DGCR Scratch Scoring Estimate

[grodney]

For what it's worth, Chuck always said the PDGA rating system wasn't valid below a certain SSA -- 41 or 42 as I recall -- and they wouldn't calculate ratings if the SSA was below that. Maybe Chuck and Roger had a math reason for this, but it always seemed pretty arbitrary to me.

 

[ERicJ]

you would have to assume that almost every pro in the field, hell even most of the advanced players, would shoot right at -18. if that were to happen it seems the pdga couldn't award a score of -18 a 1000 rating; to many people accomplished it. you would also expect that at least one pro could ace at least one hole.

i don't know enough about everything that goes in to pdga ratings to say but honestly i would think not even -18 would get a 1000 rated round. so for me it makes sense to have it like that, even though it does seem ridiculous.

The PDGA doesn't "award" round ratings to specific scores. The sum of the [propagators'] player-ratings points going in equals the sum of the [propagators'] round-ratings points coming out. [That I know of] There's no artificial limit on the number of players that could shoot a 1000-rated round. If you had the top 30 players in the world play the course and they were all over 1000-rated, and they all shot -18, then they would all get a rating over 1000 for that round.

 

[Stud Muffin]

I believe it is a good formula, because at Engler, you have the SSE at 47.5 white and 51.2 blue, and during the Gateway Open, Justin Bunnell (*rating 992) won the tournament shooting 46 white and 52 blue.

Good job guys!
:clap:

 

[optidiscic]

It's not perfect and it doesn't take into account OOB/water or unique obstacles but overall it does a remarkable job of at least giving you an idea of what to expect.....my home course comes in at 22 and the best rounds are typically 4 under so it makes sense ...sort of

 

[Rip Van]

The SSE provides an extremely useful thumbnail view of course difficulty.

The alternative is to root around the PDGA results, trying to find a tournament that plays that course, and find someone with a 1000-rated score. That's harder than it seems, because the course may not have recently hosted a PDGA event. And even if it has, it may not be anything named as obviously as "Take a Peek at Reedy Creek", so you need to know what course maps to what tournament. Even if you've figured that out, if it's a multi-course tournament, it's not obvious who's playing which course for which round. Then there are occasional data entry errors, where a group gets rated against the longs when it played the shorts, or vice versa. Finally, there's the inexplicable "disappearing ratings" syndrome, when no ratings are visible between the initial ratings posting and the final ratings posting. You'd think they'd just put the ratings in italics or red while they're not official, but keep them available the entire time.

The PDGA takes an insular approach to course ratings data. That's certainly their right, as it's their data, and apparently people were scraping course ratings info and trying to re-use it. The SSE may well become the consensus measure of course difficulty, as most people just want a ballpark figure, and the SSE is the only reasonable way to get it right now.

 

[DavidSauls]

Another alternative---perhaps better, perhaps worse---would be to have users enter the course SSA along with the other course statistics (hole length, par, year installed, etc.). Presumably someone familiar with a course would have a good idea of whether tournament layouts matched the everyday layout, and have little trouble finding a representative event.

They could get it wrong, of course, but that's true of all the rest of the info posted to this site.

 

[Dave242]

The problems with that IMO are prone to much more inaccuracy across the board. Sure there are some/many cases were it would work perfectly, but these are the factors I can think of that add the chance of error:

• Many courses never have a PDGA event - so that would leave tons of holes in the data
• PDGA events are often set up in a configuration that will confuse the person trying to interpret and enter SSA (extra holes, subset of holes on >18 holers, ground rules, longer or different temporary setup....heck, just look at the course setups being used at Worlds in Charlotte!)
• Multiple tees and/or basket locations: DGCR SSE accounts for the current set up in DGCR.....SSA can never do that.
• Weather
• The majority of DGCR users do not really understand SSA (let alone play competitively) so the data is prone to being entered in error

I think it us much cooler anyhow to have something "home cooked" based solely on DGCR's "intellectual property" (our own understand of length and what "lightly/moderately/heavily wooded" is)

 

[New013]

All of my statistics and physics classes have taught me one thing, human error is the most probable of all errors

 

[grodney]

I think it us much cooler anyhow to have something "home cooked" based solely on DGCR's "intellectual property" (our own understand of length and what "lightly/moderately/heavily wooded" is)

Well, other than the fact that the formulas for SSE were reverse-engineered from published PDGA results/ratings.

