Proposed new HHF formula

[j33716]

Over the years it has become apparent that nobody (outside of the inner circle) knows how the USPSA 100% hit factor is calculated. My problem with this is the damage it does to USPSA by having a cloud of secrecy.

As a new BOD member, I asked questions. I now have a better understanding of the process. I am not happy with the answers I got.

Therefore, I have proposed a new HHF standard.

The 100% Hit Factor shall be the 10th highest submitted score for each respective classifier. Competitors achieving 100% or better shall be awarded 100% for the respective classifier. The 100% shall be calculated prior to any new classifiers being added to the data.

• Allows for anomalies over the top
• Gives a sliding scale that never goes down
• Is a formula the computer can calculate weekly
  
  
  Whatever standard is adopted should be published in the rule book!

Jay Corn

Area 6

[Shadow]

Jay-

I like it. My only concern is the HHF would be a little low when the classifier first comes out (after its shot at the nationals?). The difference between 1st and 10th for classifiers shot at last years nationals can be 20%. But it would quicly correct with your system.

[sperman]

I assume USPSA has several years of historic classifier data. Have you looked at how this system would align with the current HHF? Would a bunch go up? down?

I like the idea of a simple and transparent system, but without some more information, I'm not sure this is the correct formula. (It might be, but I would want to see some numbers to make sure.)

[h4444]

Simple, calculate it with the standard set for each class.

If GM is top 5 percent, the bottom score at that top 5 percent becomes hhf.

That should create a sliding scale that should align with the rest of the classification system.

Also, I know there is no way in the world this would happen, but for years I have said there should be downward mobility in the classification system as well. Sandbaggers are going to game the system anyway, but why not encourage some consistency in shooting?

Now, you can 'hero or zero' classifiers with no penalty, so there is no risk to trying to grandbag, completely blow a classifier,,,,,,who cares, it won't count against you.

The one way nature of classification systems in most of the shooting sports I play is something that has bothered me, but I just go on as I like shooting.

[kneelingatlas]

New classifiers are released at Nationals each year right? Why not just set the HHFs for each classifier with the top stage score in each division? That seems pretty simple to me.

The HHF for each classifier could then be set in stone at that point, or it could ratchet up with each new high score submitted into the system.

I would vote to set it in stone to eliminate the possibility of input error or 'grandbagging' on subsequent runs at lower level matches.

[steel1212]

New classifiers are released at Nationals each year right? Why not just set the HHFs for each classifier with the top stage score in each division? That seems pretty simple to me.

The HHF for each classifier could then be set in stone at that point, or it could ratchet up with each new high score submitted into the system.

I would vote to set it in stone to eliminate the possibility of input error or 'grandbagging' on subsequent runs at lower level matches.

Even at nationals some people will go for broke and Hero it making it a impossible classifier if you use the top one.

[JFlowers]

How does a changing HHF affect standing scores and classifications? Does my 61% suddenly become a 59% when the HHF is recalculated? Or do historical scores stay the same, even when they can no longer be calculated because of the change? If they recalculate, do people lose classification?

How do you calculate HHF for the smaller Divisions? Do you link Limited 10, Single Stack, Revolver to the HHF in Open as a percentage to start and then let if develop on its own? That has not been well received in the past..

Not saying the current system can not be improved or replaced with a better one, just that a lot of thought on the impacts of a change needs to happen.

[kneelingatlas]

Even at nationals some people will go for broke and Hero it making it a impossible classifier if you use the top one.

It is what it is. I guess you could average the scores from nationals and call it a RHF (reasonable hit factor)...

[kneelingatlas]

I don't understand this idea that a run can be 'too fast' to count. I do understand that if a shooter shoots the classifier stage over and over again until the hit a record setting time (as if shooters with the ability to do so wouldn't have anything better to do with their time) they would have an unfair advantage over the rest of us who see the stage, walk it and shoot it; furthermore I understand that if classifier scores shot at lesser matches were eligible to set the HHF questions could arise regarding conditions under which the score was shot or the integrity of the scoring.

If the HHF was set at nationals, everyone would know who set it, when and they could likely watch it on YouTube.

[sperman]

How does a changing HHF affect standing scores and classifications? Does my 61% suddenly become a 59% when the HHF is recalculated? Or do historical scores stay the same, even when they can no longer be calculated because of the change? If they recalculate, do people lose classification?

Historic data remains unchanged when the HHF changes.

I don't understand this idea that a run can be 'too fast' to count. I do understand that if a shooter shoots the classifier stage over and over again until the hit a record setting time (as if shooters with the ability to do so wouldn't have anything better to do with their time) they would have an unfair advantage over the rest of us who see the stage, walk it and shoot it; furthermore I understand that if classifier scores shot at lesser matches were eligible to set the HHF questions could arise regarding conditions under which the score was shot or the integrity of the scoring.

