PDGA Round rating question

[jksdg]

Was in a tournament where all but 1 player was a propagator.

I took a sum of the differential between player rating and round rating and the result was a +286 for the 43 propagators. On average a player shot almost 7 points above their rating.

I thought it should be a zero sum game.

What could I be missing?

https://www.pdga.com/tour/event/66968#MPO

 

[jmdaire27]

If you figure out the recipe of the secret sauce (algorithm), let me know.

 

[jksdg]

The other tournament I was in had the opposite.

Do we know if a player whose score is 2.5 SD below his rating is thrown out?

How about a player on the high side?

Does it try to figure out SSA using a certain percentage of the best scores or best propogators? In this tournament the winner and best player actually shot under his rating

 

[Steve West]

Was in a tournament where all but 1 player was a propagator.

I took a sum of the differential between player rating and round rating and the result was a +286 for the 43 propagators. On average a player shot almost 7 points above their rating.

I thought it should be a zero sum game.

What could I be missing?

https://www.pdga.com/tour/event/66968#MPO

If it was truly a zero-sum game, every player's rating would inexorably trend toward the average rating of all players.

While (I think) the ratings are fairly stable for players rated around 1000, anti-clumping agents push the ratings apart for other players - to fight the trend that zero-sum averaging would induce.

Also, the average round rating is 2.2 points higher than the player rating. This reflects that the typical player gets about a throw better every 4 or 5 competitive rounds. Remember, the average player rating is about 913, so it's not hard to believe.

But, the go-to explanation is that these round ratings aren't official yet.

 

[Moose33]

Was in a tournament where all but 1 player was a propagator.

I took a sum of the differential between player rating and round rating and the result was a +286 for the 43 propagators. On average a player shot almost 7 points above their rating.

I thought it should be a zero sum game.

What could I be missing?

https://www.pdga.com/tour/event/66968#MPO

It also compares previous rounds at that course as a baseline if I understand it correctly. So if players shot on average a stroke better than the last few events that could be your variation?

Where's Chuck? He can explain it better than anyone.

 

[jksdg]

I don't think that is true. PLUS this is the 1st tournament ever there.

The one thing that could have some impact is the 2nd highest player 888. That wouldn't be enough to have that much of an impact.

 

[jksdg]

Does a 888 have the same impact as 999?

Do they both get treated as the round didn't happen?

For this tournament the ratings didn't change when is 91 was changed to 888. I wonder if everyone is going g to take a 3-4 point hit when this goes official

 

[BillFleming]

Does a 888 have the same impact as 999?

Do they both get treated as the round didn't happen?

For this tournament the ratings didn't change when is 91 was changed to 888. I wonder if everyone is going g to take a 3-4 point hit when this goes official

No and no.

I can't recall right now which is which, but one is a Did Not Finish and it does affect the rating. The other is Did Not Finish Due to Medical Reasons (or something like that) and does not affect the rating. It could be Did Not Finish, Exception Granted....or something, I don't know the exact wording....I just know one counts against the ratings and one doesn't. ((I was a course director and, with the Tournament Director we had to decide if a player deserved the DNF for Medical Reasons so that her round wouldn't count and that is how I know there is a difference between the two))

 

[ToddL]

No and no.

I can't recall right now which is which, but one is a Did Not Finish and it does affect the rating. The other is Did Not Finish Due to Medical Reasons (or something like that) and does not affect the rating. It could be Did Not Finish, Exception Granted....or something, I don't know the exact wording....I just know one counts against the ratings and one doesn't. ((I was a course director and, with the Tournament Director we had to decide if a player deserved the DNF for Medical Reasons so that her round wouldn't count and that is how I know there is a difference between the two))

Neither affect the round ratings for the field.

An 888 has a penalty for the player, but it doesn't affect any other players.

 

[biscoe]

It also compares previous rounds at that course as a baseline if I understand it correctly. So if players shot on average a stroke better than the last few events that could be your variation?

Where's Chuck? He can explain it better than anyone.

Previous rounds are not included in any way. Whether they should be or not is a different question.

 

[BillFleming]

Neither affect the round ratings for the field.

An 888 has a penalty for the player, but it doesn't affect any other players.

I thought the poster was asking how it affected the round for the player getting that number, not the entire field of players.

 

[Moose33]

Previous rounds are not included in any way. Whether they should be or not is a different question.

Hmm, that's interesting. It seems like that would be a good facet so you don't get aberrations.

 

[Steve West]

Hmm, that's interesting. It seems like that would be a good facet so you don't get aberrations.

That would likely take human intervention to determine whether the layout was actually the same or not.

Take Hillcrest Farm. Even if the tees and targets have not changed, is it the same course since the hurricane?

The potential for multiple events on the same layout is small. TDs are always tinkering with the layout and ground rules - as they should, in pursuit of continual improvement.

Even if the layout was the same, are the scores from the windy spring DDO comparable to the scores from the calmer summer Worlds?

Much of the potential enhanced information is being used anyway. The data from other events on the same layout is baked into the player ratings of those who played the other layout. Many of the same players will play both layouts.

 

[@CD-]

Was in a tournament where all but 1 player was a propagator.

I took a sum of the differential between player rating and round rating and the result was a +286 for the 43 propagators. On average a player shot almost 7 points above their rating.

I thought it should be a zero sum game.

What could I be missing?

https://www.pdga.com/tour/event/66968#MPO

Probably need to wait for the finalization of ratings since this is the first time this course has hosted a tournament. Every throw was worth ~6.67 pts so some rounding may be occurring to get whole numbers and throw off the sum calculation.

 

[Cgkdisc]

How much was OB a factor on this course layout?

 

[Steve West]

How much was OB a factor on this course layout?

If my calculations are correct, the average score was 10.5 throws higher than expected based on length alone.

 

[jksdg]

Only hole 18 had OB. It was 250 foot shot with a 23 foot island. The drop zone was 40 feet from the basket.

 

[jksdg]

The course is extremely wooded. Some holes are right on the edge of being too tight.

We thought par would be in the 1015 area and were surprised it was only 1002 (so far as not official).

 

[Steve West]

If my calculations are correct, the average score was 10.5 throws higher than expected based on length alone.

My calculations were not correct*. The average score was "only" 9.0 throws higher than expected based on length alone.

(Gotta keep in mind that it is throws/feet that is linear with ratings, not feet/throw.)

My gut feel has always been that a lot of trees is about half as punishing as one OB stroke per hole.



* Well, technically they were correct at two points.

 

[Cgkdisc]

The course is extremely wooded. Some holes are right on the edge of being too tight.

How do you know they were on the "edge" versus being too tight? One thing to look at is the correlation between the players' ratings and their round ratings from the event. A "good" course for the skill levels of the players should produce a good correlation in these ratings. I get 77% correlation in these numbers which isn't too bad overall but 85%-90% is more common on more open or well-worn tracks with minimal OB.

Steve may not want to do it, but correlations could be done on individual holes to determine some holes that might be "too lucky". Separate from narrowness of fairways, recoverability is also a key element in testing skill versus luck. Typically, the rough areas just beyond the newer fairways through the woods either haven't been cleared sufficiently or beat down enough for more skilled players to execute a recovery shot to save a stroke versus getting randomly punished the same as lower skilled players.