Beginner Throwing Distance

[Cgkdisc]

There are three types of holes that don't reward power/distance. One is a downhill drop shot. In theory, a ski hill course that starts at the top where all the holes are downhill and reachable by everyone in the field with putters, would neutralize power/distance the most. Even then, power matters for putting in the 25-70 ft range for throwing lasers versus more arching putts required by less powerful throwers.

Another is a dogleg or bend in the fairway that can't be defeated by going over the top to where all players in the field are able to reach the dogleg with mid-ranges and then the hole from the dogleg. The third is a hole with an OB area that has to be crossed, everyone in the field can reach the front edge of the pond off the tee and the pond is wide enough that no one can clear the pond from the tee or at least wants to risk it.

Those types of holes are a small percentage of all holes being played. So, it should be no surprise that Steve's graph shows how important power/distance is in our game as currently designed whether thru woods or mostly open.

 

[Rastnav]

So does this imply that tight, difficult, short courses, by their very nature in not allowing players to outcompete using their distance advantage, will not tend to show results that conform to ratings?

Also, is it possible that distance and other skills correlate such that the higher the rating, the better the player is, on average, in all phases of the game including distance?

I believe that the answer to this is sort-of "yes", with the specific term for it being "compression". Basically, each stroke gained is worth fewer ratings points.

That doesn't necessarily mean it's harder to shoot your rating though. I think that would depend on the precise mix of propagators. If you were the lone 1000 rated player in a tourney, shooting -18, and the next highest rated player is 950 and shoots -17, with a fairly even/expected distribution of scores and ratings below that, I don't believe your round would be 1000 rated.

Conversely, take that 1000 rated player out of the mix, and that 950 rated player probably ends up with a 950 rated round. Basically, I don't believe you don't "want" to be the highly rated outlier when playing on a course that will end up compressing scores.

https://www.pdga.com/ratings/guide

The reason that the ratings points per throw change over the range of course difficulties is an effect called "compression." On an easy course, top players can only shoot so well, given they are limited to scoring no better than a 2 on virtually every hole. However, on these courses, where the average hole is likely to be wide open and less than 250 feet, even lower rated players can also shoot lots of 2s. This "compresses" or narrows the range of scores for players of widely varying skills in that round. On the other hand, a difficult course with an SSA over 60 will spread the scores farther apart in each round when compared to a course with a scoring average of around 50 for scratch players.

 

[Steve West]

So does this imply that tight, difficult, short courses, by their very nature in not allowing players to outcompete using their distance advantage, will not tend to show results that conform to ratings?

Also, is it possible that distance and other skills correlate such that the higher the rating, the better the player is, on average, in all phases of the game including distance?

Frankly, the idea that certain top players perform better on certain types of courses is only a conjecture at this point. When I try to find evidence for it, I get evidence against it.

Here are three reasons the conjecture might be untrue:

First, all top players need all the skills that are being tested on tour. So they play alike.

Second, all courses used for big events are actually quite similar to each other. This is because we look for balance and variety. It's like always taking one of each item at the buffet. That makes for the most interesting meal, but then all your meals are the same as each other. So the courses play alike.

Third, courses are designed for excitement: "Anything can happen up until the last putt!" This can only be accomplished by dampening the differences in skills. So the resulting scores are alike.


For lower-rated players, yes, there are short vs. long throwers, accurate vs. not, etc. and the lower rated players are only good at some of these skills. Also, courses that aren't on tour vary more widely, because there are some short, bad, and dull courses being played.

 

[ChrisWoj]

I'm developing a Disc Golf Course Designer (DGCD) Purple Skill layout guideline. I've already implemented this on some courses and will soon be implementing it on a few more as follows:

Total 18-hole Course Length between 2250' and 3150' which falls between 125'-175' (38m-53m) average per hole.

Par 3 = 70' to 147' effective* hole length (21m-45m)
Par 4 = 148' to 279' effective hole length (46m-85m)
Par 5 = 280'+ (85m+) effective hole length. However, holes beyond this length not recommended when designing new Purple level layouts but may be necessary when fitting a purple layout on an existing course.

