Ratings School...

[brutalbrutus]

Aka Chuck over thinks a problem...

http://www.pdga.com/leagues-vs-tournaments-pt-1

 

[Vonbeezy1]

Aka Chuck over thinks a problem...

http://www.pdga.com/leagues-vs-tournaments-pt-1

Stopped reading after it said the SSA changes after every round lulz

 

[DavidSauls]

It makes sense, as far as it goes, but I suspect there are other factors in the SSA difference between league and tournament play. Among these are

---pace of play, thus a longer time between shots, thus harder to stay in a groove.
---a lot more local players, very familiar with the course, in league play.
---and perhaps something else that hasn't come to mind yet.

 

[jeverett]

Stopped reading after it said the SSA changes after every round lulz

I don't follow. SSA's are indeed [somewhat] different every round. How much different depends on lots of factors, such as number of propagators, changing course conditions, etc. The basic issue is over-extrapolation: because of all kinds of variables, the existing SSA system does not produce particularly precise (replicable) SSA values, compared to the particular scale used. i.e. if two rounds played on the same course and layout produce SSA values say 10 rating points apart (a fairly common occurrence), ultimately this means that the system itself cannot say for certain which is the 'correct' SSA value (i.e. the margin of error on each value is quite large compared to the scale used.

Or are you suggesting that SSA's *shouldn't* change? i.e. that the rating system should fundamentally work differently than it does?

 

[Cgkdisc]

It makes sense, as far as it goes, but I suspect there are other factors in the SSA difference between league and tournament play. Among these are

---pace of play, thus a longer time between shots, thus harder to stay in a groove.
---a lot more local players, very familiar with the course, in league play.
---and perhaps something else that hasn't come to mind yet.

It would be hard to ever determine for sure the contribution of different factors. The point is this combination of differing factors exists and virtually always influences scores in the same direction.

 

[iacas]

Stopped reading after it said the SSA changes after every round lulz

Why?

Also, golf is likely going to a "daily" course rating/slope system. Courses will still be rated (golf courses can't change as much as the A to the C pin position on a disc golf course), but on a day when a bunch of players shoot higher scores than their handicaps because of a bunch of wind or something, the rating/slope will be a little bit higher. (Assuming enough players with established handicaps play that day.)

 

[jeverett]

I'm actually most interested in a side-topic that the article didn't mention: the impact of the lower SSA's of league rounds on the ratings points-per-throw (i.e. the overall scoring spread of the rounds). In the PDGA system, the overall slope of the ratings (the 'points-per-throw') is computed solely based on the SSA of the round, with a pair of linear formulas converting SSA to points-per-throw (i.e. a lower SSA produces a smaller points-per-throw value). Do the formulas the PDGA system uses hold up under these kinds of SSA differences? Does the difference in SSA observed between league and tournament rounds produce an equal linear effect on the observed slope of the scoring spread? i.e. Are the league rounds, in addition to producing a lower SSA value, producing a lower overall scoring spread?

 

[jeverett]

I'm actually most interested in a side-topic that the article didn't mention: the impact of the lower SSA's of league rounds on the ratings points-per-throw (i.e. the overall scoring spread of the rounds). In the PDGA system, the overall slope of the ratings (the 'points-per-throw') is computed solely based on the SSA of the round, with a pair of linear formulas converting SSA to points-per-throw (i.e. a lower SSA produces a smaller points-per-throw value). Do the formulas the PDGA system uses hold up under these kinds of SSA differences? Does the difference in SSA observed between league and tournament rounds produce an equal linear effect on the observed slope of the scoring spread? i.e. Are the league rounds, in addition to producing a lower SSA value, producing a lower overall scoring spread?

Edit on the above: Oops. it should be 'larger' points-per-throw value. A lower SSA produces a larger points-per-throw value.

 

[cjman]

Hopefully part 3 will outline the process for bringing in a new rating system.

 

[Steve West]

Does the difference in SSA observed between league and tournament rounds produce an equal linear effect on the observed slope of the scoring spread? i.e. Are the league rounds, in addition to producing a lower SSA value, producing a lower overall scoring spread?

A smaller SSA will produce a narrower scoring spread (of actual scores) just because there are fewer throws to be spread around.

You're really asking about whether league rounds have a narrower scoring spread than would be expected for a tournament round with the same SSA, right?

What's your hypothesis on why the linear effect would be different between leagues and tournaments?

Isn't it just as possible league rounds produce a wider scoring spread than a tournament round with the same SSA?

For example, if the two were held on the same course, for the league round to have the same SSA as the tournament round it was probably windy or there was something that made the course physically more difficult on league day (to offset the lack of tournament pressure). Anything that made the course harder for everyone (like a steady wind, or the league was held before the foliage got trimmed) would generally widen the scoring spread because better players are less affected by adverse conditions.

 

[jeverett]

A smaller SSA will produce a narrower scoring spread (of actual scores) just because there are fewer throws to be spread around.

