In the ratings guide it informs me that "Almost all of your rounds are counted, but those more than 2.5 standard deviations below your average are dropped (about 1 in 50)."

My question is, what is this standard deviation value? My rating is 1005 and one of the rounds that was counted was a 928. How is this still counted?

Standard deviation s is calculated by:
http://www.tripledisc.com/preview/msdgc/sd1.jpg
&
http://www.tripledisc.com/preview/msdgc/sd2.jpg
where N is the number of rounds used in the calculation & x is the rating for each round.

the ratings guide mentions (in conjunction with the standard deviation idea) any round 100pts below your rating or more will be dropped. one can then assume anything less would be kept.

I'm assuming that its S.D. including the offending score, right?

The reason I ask is that with a small number of rounds, the one bad round has to be really really bad. Look at Dixon's 10 rounds not including the 928:

Average of 1014 with a S.D. of 18.5. That says that you wouldn't include any round with a score of lower than 968.

Add in the 928 and the S.D. jumps to 31.1 and the cutoff drops to 927 or 928...

If it had been a 925 it wouldn't have counted...

Come on Ray, don't just show off, do the math! (And don't cop out with "I don't have enough data" or "Too many unknown values") you can do complex math solve for as many unknown values as you were given, no whining!!

Come on Ray, don't just show off, do the math! (And don't cop out with "I don't have enough data" or "Too many unknown values") you can do complex math solve for as many unknown values as you were given, no whining!!


I will happily provide the tools, Shark, but no farging way I'm gonna look up all of that dude's included rated rounds & do the calc for him. As you well know, I'm just not that nice of a guy!

Apparently, reallybadputter is. :D

easycalculation.com for the intelectually lazy like me.

I took a quick look at his ratings and saw that there were only 11 rated rounds and they were all insanely tightly bunched. With 9 rounds between 998 and 1027, a 1050 and a 928.

With only 11 numbers it took 15 seconds to type them into Excel...

=stdev( {highlight the list} ) and you have your answer...

The interesting thing is that will make you pull your hair out with the ratings math:

If the round that he shot a 1050 on, he had only shot a 1025 rated round, the 928 wouldn't have counted and he'd have a 1012 rating...
:confused:

I took a quick look at his ratings and saw that there were only 11 rated rounds and they were all insanely tightly bunched. With 9 rounds between 998 and 1027, a 1050 and a 928.

With only 11 numbers it took 15 seconds to type them into Excel...

=stdev( {highlight the list} ) and you have your answer...

The interesting thing is that will make you pull your hair out with the ratings math:

If the round that he shot a 1050 on, he had only shot a 1025 rated round, the 928 wouldn't have counted and he'd have a 1012 rating...
:confused:



Sucks eh? I had to shoot a 935 to get it dropped and I have 70 rounds and am 1005. Just a bummer that shooting a real good round could come back to bite you in the arse. Maybe they should go back to dropping a % of your lowest rounds but instead of 15% make it <5%

I'd like to address the issue brought up here with Roger. The problem is the computing cycles that might be required to test Std Devs if a round rating is excluded from all other rounds versus just calculating the overall SD like is done now. Obviously, having one bad round can dramatically increase the SD and pull that round back into the calcs when the person was pretty consistent on all other rounds.

I took a quick look at his ratings and saw that there were only 11 rated rounds and they were all insanely tightly bunched. With 9 rounds between 998 and 1027, a 1050 and a 928.

With only 11 numbers it took 15 seconds to type them into Excel...

=stdev( {highlight the list} ) and you have your answer...

The interesting thing is that will make you pull your hair out with the ratings math:

If the round that he shot a 1050 on, he had only shot a 1025 rated round, the 928 wouldn't have counted and he'd have a 1012 rating...
:confused:



Sucks eh? I had to shoot a 935 to get it dropped and I have 70 rounds and am 1005. Just a bummer that shooting a real good round could come back to bite you in the arse. Maybe they should go back to dropping a % of your lowest rounds but instead of 15% make it <5%



Yeah, but you throw your share of rounds in the high 960s and 970s in with that 1005 average.

Your standard deviation is probably very typical. The original poster had 9 rounds with a really tight SD and then one blow up and one really hot round.

Of course since he only had 11 rounds, the stats swing pretty quickly...

The computing cycles wouldn't be too bad (O(n)) if you did something like this to calculate the ratings drop line:
if num_rounds>8, then compute average/std dev over the data with the highest and lowest rated round excluded.

It likely would add up to two more passes than the two Roger does now. But I agree that we need to address the issue in an effective way.

and this round double sucks because it is one of the most recent 25% so it gets counted twice. guess i shoulda played one throw worse on that day or missed a couple of putts in my hot round a couple of weeks before and i would be better. i guess?

yep, mine is just the opposite- if i'd shot one stroke better my last round my rating would be 10+ points worse.

and this round double sucks because it is one of the most recent 25% so it gets counted twice. guess i shoulda played one throw worse on that day or missed a couple of putts in my hot round a couple of weeks before and i would be better. i guess?






Doubling your most recent 25% isn't dumb. It makes the ratings better reflect how well you have been shooting recently and gives your recent rounds more weight than the ones from a year ago. I have a whole tournament avg 960 that is doubled and it helped me drop two points instead of going up 4.

Be glad PDGA is not using Six Sigma, nothing would ever get dropped.

guess i shoulda played one throw worse on that day or missed a couple of putts in my hot round a couple of weeks before and i would be better. i guess?



if you didn't answer your own question, the answer is, yes.

who says you can't have fun with mathematics? pDGA ratings make me laugh all the time :D

accredited mathematics professors would likely theorize the accuracy of pDGA ratings is lowered by reliance on very limited data.

others say otherwise:

http://blog.joehuffman.org/content/binary/CalvinMath.jpg