Ratings "propagators"

From: http://pdga.com/competition/ratings/CalculatingYourRating.pdf

"All we need are 10 players with ratings over 799 to be playing the same course to determine your rating for that round. It's not just any player with a rating over 799, they have to be players whose rating is based on at least 8 rounds of information. We call these players "propagators."

So, how are these people chosen? At some events, there may be 30 players that qualify. Do you choose the best 10 scores of the round? The worst? middle? a mix?

If I shoot 45 and am chosen as a propagator my rating will be lower than if you picked a guy that shot 50.

Perhaps I am missing a critical step in the process.

[jconnell]

I believe 10 is the minimum, and that if an event has 30 like your example, then they'd use all those that qualify as "propagators".

The more "true" ratings (propagators) used to figure an event's ratings, the more "true" that rating will be.

I'm sure that someone official from the Ratings committee will clarify/correct me if I'm wrong. They know better than I.

--Josh

[ck34]

The critical piece of information is the words "at least" need to be in that statement. We need at least 10 props playing a round for the ratings to be calculated. All props scores in that round are looked at. However, anyone whose round rating is more than 50 rating points below their current rating is excluded as a propagator for that round.

However, anyone whose round rating is more than 50 rating points below their current rating is excluded as a propagator for that round.



But, those whose round rating is higher are accepted, thus reducing that rounds affect on their rating by skewing the rounds WCP?

[brianberman]

You make a good point Brian. All individual ratings information in our sport is too high. I think that if we are going to knock 10% off the bottom we should do the same to the top 10%. This goes for the WCP figuire for each round as well. If they drop them b/c they are too low
(anyone whose round rating is more than 50 rating points below their current rating is excluded as a propagator for that round. ) then they should drop them if they are too high as well. This would drop all of our ratings and make the divisional chart make more sense as well.

[brianberman]

Them maybe I could play intermediate and not feel like an *****.
:D
Just kidding I can never go back.

You make a good point Brian. All individual ratings information in our sport is too high. I think that if we are going to knock 10% off the bottom we should do the same to the top 10%. This goes for the WCP figuire for each round as well. If they drop them b/c they are too low
(anyone whose round rating is more than 50 rating points below their current rating is excluded as a propagator for that round. ) then they should drop them if they are too high as well. This would drop all of our ratings and make the divisional chart make more sense as well.



If someone plays 50 points above their rating and they are used as a propagator then everyone else's rating would go down for that round in relationship to that person--unless, of course someone played even more above their rating. When someone plays below their rating it bumps other ratings higher.

I think you are confusing being a propagator and whether a particular round is dropped from your average rated rounds (which is your rating).

On a personal level, your worst rated rounds are dropped once you have the minimum in the data base. The rounds that are kept are averaged to make your rating.

If you are a propagator and you play 50 points or more below your rating, then you aren't used to propagate ratings for that round. You will, however still get rated for that round. That rating may then get dropped if it is one of your lower rated rounds, but that is something else entirely.

If they drop them b/c they are too low
(anyone whose round rating is more than 50 rating points below their current rating is excluded as a propagator for that round. ) then they should drop them if they are too high as well. This would drop all of our ratings and make the divisional chart make more sense as well.



Yeah, I think you are confused. By not dropping the top scores to determine the WCP, it hurts your ratings. If your smoking round of 45 is used and you have a 900 rating, in theory that means that a 900 rating player can shoot a 45. So WCP is set at 35 which we all know is BS.

At least, this is how I am reading this. Chuck, care to give us an example that uses numbers to make it clearer?

[gnduke]

It does mean the ratings can be skewed if 1/3 of the gators play 50 points above thier ratings, 1/3 play within 50 points of thier ratings and 1/3 play 50 points below thier ratings, then the ratings will be lowered for the entire field.

If however, 1/10 of the field shoots 50 points above and 1/5 shoot below, then the calculations will be closer to correct.

[brianberman]

I am confused b/c this is a confusing subject. I just think that in order to smooth the bell curve, in individual rounds and in ratings, that we should slice off the fat from both ends.

If your course's SSA is 54 on record and there is a tournament, if several gators shoot above their rating and several gators shoot below their average but the majority are around their rating, if we take the majority who are playing at their rating to figure out the WCP then we get a more honest number vs. only taking out the "belows".

[brianberman]

This will make the ratings more honest and it will be harder to approach 1000 or higher.

Some of the "Gods" excluded of course

If we clip the top and bottom off to get individual ratings all numbers will be honest.

ha

I just think that in order to smooth the bell curve, in individual rounds and in ratings, that we should slice off the fat from both ends.



That is my point. If you are going to ditch the scores of the guys that played like crap, you need to ditch the scores of the guys that played out of their mind.

I am confused b/c this is a confusing subject. I just think that in order to smooth the bell curve, in individual rounds and in ratings, that we should slice off the fat from both ends.