I'm actually interested in proving that, though. The PDGA system actually pegs a linear equation to the effect, and I want to know that the particular linear equation they use is the mathematically correct one. i.e. is the relationship between SSA and narrowing scoring spread purely based on there being fewer throws to to spread around? Or could there be other variables at work too? The PDGA method assumes there are none, and data like this (the same layout producing two different SSA's) as well as the data you mentioned (two different layouts producing the same SSA) should both be useful in trying to determine if the PDGA method is accurate or not.

 

[filobedo]

I'm actually interested in proving that, though. The PDGA system actually pegs a linear equation to the effect, and I want to know that the particular linear equation they use is the mathematically correct one. i.e. is the relationship between SSA and narrowing scoring spread purely based on there being fewer throws to to spread around? Or could there be other variables at work too? The PDGA method assumes there are none, and data like this (the same layout producing two different SSA's) as well as the data you mentioned (two different layouts producing the same SSA) should both be useful in trying to determine if the PDGA method is accurate or not.

opcorn:opcorn:opcorn:opcorn:opcorn:

The model is accurate because the bullies that created it said so. The rating system is a joke. I could have two entry level employees create something in a couple of hours that makes more sense.

 

[Steve West]

..is the relationship between SSA and narrowing scoring spread purely based on there being fewer throws to to spread around? Or could there be other variables at work too?

What other variables are possible? And how would you incorporate them into the ratings?

 

[iacas]

What other variables are possible? And how would you incorporate them into the ratings?

Golf has a two-pronged approach. It's still linear, but they have the course rating (like the SSA), and then they have the slope, which effectively changes the course rating as players get better or worse.

 

[gdub58]

The model is accurate because the bullies that created it said so. The rating system is a joke. I could have two entry level employees create something in a couple of hours that makes more sense.

Well then, by all means please enlighten us with an overview of your new system. How many rounds will be required to get a rating? Will there be any restriction on the types of courses or the composition of the field to be considered a "rateable" round?

I don't think I've heard anyone proclaim that the current system can't possibly be improved upon, but it's weak to just say that "it's a joke" without offering an alternative. Perhaps you've done that elsewhere, and if so then mea culpa.

Don't get me wrong, I have on more than one occasion thought that the rating given to one of my tournament rounds didn't seem to line up with how I thought I played, but since the only thing it is used for at this point is to restrict me from playing in some amateur divisions I don't assign it as much importance in my life as others might...

 

[jksdg]

I'd like to hear what specificily is wrong with the current rating system is wrong or needs fixing.

I have both a GHIN handicap and a PDGA rating. The PDGA rating is a much better indicator of a player vs. GHIN handicap.

If I were to nitpick I'd think the only issue would be the non linear difference between different types of players. Often times the type of course played amplifies this.

For me...if I wanted to maximize my rating (other than throwing a bad round to -2.5 SDs!!)
a. course selection is crucial, finding a course that can hide weaknesses
b. tournaments with lots of pros, taking advantage of non linear difference between me and pros. Especially on higher SSA courses.
c. avoid tournaments on courses i am unfamiliar with and there is a high local presence. Especially leagues where it is all locals

For the record I am not doing these things. None of these ways of "gaining the system" is enough to have me think ANYTHING needs to be changed regarding the rating system.

On a separate note I am not ready to subscribe to the notion scoring is better in leagues due to the lower pressure environment. Personally me and my family all play better in tournaments because the concentration level is higher and we tend to play smarter. In addition, we take more time with putting. In a causal round taking 20 seconds before a putt would give you the stink eye.

 

[jksdg]

I don't think I've heard anyone proclaim that the current system can't possibly be improved upon,

I might be in that camp. I can't think of any changes I'd make otr things to be eliminated.

2.5 SDs rule makes sense. Weighting recent rounds higher makes sense. 12 month look back seems right.

 

[DavidSauls]

I'd like to hear what specificily is wrong with the current rating system is wrong or needs fixing.

.

The biggest knock against it is that it doesn't do things that it wasn't designed to do, anyway.

Like create a fixed, permanent number (SSA) for a particular course. Or produce a single round rating so specific and accurate that it means something all by itself.

 

[slowplastic]

...but why can't we just have like...a number that's par and we're good if we shoot under it and not good if we shoot positive? And with no par 2's cuz that's a joke? And no matter the weather impacts it should still work...

Interesting article, makes sense. Unfortunately it also confirms that if I shoot a hot round one afternoon it's probably not what I'd do if there was pressure...

 

[Cgkdisc]

...but why can't we just have like...a number that's par and we're good if we shoot under it and not good if we shoot positive? And with no par 2's cuz that's a joke? And no matter the weather impacts it should still work...

Interesting article, makes sense. Unfortunately it also confirms that if I shoot a hot round one afternoon it's probably not what I'd do if there was pressure...

It's similar to comparing your speed difference above the limit when you actually see a cop versus your speed difference above the limit on a stretch of road where you've seen occasional traps, and your speed difference above the limit where it's unlikely a cop will ticket you (i.e. going 70 in a 55 zone around Chicago) or have a radar detector. These are different levels of pressure on your "performance."