If the HHF was set at nationals, everyone would know who set it, when and they could likely watch it on YouTube.

Let's say a GM has an exceptional run on a given classifier. He just goes out of his mind and sets a new standard. For the next 6 months, GM's shooting that classifier are scoring in the high 80%, M's in the high 70%, etc. Almost anyone who shoots that classifier is going to be more than 5% below their classification, making it not a good representation of a shooter's skills. If the classifier system is going to work the way it was designed, the HHF's need to be set so that a shooter's run is representative of their skill.

ETA, I think we have a few classifiers like that in the system already, but we definitely want to try and limit creating any more.

[ZackJones]

Simple, calculate it with the standard set for each class.

If GM is top 5 percent, the bottom score at that top 5 percent becomes hhf.

That should create a sliding scale that should align with the rest of the classification system.

I like this idea. It would allow the HHF to move up or down as more scores are reported. I think it could handle the situation described above where a GM shoots it in Jedi Hero Mode and in the following months others shoot a much lower HF.

[kneelingatlas]

How does a changing HHF affect standing scores and classifications? Does my 61% suddenly become a 59% when the HHF is recalculated? Or do historical scores stay the same, even when they can no longer be calculated because of the change? If they recalculate, do people lose classification?

Historic data remains unchanged when the HHF changes.

I don't understand this idea that a run can be 'too fast' to count. I do understand that if a shooter shoots the classifier stage over and over again until the hit a record setting time (as if shooters with the ability to do so wouldn't have anything better to do with their time) they would have an unfair advantage over the rest of us who see the stage, walk it and shoot it; furthermore I understand that if classifier scores shot at lesser matches were eligible to set the HHF questions could arise regarding conditions under which the score was shot or the integrity of the scoring.

If the HHF was set at nationals, everyone would know who set it, when and they could likely watch it on YouTube.

Let's say a GM has an exceptional run on a given classifier. He just goes out of his mind and sets a new standard. For the next 6 months, GM's shooting that classifier are scoring in the high 80%, M's in the high 70%, etc. Almost anyone who shoots that classifier is going to be more than 5% below their classification, making it not a good representation of a shooter's skills. If the classifier system is going to work the way it was designed, the HHF's need to be set so that a shooter's run is representative of their skill.

ETA, I think we have a few classifiers like that in the system already, but we definitely want to try and limit creating any more.

Then don't call it the High Hit Factor, call it the "reasonable hit factor for a GM; within 5%".

It doesn't really matter to me what objective, verifiable, repeatable methodology is in place, just that there is one.

If this "Super GM" is a real problem, then just average the scores shot by all the GMs on that stage at Nationals and make that the 100% score.

[Chills1994]

Over the years it has become apparent that nobody (outside of the inner circle) knows how the USPSA 100% hit factor is calculated. My problem with this is the damage it does to USPSA by having a cloud of secrecy.

As a new BOD member, I asked questions. I now have a better understanding of the process. I am not happy with the answers I got.

Therefore, I have proposed a new HHF standard.

The 100% Hit Factor shall be the 10th highest submitted score for each respective classifier. Competitors achieving 100% or better shall be awarded 100% for the respective classifier. The 100% shall be calculated prior to any new classifiers being added to the data.

• Allows for anomalies over the top
• Gives a sliding scale that never goes down
• Is a formula the computer can calculate weekly

  Whatever standard is adopted should be published in the rule book!

Jay Corn

Area 6

Thanks Jay! I will be back later, I think, to write more.

[Ron Ankeny]

Statistics aren't rocket science. I too think outliers need to be eliminated when figuring the HHF and I would assume that is already happening. What would happen under your proposal when there are a dozen or more outliers that are artificially high due to the hero or zero mentality? Would the 10th outlier (which should be eliminated) become the HHF?

[Nik Habicht]

Over the years it has become apparent that nobody (outside of the inner circle) knows how the USPSA 100% hit factor is calculated. My problem with this is the damage it does to USPSA by having a cloud of secrecy.

As a new BOD member, I asked questions. I now have a better understanding of the process. I am not happy with the answers I got.

Therefore, I have proposed a new HHF standard.

The 100% Hit Factor shall be the 10th highest submitted score for each respective classifier. Competitors achieving 100% or better shall be awarded 100% for the respective classifier. The 100% shall be calculated prior to any new classifiers being added to the data.

• Allows for anomalies over the top
• Gives a sliding scale that never goes down
• Is a formula the computer can calculate weekly

  Whatever standard is adopted should be published in the rule book!

