil and sweat!

[Wrestling Scholar]

What would be the result if only the top 3 from regional moved on to SS as in years past?

Should we go back to top 3?

Does taking top 4 make it exceptionally easier to go to SS?

 

 

I would say yes. It is exceptionally easier to qualify for SS when you take 4 instead of 3.

[MOWrestler]

I would say yes. It is exceptionally easier to qualify for SS when you take 4 instead of 3.

I think technically it is twice as easy (at least without wrestlebacks).

 

Take top 3 of 8 means Win,Loss, Win gets you to 3rd. Take top 4 means Win, Loss, Loss gets you in.  So you have to win twice as many matches to get in IF you have no wrestle-backs AND only take 3.

 

Not sure why this  matters...

[Westforkwhite]

What would be the result if only the top 3 from regional moved on to SS as in years past?

Should we go back to top 3?

Does taking top 4 make it exceptionally easier to go to SS?

 

Also

Has moving from 103 to 106 lowered the number of forfeits?

We are seeing very low success rates for the 4th place finishers, so it suggests we'd have most all the kids with realistic SS advancement potential in the field.

Should we go back to Top 3? Yes, but only if there were full wrestlebacks at Reg.

Yes, as it appears there is a substantially dropoff (on average) from the 3 to 4 spot (at least in terms of SS performance)

 

103 to 106 - Very good question, unfortunately I don't know the answer, but I think it would also be interesting to see how removing a middle weight has impacted forfeits.

[MOWrestler]

We are seeing very low success rates for the 4th place finishers, so it suggests we'd have most all the kids with realistic SS advancement potential in the field.

Should we go back to Top 3? Yes, but only if there were full wrestlebacks at Reg.

Yes, as it appears there is a substantially dropoff (on average) from the 3 to 4 spot (at least in terms of SS performance)

Would you then also make true wrestlebacks @ the SS level?

[Westforkwhite]

Would you then also make true wrestlebacks @ the SS level?

Absolutely, I was only limiting within the frame of his question. I think wrestle backs at every stage would be great, but if I had to choose just one it would be hard to pick between state and semi state.

[maligned]

I get what you are saying, but how then do you evaluate the accuracy of the system after the fact?

The one and only metric that matters in evaluating the system as a whole is accuracy in getting the best teams from the following season into the event. Again, because we  are giving huge points for semi-state second round and state advancement AT SEMI-STATE, we are strictly and only concerned with assigning those points for advancing to the round of 64 (first round of semi-state) correctly at regional. For this, we have to look at each site's comprehensive crowdedness in the Round of 256 (essentially sectional) and the Round of 128 (first round of regional). Once we've assigned those Round of 64 points correctly, we move on and give big points for the Round of 32 (semi-state 2nd round) and the Round of 16 (state qualification) the FOLLOWING WEEK. 

 

Here are the metrics we care about in evaluating how well we've assigned points throughout the system from sectional non-placer through state champion:

 

2016-17 Season-Ending Duals Top 10s (one year after the scoring system is used):

 

3A

1. Brownsburg-invited

2. Perry-invited

3. Avon-invited

4. Portage-invited

5. EMD-invited

6. Penn-invited

7. Cathedral-lost at vote-in, but the system had them 11th

8. Castle-invited

9. Carroll-invited

10. Carmel-invited

(if selection committee had "guessed" correctly, all 10 of the season-ending top 10 would have been at the event. They took Chesterton and Columbus East instead of Cathedral--hard to blame them although we see it was a slight miss.)

 

2A

1. Yorktown-invited

2. Bellmont-invited

3. North Montgomery-invited

4. Edgewood-invited

5. Evansville Memorial-invited

6. Jimtown-invited

7. Garrett-invited

8. Indian Creek-lost at vote-in; system had them 14th despite 4 FFs at sectional

9. Southridge-lost at vote-in; system had them 15th

10. Culver Academies-invited

(Committee took Delta ahead of IC and Southridge; system "error" was #9 Northwood, who lost a returning state qualifier to a move among other attrition.)

 

1A

1. Prairie Hts-invited

2. Shenandoah-invited

3. Oak Hill-invited

4. Monrovia-invited

5. North Posey-invited

6. Eastern-lost at vote-in; system had them 13th despite FOUR sectional forfeits

7. Adams Central-invited

8. Cass-system had them 20th, with 2 forfeits and returning only 9 (nearly impossible they would qualify because 1A teams with that profile are never top 10 teams the next year)

9. Knightstown-invited

10. Central Noble-invited

(By far the most unpredictable class, but the 2 years of the current Category system have been very accurate even in this class.)

 

Totals: 15 of 15 from the year-later, season-ending Top 5s and 9, 8, and 8 of the Top 10s were invited. Only #8 from 1A, an extremely unpredictable class, was a genuine scoring system miss with no chance to be critiqued deeply and voted in. 

 

I've mentioned this before, but we use a mathematical model that's able to "control" completely for sectional/regional path (i.e. take it out of the picture) and create predictions for the number of regional qualifiers and semi-state qualifiers that teams of any given duals ability should have. We use FOUR SEASONS and 1,225 full season results simultaneously to run this. The fact that deserving teams are being given regional points for 5th or 6th place guys at some sectionals and undeserving teams are not being given semi-state first round points for a regional 4th placer, for example, at some sites is strategically instrumental into why we're able to produce the effective results I've shown above. The continued argument that "who got the most to the Round of 32 or the Round of 16 should be our predictor of which area's Round of 256 and Round of 128 is richer" argument is unfortunately very limited. We've been doing deep research into ways to control for sectional/regional path, including many of the ways people have suggested on this site, for 7 years. We've run lots of past-year tests for comparison and nothing works nearly as intuitively as what we're doing right now for getting the right group of teams into the event.

Edited March 15, 2017 by maligned

[Wrestling Scholar]

The one and only metric that matters in evaluating the system as a whole is accuracy in getting the best teams from the following season into the event. Again, because we  are giving huge points for semi-state second round and state advancement AT SEMI-STATE, we are strictly and only concerned with assigning those points for advancing to the round of 64 (first round of semi-state) correctly at regional. For this, we have to look at each site's comprehensive crowdedness in the Round of 256 (essentially sectional) and the Round of 128 (first round of regional). Once we've assigned those Round of 64 points correctly, we move on and give big points for the Round of 32 (semi-state 2nd round) and the Round of 16 (state qualification) the FOLLOWING WEEK. 

I don't think WestForkwhite gets this point.  But this is the essential point.   His output argument that you can measure the toughness on the regional based on how only the top one or two guys do later in at SS and State is weak.   You're measuring the totality of the top 8 in Regional,   he wants to base his measurement on the success of the top one or maybe two guys.   I guess good to debate, but I think the jury is back and WestForkwhite just lost.