 

[Dave242]

well yes.......by definition they needed to be since "Scratch" is defined by them. But.....there is very little special scaling or slope-curving in the formulas and none zip nada course-by-course adjustable trimming/tweaking ability.

Now, what was reverse engineered is the method developed here to derive personal round ratings based off of SSE (or SSA). timg has not implemented that feature into the scorecards......but it will be incredibly cool if/when he does.

 

[timg]

I'm planning to create round ratings. Just haven't had the time. It'll happen!

 

[grodney]

well yes.......by definition they needed to be since "Scratch" is defined by them. But.....there is very little special scaling or slope-curving in the formulas and none zip nada course-by-course adjustable trimming/tweaking ability.

Oh yeah, the SSE was based off the very public estimate formula, wasn't it. I forgot that. But yeah, as you said "scratch" as used in SSE is a PDGA thing anyway.

 

[Dave242]

Yup....the basic formula is SSE = Length/285+1.67x#Holes (which is good for average length courses (~5,000' for 18 holes) in moderately wooded areas. That is a PDGA formula.....but certainly no secret sauce involved in that. If anything, it is harmonizing with the PDGA methodology and driving people to their worldview and standards.

 

[jeverett]

Oh yeah, the SSE was based off the very public estimate formula, wasn't it. I forgot that. But yeah, as you said "scratch" as used in SSE is a PDGA thing anyway.

Yes, to be very precise, we're "estimating" or "approximating" what a 1000-rated (scratch) round might be for a specific course and course layout, using a generalized formula that takes into account (only) current course layout length, number of holes, and the DGCR course 'landscape' field.

The conversion between this scratch score estimate and a player's round round rating (not currently implemented here on DGCR) is available on the original DGCR thread, and was reverse engineered (by myself) using a sampling of real PDGA SSA data. It is 99.999% accurate to the actual PDGA formulas, given the stipulation that I threw out one or two rounds of data that apparently were submitted to the PDGA incorrectly (the round data was labeled with the wrong course or layout). Overall, though, the formulas will only be as accurate as the SSE is accurate, as the formulas take the SSE value and convert it into a points-per-throw-difference metric.

 

[jenb]

While I agree that your conversion of the SSE to to a round rating is quite good, I think the results of applying your formula reveal that the SSE being calculated is way off.

Take Cherry Creek DGC in Canton, TX for example. For the blues (length 5346 feet), the DGCR SSE is a ridiculous 45.1. Using Chuck's raw formula that doesn't take vegetation into account, you would get a more accurate result at 48.76 SSE. Using actual tournament results, the PDGA app reveals that the true SSA of this course from the blues is exactly 50. That makes a huge difference when calculating ratings for people scoring in the 60s.

And look at the SSE for the reds at Cherry Creek. 38? Seriously? You're gonna expect 100 scratch players to average 16 under on that course from the reds? I don't think so.

Bottom line. I think you would be better off using Chuck's formula for average vegatation, without trying to take vegetation into account. And the expectation that players will throw over 300 feet on average without dense vegetation seems highly unrealistic to me. The formula averages distances drives and approaches, then adds 1.67 putts per hole. You seem to think the average of a drive and approach would be 335 or something like that. Get real.

 

[Dave242]

Cherry Creek looks moderately wooded to me from the pictures. Would you agree? Plus, read the course description. If so, make that adjustment and see what happens to the SSE.

 

[optidiscic]

the big problem comes when you get a course with open holes and wooded holes and it gets labeled moderately wooded....even if it's 3 wide open holes and 15 tightly wooded holes. I also don't see any emphasis on elevation and oob which can alter the formula as well. There is a lot of flaws. But it is better than nothing.
For the most part it works but there really is no way to make it perfect.

 

[Dave242]

Why is that a big problem? If you notice the problem of a course being listed incorrectly..........fix it. :clap::clap::thmbup:

 

[jenb]

Why is that a big problem? If you notice the problem of a course being listed incorrectly..........fix it. :clap::clap::thmbup:

And just how do I do that? Do I have to pay money or something? Eff that.

 

[Dave242]

If you are a logged in user (I imagine that is required) go to the tab on the course's page labelled "update course info" and make the change. That tab is about half way down the page on the right side.