If your course's SSA is 54 on record and there is a tournament, if several gators shoot above their rating and several gators shoot below their average but the majority are around their rating, if we take the majority who are playing at their rating to figure out the WCP then we get a more honest number vs. only taking out the "belows".



How many tournaments have propagators shot more than 50 points above AND other propagators shot more than 50 points below their ratings? If you were to throw out the top, how much would that change everyone's rating? I think you'd have to answer those two questions before anyone would consider making the changes. It would be a lot of work to figure that out. Good luck! ;)

[bruce_brakel]

I like that. I was a gator both rounds on Sunday. I even let a little bird clean my teeth!

[spartan]

sorry. i saw GAYTERDS! and thought it was a thread about something else. ;)

GO VOLS!!

[widiscgolf]

Sweet Tournament guys at Crystal Lake!! A few of us from Wisconsin came down and had a blast! Great course, great payout, and great tourney!

So when you getting those scores posted to the PDGA? Hehe

[ck34]

The tails of the distribution curve are not equally probable. It's much easier to shoot a score that's more than 50 points below your rating than more than 50 above. I can easily decide to play risky and go for ace runs and "magic" shots 350 over OB when I'm out of it. Not the opposite though. Wouldn't you just love to say, "OK, I'm sick of how I'm playing. I'm just going to shoot a round 100 points above my rating today," and actually do it?

It's possible 50 points isn't far enough down and it should maybe be 60 or 70. We'll look at it and see. We do look closely at large deviation rounds, both high and low, to see if there are obvious errors in event processing. But that's a manual review that could probably be more automated with a little code work.

[james_mccaine]

Spurrier envy.

It seems logical that +50 people should be excluded also. They are in effect the "curve busters" that lower all the other ratings. These people are probably quite often players whose rating is lagging way behind their ability, so they make lousy gators anyway. However, I suspect that the ratings folks have considered all this and have good reasons to include these folks.

[cevalkyrie]

Jon is usually very prompt with getting scores reported. He busted him hump to get the Streamwood Starter scores turned in before the ratings update. He got them in but they were not included. Give him a week or so to get them updated.

Thanks for coming! We've enjoyed the WI support for the first two events. Please pass the word.

[widiscgolf]

I was just giving him crap. That's all cool. Again great tourney!

Sweet Tournament guys at Crystal Lake!! A few of us from Wisconsin came down and had a blast! Great course, great payout, and great tourney!

So when you getting those scores posted to the PDGA? Hehe



Thanks Josh. I responded here (http://www.pdga.com/msgboard/showflat.php?Cat=&Number=216057&page=0&view=&sb=5&o=&fpart=&vc=1) to correct the thread drift.

[brianberman]

So in SQL you run the current ratings to give you numbers to work with, create a query that gives you a differential number that can be used to remove the top 5% and bottom 15% of the gators. Then take 80% remaining that are not flagged as excluded to calculate wcp for the round and re-run the original scores for ratings. I think this will take the fluke rounds out of the equation and give us a truer center.

If we did the same to removing part of an individuals top rounds in the rating system I think that this could possibly really give us an honest rating.

[hitec100]

What's the deal with your sig line? Take it down a notch, will ya? There's no need for you to be insulting to half the country who has a different point of view from you.

[girlie]

I find it to be right on track /msgboard/images/graemlins/smirk.gif

[sandalman]

What's the deal with your sig line? Take it down a notch, will ya? There's no need for you to be insulting to half the country who has a different point of view from you.

actually its slightly less than half the country... besides, juck em if they cant take a foke. :eek:

[neonnoodle]

What's the deal with your sig line? Take it down a notch, will ya? There's no need for you to be insulting to half the country who has a different point of view from you.

actually its slightly less than half the country... besides, juck em if they cant take a foke. :eek:



Actually it is way less than half the country. It isn't even have the registered voters.

What is the "White dude" thing in reference to though?

The best symbol for Republicans would be a pile of **** in a bed. That or a blindfold and money symbol.

Not this again...and the election is still over 3 months away!

[neonnoodle]

Yeah! That is the Bush campaign slogan:

"Don't Talk About It"

You're right though, let's keep this on the Politics thread.

[hitec100]

The best symbol for Republicans would be a pile of **** in a bed. That or a blindfold and money symbol.



Nick, you can't be ignored. So how about setting a good example instead of the one above?

[neonnoodle]

I don't know why the board word filter blanked out "bags". :D

I'll see what I can do about removing my ignore protection for you Playa Hatas out there...

[neonnoodle]

Ignore away!

Although that offer is tempting, I keep my "never have used the ignore feature" status intact.

Just please move your polictical bullbags somewhere else. I was actually reading ratings info here.

[neonnoodle]

Did Mark say something... :D

[gnduke]

Sounded like he was posting something against thread drift.

Must be an imposter posting under his ID. :cool:

Oh, my sides hurt from laughing. Y'all are freak Hi-larious.

:mad:

[neonnoodle]

Wha...? /msgboard/images/graemlins/smirk.gif

[brianberman]

Thanks Linds

Sometimes the truth hurts?