Jay Corn

Area 6

And that part gives me significant pause -- as we've already lost classifiers that were not set correctly, or allowed to be gamed. When that later came to the attention of DNROI, after the results had already corrupted the database, the classifiers were inactivated......

There's going to need to be a way for classifier HHFs to go down, if an improperly set classifier drives the HHF too high.....

....or if a series of improperly set classifiers slowly edge it up.....

[kneelingatlas]

I agree with you, which is why I'm in favor of setting the 100% score at Nationals, under official conditions, with each compeditor seeing the stage for the first time. It might be fun to keep track of the highest score ever shot on the classifier, but all classifications should be calculated off the same 100% score for any given classifier stage.

[sperman]

What happens if the classifier isn't shot by every division at Nationals, or like several of the 13 series classifiers, the published classifier doesn't match how it was shot at Nationals?

[kneelingatlas]

I guess I didn't think about the fact there are different nationals for different divisions...

[steel1212]

Even at nationals some people will go for broke and Hero it making it a impossible classifier if you use the top one.

It is what it is. I guess you could average the scores from nationals and call it a RHF (reasonable hit factor)...

That would be more realistic. I just know that it's happened and is being used and is damn near impossible lol.

[j33716]

When a classifier becomes a classifier has nothing to do with the classifier 100% level

A stage is not a classifier until it has been shot enough to be considered for a classifier.

It is a different rule and requirement.

Edited January 20, 2014 by j33716

[Chills1994]

If you could assume that the hit factors for each classifier stage followed a normal distribution:

You could make whatever the mean/median hit factor the dividing line between the B's and the A's. The first stand deviation above the mean/median would be all the A class shooters.

Going down to one stand dev from the mean/median would be all the B class shooters.

That group of shooters with HF's between +1 and +2 standard deviations would be the Masters. If you're higher than +2 stand devs, then you're a GM.

For those between -2 and -1 stand devs, they would be C class.

Farther out than -2 stand devs, you're D class.

The great thing statistically speaking with using the median, and to a lesser extent the mean is that you lessen or blunt the impact of the Hero-go-for-broke classifier hit factors...

Kinda like how using the phrase "the median household US income is ________ per year" is more accurate

because the average or mean gets really wonky with the few Oprah Winfrey, Bill Gates, and Warren Buffet incomes.

Probably at some point, I did know how to convert a whatever score or say a + 1.83 standard deviation placement into a percentile or percentage. And that's how you would add the percentilesor percentages from several classifiers to get a shooter's average, and classify him appropriately.

Edited January 20, 2014 by Chills1994

[beltjones]

Jay,

I like that you are taking a fresh look at HHF's, but I would suggest that - based on your proposal - you don't have the statistical background to really fix the system. Your "bottom of the top 10 scores" idea is pretty bad from a statistical standpoint.

It would be better to do as Chills1994 suggests and use a normalized distribution to calculate the frequency at which a certain score is shot, then use that to classify people based on percentiles as defined in the rule book (ie, GM's at 95% and so on).

[MarkCO]

^ I agree. The 10th place setting the HHF does not sound like a good idea.

Actually, I don't have a problems with HHF being set by the high score at a Nationals, or an average of the top 5, and then locked for a year with some sort of "certification". The certifications would be in place for those times when there is a problem, like it was not set up right, or the super-squad shot in a downpour or dustdevil. If a classifier HHF is not re-set in 3 years by again being shot at a Nationals or Area match with at least 10 GMs, it goes bye-bye. If the HQ staff wants to keep a classifier on the books, it has to be shot at an Area or Nationals level match and the HHF can then be re-set. The skill level has jumped a huge amount in the last 10 years, and we still have 15 year old classifiers active. No-one, except a chosen few, understand if that is even an issue.

I've just seen WAY too many improperly scored, and set-up classifiers, and people re-shooting 6 times until they get the score they want, to make the current system worth much. You can't prevent sand/grand baggers, but you can maintain integrity of the HHF by only allowing it to be set at majors.

[ZackJones]

Here's a novel thought: Let's do away with classifications. Yeah I know it sounds crazy but think about it for a couple of minutes.

At a match we only compete within our declared division so the shooter that shoots the best match within the division is going to win that division. It doesn't matter if that shooter is classified on paper as GM, M, A, B, C, D, or U. The best shooter on that given day is going to win.

[MarkCO]

Once I met a "classification" goal, then they started to mean less to me, however, the system is still a good measuring stick and provides those who only compete at a local level a more realistic determination of their shooting skill against the entire membership. It did provided some motivation to me as a shooter, artificial as it may be. But for a hobby, it is all artificial anyway. Doing away with the Classification system would be folly, and crazy.