Recommend at least 2 holes effectively under 108' (33m) per 9 holes
Recommend no more than 2 holes effectively over 246' (75m) per 9 holes
Recommend that every hole include at least one shot-shaping challenge around, between or over vertical obstacles even if they must be artificially added on terrain lacking them.
Recommend not requiring players to throw their tee shot more than 50 feet to clear a tough-to-traverse gully or body of water and always provide a reasonable route to throw around one.

* Effective Length estimate is the measured length of the hole plus or minus the vertical elevation difference between the tee and basket times 3. Example: 100 ft hole where pin is 5 feet higher than tee would have an Effective Length of 100' + (5' x 3) = 115 feet. (If downhill, it would be 85 ft). The elevation difference factor of 3 seems to work well with elevation differences up to 12%-15% of the horizontal distance. Beyond that we have no formulaic factor.

The goal for purple layouts is to serve several player groups:
1. Beginners throwing golf discs
2. Families throwing catch class discs (Frisbee(R) style)
3. Ultimate players throwing Ulty discs (user group on the increase)
4. Par 2 mid/putter challenge for experienced players

Sounds like your favorite course, Parmelee Park!

 

[Cgkdisc]

Sounds like your favorite course, Parmelee Park!

Kate liked Parmalee but still thought it was too long for Purple level.

 

[ChrisWoj]

Kate liked Parmalee but still thought it was too long for Purple level.

Damnit, we can only seem to build bomber layouts in Toledo.

 

[ChrisWoj]

Oops. Add two throws for close range up and down.

2+((200-143)/258) = 2.22, rounded up to 3.

Good catch.

Not pars, these were fit to average scores.

Can you define throw length and close range?
Is throw length related to par? birdie? Is it player can birdie X out of X times? par? For close range, able to get up and down how often? I'm just not sure I'm understanding the totality of what this represents.

 

[Doofenshmirtz]

For lower-rated players, yes, there are short vs. long throwers, accurate vs. not, etc. and the lower rated players are only good at some of these skills. Also, courses that aren't on tour vary more widely, because there are some short, bad, and dull courses being played.

Let's get rid of touring pros and take 1000 rated and 950 rated players. If I hold a hypothetical tournament with two courses, a long, mostly open course (8,500 feet) and a very short, technical, heavily wooded course (4400 feet). Can I expect, with a large field, say, 100 players of each rating, the results to average out with the higher rated group averaging the number of throws better than the lower rated group such that those averages generally conform to the difference in rating on both courses? I.e., with the 1000 rated group play 50 rating points better than the 950 rated group on both courses?

 

[Doofenshmirtz]

There are three types of holes that don't reward power/distance. One is a downhill drop shot. In theory, a ski hill course that starts at the top where all the holes are downhill and reachable by everyone in the field with putters, would neutralize power/distance the most. Even then, power matters for putting in the 25-70 ft range for throwing lasers versus more arching putts required by less powerful throwers.

Another is a dogleg or bend in the fairway that can't be defeated by going over the top to where all players in the field are able to reach the dogleg with mid-ranges and then the hole from the dogleg. The third is a hole with an OB area that has to be crossed, everyone in the field can reach the front edge of the pond off the tee and the pond is wide enough that no one can clear the pond from the tee or at least wants to risk it.

Those types of holes are a small percentage of all holes being played. So, it should be no surprise that Steve's graph shows how important power/distance is in our game as currently designed whether thru woods or mostly open.

So 200 - 280 ft tunnel holes reward distance just as much as 450', wide open holes? On my three home courses, I have 17 holes that fit the short and tight description.

My question was aimed at the fact that the graph shows the importance of distance without taking other skills into consideration. I'm completely prepared to accept that either Steve's conclusion is true or that, by the way that it is set up, it is designed to show the importance of distance to the exclusion of other skills. Based on Steve's answer regarding the similarity of courses on tour, it could be that Steve's graph is specific to a certain kind of course in its implications regarding distance, and I may not have caught where Steve indicated that the 1.8 million holes were only tour holes played.

BTW, I find the data in the chart fascinating and very useful even if I am not certain about conclusions to draw from it about the exact prominence of distance in determining rating (although it would appear to have definite importance). The chart fairly directly answers my original question with far more specificity that I realized that I needed.

