Hole and Course Performance Statistics

[Steve West]

For 2020 USDGC:

 

[Steve West]

Just for fun, I summarized over a million scores. Both Rating and Length were rounded to the nearest 10.

 

[ru4por]

Just for fun, I summarized over a million scores. Both Rating and Length were rounded to the nearest 10.

Any hypothesis?

 

[Steve West]

Any hypothesis?

Short holes aren't as good at separating players by skill as long holes.

What do you see?

 

[BogeyNoMore]

Abstract art?

 

[Steve West]

Some numbers:

On a 200-foot long hole, two players need to be 250 ratings points apart to expect a one-throw difference in their average scores.

On an 875-foot long hole, two players that are 100 ratings points apart will have a one-throw difference in their average score.

 

[Steve West]

A couple of more spin-offs from me figuring out how to handle so much data.

 

[Steve West]

More stupid - but surprising - trivia. Average length of throws by rating.

(Total length of holes divided by total scores.)

 

[ru4por]

More stupid - but surprising - trivia. Average length of throws by rating.

(Total length of holes divided by total scores.)

Is the distance supposed to be meters?

 

[Cgkdisc]

Is the distance supposed to be meters?

Just feet. You get a 2 on a 340 ft hole and it shows up as a 170 ft average data point.

 

[ru4por]

Just feet. You get a 2 on a 340 ft hole and it shows up as a 170 ft average data point.

ahhhhh.....got it now. :doh: Thanks Chuck.

 

[Karl]

More stupid - but surprising - trivia. Average length of throws by rating.

(Total length of holes divided by total scores.)

What do you find surprising? Seems anticipated, since better players DO make more long putts and CAN throw longer.

 

[Steve West]

I've been working on a way to compare events in terms of their ability to give a group of players a better scoring spread. It's based on event vs. event matchups. Taking only the players that played both events, I calculate the information value of the set of final scores given to those players.

I then solve for a set of factors which reproduces the total information generated by each event across all matchups.

For example, say there were 8 players who played both Event A and Event B. Event A gave each player a unique score, while at Event B each player was tied with one other player for four different scores. Event A generated a scoring spread of 8, which is 3 bits of information. Event B generated a scoring spread of 4, which is 2 bits of information. Event A generated 150% as much information as Event B. If these were the only two events, Event A would get a factor of 1.2 and Event B would get a factor of 0.8. [1.2/.8 = 150%.]

I then multiplied the factors by 1000 because that makes them look like the PDGA player ratings, which are relatable.

Here are the results for the first half of the year. Do these make sense? Or would we be able to know whether they do or not?

Event = Skill Sorting Power
2021 High Plains Challenge Sponsored by Discmania = 1077
2021 Republic Wireless presents the 38th Annual Dogwood Crosstown Classic, Driven by CADL = 1076
2021 St. Patrick's Classic - Powered by Innova Champion Discs = 1062
2021 The 303 Open = 1052
2021 Treasure Valley Cup Driven by Innova Discs and 208 Discs = 1048
2021 PDGA Professional Disc Golf World Championships Presented by Grip6 = 1046
2021 Dynamic Discs Open - National Tour = 1041
2021 Kitsap Classic - Presented by Dynamic Discs, Hosted by A Advanced Septic & Construction = 1036
DGPT - Waco Annual Charity Open presented by Prodigy Disc = 1034
DGPT - Las Vegas Challenge presented by Innova = 1023
2021 Utah Open Driven by Innova = 1019
Memorial Championship presented by Discraft = 1013
Shelly Sharpe Memorial presented by SpinnersontheGreen.com = 1009
2021 DGPT - OTB Open presented by Innova = 1004
2021 Discraft's Lower Peninsula Open + WGE = 1000
Titan Open 2021 = 1000
2021 DGPT - Play It Again Sports Jonesboro Open powered by Prodigy = 996
2021 Joe Machens Toyota's Mid America Open a DGPT Silver Series Event presented by Discraft = 990
2021 Discraft presents The Open at Belton a DGPT Silver Series Event = 989
2021 Tennessee State Disc Golf Championships = 988
2021 Santa Cruz Masters Cup Presented by Innova - National Tour = 985
2021 28th Annual City of Mobile Championships Presented by Dynamic Discs = 982
2021 Bowling Green Open = 980
2021 The Magic City Mega Bowl presented by Dynamic Discs = 980
2021 DGPT - Portland Open presented by Dynamic Discs = 980
2021 The Vintage Open a DGPT Silver Series Event = 979
2021 26th Annual Texas State Disc Golf Championship Presented by Latitude 64 - National Tour = 973
2021 Moonswitch Presents the 43rd Annual Norman Pro/Am driven by Innova = 968
2021 The Challenge @ Goat Hill Park - A DGPT Silver Series Event - Presented by Innova (MPO/FPO only) = 965
2021 Discraft presents Sun King's Throw Down the Mountain IX (Weekend 3) = 963
2021 Huk Central presented by Southern Wesleyan University - Pro = 963
The 38th Annual Victoria Open driven by Innova = 942
2021 Fort Steilacoom Open - Driven by Innova = 934
2021 Michigan Disc Golf Hall of Fame Classic = 926
2021 The Rumble Presented by Iron Lion Disc Golf Supply = 902

 

[Steve West]

Final results of the Scoring Spread Match-ups for 2021.

http://www.stevewestdiscgolf.com/ScoringSpreadMatchup.pdf

 

[Steve West]

Feet per SSA Throw

Looking at the results of A-tier+ rounds, I divided the length of the course by the Scratch Scoring Average (SSA, or approximately what a 1000-rated player would score).

Not surprisingly, the average throw goes farther on longer courses. Longer courses tend to be more wide-open and short courses just don't have room for huge throws.

Here is how the length of each throw varies by length of the course.

Note this is fairly linear, which means that the score on a course generally goes up with the square root of the length of the course. (The math is left as an exercise for the student.)

This uncovers an inefficiency in creating higher-scoring courses.

For example, a typical course with an SSA of 72 would be 178% as long as a course with an SSA of 54. The SSA=72 course would occupy 182% as much land as the SSA=54 course.

(Fairways get only a little wider on longer holes, so the total area of a course does not increase with the square of the length.)

The 18 extra throws require about 2.5 times as much land per throw as the first 54 throws.

(Unless we start making longer courses just as tight as shorter courses.)

 

[Steve West]

Here are the feet per SSA throw for the top and bottom ten. A shorter feet per throw means more trees, doglegs, or uphill. A longer feet per throw means fewer trees, fewer doglegs, or downhill.
(Other rounds on the same course at the same event are excluded from this list.)

90 = Swordstone; 18 holes; Par 59; 5,614 ft. at NADGT-National Championships (finals)
93 = Lake Ponca - Bois D'Arc - Bois D'arc; 18 holes; Par 55; 4,869 ft. at Ponca City Open Sponsored By Dynamic Discs (rd2)
94 = Cedar Creek West - Cedar Creek West Shorts; 18 holes; Par 55; 4,605 ft. at Cedar Creek Open (rd3)
95 = Tittabawassee Trails - Pro states 2021; 18 holes; Par 55; 4,854 ft. at Michigan State Professional Championships (rd3)
98 = Robert Morris - RoMo Short Tees; 18 holes; Par 54; 4,293 ft. at Michigan Disc Golf Hall of Fame Classic (rd2)
100 = Inverness Disc Golf Park - Inverness Mega Bowl Mixed Tees Green; 18 holes; Par 57; 5,457 ft. at The Magic City Mega Bowl presented by Dynamic Discs (rd3)
100 = Mont du Lac - White Cedar - White Cedar Long Pins; 18 holes; Par 60; 5,571 ft. at Lake Superior Open (rd2)
102 = Orange Crush - Orange Crush - Short Tee to B Pin; 18 holes; Par 59; 5,224 ft. at Seth Burton Memorial (rd3)
102 = Cedar Creek East - Cedar Creek East; 18 holes; Par 56; 4,805 ft. at Cedar Creek Open (rd2)
102 = Cedar Creek Central - Cedar Creek Central; 18 holes; Par 58; 5,405 ft. at Cedar Creek Open (rd3)
- - - - - -
162 = Jones Gold - Jones Gold Long Layout Rd1; 18 holes; Par 69; 10,511 ft. at Dynamic Discs Open - National Tour (rd1)
163 = Glendoveer - FPO; 18 holes; Par 70; 10,522 ft. at DGPT - Portland Open presented by Dynamic Discs (rd3)
163 = Toboggan Longs; 18 holes; Par 63; 10,159 ft. at DGPT - Discraft's Great Lakes Open (rd3)
165 = Ramirent Nevas Golf; 18 holes; Par 61; 2,814 m at NBDG Tyyni sponsored by Latitude 64 (rd1)
165 = Airborn Disc Golf Preserve - DGPT Black Bear FPO; 18 holes; Par 66; 9,874 ft. at DGPT - The Preserve Championship designed by Leiviska Disc Golf (rd1)
169 = Glendoveer - FPO Rd1; 18 holes; Par 70; 10,522 ft. at DGPT - Portland Open presented by Dynamic Discs (rd1)
174 = Airborn Disc Golf Preserve - DGPT Black Bear MPO; 18 holes; Par 65; 10,806 ft. at DGPT - The Preserve Championship designed by Leiviska Disc Golf (rd3)

 

[DavidSauls]

Isn't that largely because all throws are lumped together -- drives and long fairway shots and putts and tap-ins?

 

[Steve West]

Isn't that largely because all throws are lumped together -- drives and long fairway shots and putts and tap-ins?

Well, yeah, the final throws on every hole are going to be fairly similar in length. The length of the drives and fairway throws is what causes the variation.

What's land-inefficient is that the way we've been making really high-scoring courses is to add long throws.

 

[DavidSauls]

Do we add long throws just to make them high scoring, or add long throws to have more long throws?

I find long 2nd (& 3rd) shots add a lot to a course. The tee shot is the same, everytime you play a particular hole (other than adjustments for wind). Putting is more or less the same, or at least similar to other holes. But those fairways shots are different each time you play a hole, depending on where your drive landed -- except on wide open holes, of course.

I think the throws/distance ratio would be more informative if you removed the final throw from each hole.

 

[Steve West]

Do we add long throws just to make them high scoring, or add long throws to have more long throws?

I find long 2nd (& 3rd) shots add a lot to a course. The tee shot is the same, everytime you play a particular hole (other than adjustments for wind). Putting is more or less the same, or at least similar to other holes. But those fairways shots are different each time you play a hole, depending on where your drive landed -- except on wide open holes, of course.

I think the throws/distance ratio would be more informative if you removed the final throw from each hole.

I'm not against adding long throws. I'm just saying we should be prepared for a disproportionate increase in land needed for higher scoring courses if we get those higher scores by making courses longer.

If the point was to find the actual average length of throws, then subtracting a constant would indeed be more accurate. (Based on previous work, it would actually be 30 throws per 18-hole round.) However, the point of this was just to illustrate the range of technicality vs. length. The ranking would not change.