One of the comparisons that keeps gnawing at me was the overlapping careers on the PGA tour of two golfers who had plenty of distance off the tee. Just as in golf, in disc golf a longer drive makes the upshot easier and typically results in shorter putts. But, if distance off the tee were that important by itself, one might figure that Tiger Woods and John Daly should have had similar career success. Yet, there was a time when Tiger Woods was hitting something like 98% of his putts inside of 10 feet compared to a tour average of 55%. Over the course of three seasons (2002-2005) Tiger missed only three putts out of 1,543 inside of 3 feet. Even at such close range, that is a staggering statistic. Daly was the distance leader of the PGA 11 times. Tiger never was. This small comparison isn't meant as a statistical analogy for disc golf, buit it does make me wonder if putting or approach graphs could not be compiled that would have similar curves. Of course, maybe they wouldn't and they would validate Steve's conclusion.

 

[Cgkdisc]

Let's get rid of touring pros and take 1000 rated and 950 rated players. If I hold a hypothetical tournament with two courses, a long, mostly open course (8,500 feet) and a very short, technical, heavily wooded course (4400 feet). Can I expect, with a large field, say, 100 players of each rating, the results to average out with the higher rated group averaging the number of throws better than the lower rated group such that those averages generally conform to the difference in rating on both courses? I.e., with the 1000 rated group play 50 rating points better than the 950 rated group on both courses?

In general, yes. However, I could design the 8500 ft course in a way that those with 1000 rated arms would have an "unfair" edge and separate themselves from the 950 pool more than 50 points. The other thing to consider is that the 950 crowd is only giving up 0.25 and .33 strokes per hole. If the shorter course has a lot of "lucky" fairways and weird angles, it will bring the two pools closer than 50 points.

 

[txmxer]

I don't have the numbers, just observation, but when you see someone gave a hot round, it's almost always the player is clicking in all areas.

When Ricky or Paul have an off round, there will be surprising missed putts, tree kicks where the line was off by a few inches, etc.

Of course all of the top rated players throw long, but not all of the long throwers are top rated.

 

[Steve West]

Kate liked Parmalee but still thought it was too long for Purple level.

That's because she's an Orange level player.

 

[Steve West]

Can you define throw length and close range?
Is throw length related to par? birdie? Is it player can birdie X out of X times? par? For close range, able to get up and down how often? I'm just not sure I'm understanding the totality of what this represents.

Throw length is how far a typical player could typically throw. Likewise for CR.

If you take a random hole and use the throw length and close range (for a specified rating) to predict the score, it will be the same as the actual score from a player of that rating 50% of the time, and 25% of the time they'll get a higher score, and 25% of the time they'll get a lower score.

 

[Steve West]

Let's get rid of touring pros and take 1000 rated and 950 rated players. If I hold a hypothetical tournament with two courses, a long, mostly open course (8,500 feet) and a very short, technical, heavily wooded course (4400 feet). Can I expect, with a large field, say, 100 players of each rating, the results to average out with the higher rated group averaging the number of throws better than the lower rated group such that those averages generally conform to the difference in rating on both courses? I.e., with the 1000 rated group play 50 rating points better than the 950 rated group on both courses?

That is a different question than what I was talking about. I was talking about the conjecture that some players would consistently do better on one of those courses (or the other) than other equally rated players.

 

[Tinkles]

Here are some early results from some unpublished research. These are based on 1.8 million actual scores in tournaments, with the best fit for:

Score = [(Hole Length - Close Range) / Throw Length], rounded up.

Steve would this be supportive evidence that when calculating ratings for an event, that the value of strokes over/under par should follow a an inverse bell curve to an extent?

Is there a baseline rating at which it makes more sense to calculate round ratings from other than 1000?

 

[Steve West]

Steve would this be supportive evidence that when calculating ratings for an event, that the value of strokes over/under par should follow a an inverse bell curve to an extent?

Is there a baseline rating at which it makes more sense to calculate round ratings from other than 1000?

Probably yes to both. But I try to avoid thinking about a better ratings system. Especially not by looking at one flaw at a time. Any better system would need a comprehensive, defined-outcome-driven, holistic development. I'd guess it would take a team of about five people working (paid) full time for a year. More than I want to do, but any tinkering short of that will likely not be bug-free.

 

[Doofenshmirtz]

Throw length is how far a typical player could typically throw. Likewise for CR.

If you take a random hole and use the throw length and close range (for a specified rating) to predict the score, it will be the same as the actual score from a player of that rating 50% of the time, and 25% of the time they'll get a higher score, and 25% of the time they'll get a lower score.

Finally getting around to look at the numbers to try to get a better grip as they relate to certain hole distances. The first hole that came to mind is a local hole that is 600' long and mostly open. Using your chart (hopefully correctly) and the above comment, I get the following.

600 feet Hole

Rating Calculation

1050 2 + ((600 - 171)/496) = 3
1000 2 + ((600 - 160)/379) = 4
950 2 + ((600 - 151)/307) = 4
900 2 + ((600 - 143)/258) = 4
850 2 + ((600 - 135)/222) = 5

Because I am getting the course owner to put in alternate tee boxes with the intention that some be blue level, some gold level, I have compiled the availabe tournament scores for the course. The 600 ft hole sees MPO (with scores averaging below 1000) scoring mostly 3s and no 5s. MA1 (with scores averaging above 900) scores mostly 4s, with about 20% 3s and a few 5s. Of course, these are small samples.

This made me wonder whether there was an overlap in distance where 1000 and 850 would be expected to score the same. At 576 feet, the distance of another local hole (adjacent course) average score would be the same for both 1000 and 850 rated players.

576 feet hole

Rating Calculation

1000 2 + ((576 - 160)/379) = 4
850 2 + ((576 - 135)/222) = 4

I don't have 850 rated player numbers for this hole, but I can't imaging the distribution being the same for these two player ratings.

I'm not quibbling with the statistics, but I am wondering whether you quote above might better be stated:

"If you take a hole whose length equals the sum of the throw length and close range distance for a specified rating, the predicted score will be the same as the actual score from a player of that rating 50% of the time, and 25% of the time they'll get a higher score, and 25% of the time they'll get a lower score."

Would there be some sort of curve, not represented in your chart, that might represent the same data without overlapping 850 and 1000 rated players on such holes, or that might acknowledge the very different expected distribution of their scores, or maybe a chart that shows separate scoring curves for the different ratings at various distances?

Of course, maybe I'm just wrong.

 

[Cgkdisc]

Start with a course with an SSA of 50.4 where the rating points per stroke is 10.0. A 950-rated player would be expected to average 5 strokes higher than a 1000-rated player on this 18-hole course, a 900-rated average 5 strokes higher than the 950-rated, etc. 5 strokes divided by 18 equals 0.278 strokes per hole average difference between players 50 rating points apart although it will vary around that number based on the various hole lengths on the course.

Let's take a hypothetically "perfect" hole on that course with the 0.278 ~0.28 difference per hole. If we start with a hole where the 1000-rated average say 3.55 (and Steve's math considers it a par 4); the 950s would average 3.55+0.28= 3.83, still a par 4; the 900s would average 3.83+0.28 = 4.11, still a par 4; the 850s would average 4.11+0.28=4.39, likely still a par 4 with Steve's math.

Point being that the appropriate par on every hole will be the same for players spanning around 150 rating points, although their scoring distributions in this example average 0.84 strokes, almost a stroke higher for the 850s than the 1000s. If you believe that holes should be birdieable, the designer or TD looking at this math "might" consider boosting the 850-player par to a 5 if it were known that no 850-range player ever got a birdie 3 on this hole when par was set at 4.

 

[txmxer]

Here's something interesting.

According to Steve's chart/formula, low rated players are expected to be much more accurate as a percentage of throw length. i.e., a 1050 rated player is expected to throw ~500', but be accurate (close range) from only 34% of the maximum throw length, where as an 800 rated player is expected to be in close range at 66% of their maximum.

 

[txmxer]

for comparison sake, the draft PAR guidelines approach this differently.