1880 Doubles Oct 29th

[kscustom]

Wish I could have made it some fam stuff went down and i just couldn't get way

[Schoen-hopper]

I'm willing to listen on how the ratings system can be improved. But I really don't feel we had any problems with the ratings at this event.

Lots of handicap systems are calculated differently. The ACDGC's system works by taking 80% of the difference of the scoring average and 52. For players with more that 3 rounds: round scores that are 9 or more above the scoring average (scoring averages <60), 10 or more for SA 60 to 70, and 11 or more for SA 70+... are thrown out. Some players aren't happy to have any round thrown out, but some kind of non-subjective limit has to be set to keep players from padding their handicap, purposefully or not.

USGA Ball golf handicap is the most well known and respected system. And it's quite complicated. The round score is first adjusted, hole by hole, with a method called "equitable score control". For handicappers below 20+, and 8 is the highest score used. For 10 to 20 handicap, 7 is the highest. Below 10, double bogey is the highest. To figure the handicap, the course rating and slope rating are needed. The CR is what a scratch player would expect to shoot. The SR is not the bogey rating, but it is number that represent difficulty for a bogey golfer. Both are generated previously by professionals.

Then the following formula is applied: (Adjusted Score minus Course Rating) x (113 divided by Slope Rating). This gives a differential which factors in both how well you played and the difficulty of the course. To get the final handicap, the BEST 10 of the most recent 20 differentials are averaged, multiplied by .96, and then truncated to one digit. If you have less than 20 rounds, there is a schedule for how many rounds you can use. But you don't even get a handicap until after your 5th round, and then it uses only the best of your scores. I think this is to "cheat-proof" the system. Using only the best scores, you can see how different this system is from an "average". However, when I calculate 80% of my scoring average (calculated from the par average), the number equates fairly well with the handicap. For some players, this would probably be closer to 75%.

Back to K-Ratings and 1880 Doubles. The system is intended to mimic PDGA ratings, is it not? How does the PDGA calculate the ratings then?
http://www.pdga.com/discussion/showthread.php?t=24143
Chuck's ratings guide link on the second post down no longer works. But check out the BeastMaster's "Ratings Predictor" spreadsheet on the forth post down.
If I understand it correctly, it works like this.
1) Don't use round ratings over a year old unless the year has less then 8 rounds.
2) Fill in your rounds and calculate the average.
3) Calculate the Standard Deviation. http://en.wikipedia.org/wiki/Standard_deviation
4) Take 2.5 times the SD to find the Range.
5) Subtract the range from the average to find the MINIMUM ALLOWABLE Round Rating.
6) Double weight the most recent 25% of the rounds.
7) If not all the rounds from a certain date are to be doubled, double the one with the highest rating.

When looking over the K-Ratings and the PDGA ratings for the same players, the K-Ratings appeared to average about 10 points higher per player. I didn't follow the process all the way to the end, I think Brandon may have made a 10 point adjustment for this. Some different thoughts on this, but for it to be this consistent and prevalent, something was different. It appears to be the throw-out round factor. The "range" for an average 950 rated player might be about 70 rating points. The more a players score varies, the higher the standard deviate and range. So some players will have a higher or lower number, but in general, the range gets higher the lower the player rating is. You can see the PDGA throw-out rule is a little tougher than the ACDGC's. 70 rating points or roughly 7 strokes. Could be 5 strokes or less for a good pro and your round is thrown out.

Doing all of that for each of 250 players would be insane. Better just to give a 10 point adjustment.

Don't know what the moral of the story is. Don't stay up so late? I do support Brandon's decision to use the numbers as much as possible. As he said, if he adjusted the numbers for the players he did know, it wouldn't be fair for the players he didn't know. If you start fudging the numbers, at some point you got to ask yourself why you even bothered if someone thinks their made up ratings for a given person are better. In the case of J.D. and Casey, you can see the PDGA doubles the most recent 25% of rounds. That's another factor to consider with the K-Rating system. But given the acceleration level of their improvement and the cut off date, I don't think it's possible for the rating to catch up with their current skill. An easy example to see, but the same is true with anyone else to varying degrees.

[Schoen-hopper]

That may have sounded right, but I just realized (again, after forgetting the first time we ran through this) that since the non-adjusted K-Ratings are generally higher than the corresponding PDGA ratings, removing the scores that are too high (round rated too low) would just worsen the problem. While it's not a perfect system, it would seem the K-Ratings tendency to be higher would help to prevent underrated players from teaming up with an advantage.

[Swede]

I say the rating should reflect your current level. We all know JD is playing at a 970 level now.

We do? JD's average rating for his last two PDGA and KDGA events is 951. If you just take his last KDGA and PDGA events, his average rating is 934. I think there is a difference between potential and an average. The K-Rating is an average.

If the TD has the power to give a rating to unknowns based on feedback from those who know a player, he should also have the power to assign a current rating.

The whole point of the K-Ratings is to be objective--not subjective. The only time other people are consulted about a players ability is when there is no data available. The alternative to that would be not letting anyone play 1880 that doesn't have any KDGA or PDGA rounds in a season.

Throw out all the low rounds, even if it is almost half as would be in my case.

JD's Doo Dah went like this; 922, 922, 921, 991. Are you saying we should throw out all the 920's and only keep the 991? His lowest rating is a 901, his highest is that 991. Statistically, you'd be more likely to throw out the 991 and keep the 920's.

Are you advocating we only keep the top 25% or so of ratings? So you'd have a rating reflecting the players best instead of their average? I'm sure that is doable. Though I'm not sure it is any fairer...

At the G-master level there is no incentive to score well once Harv has you beat. I can bring a 950 game if needed, as can many of those between 900 and 925. Your K-ratings for G-masters show a range from low 900's up to 950ish, but we all are the same. Just a difference of where and who we played against.

I should admit that I don't play this way. I am extremely competitive and it pisses me right off to play bad. My low scores are embarrassing to me. I don't think I'm alone on that. But that might explain why we are bringing different perspectives and expectations to the ratings. I tend to look at the average rating as an ability indicator. You are looking at the top end.

My take is that numbers don't show desire. They don't show opponents, or moods, or bad breaks. Some players are consistent, some are erratic. Some blow off the rest of a round because they are pissed or don't care, some don't have the ability to get their head back in the game so they spiral downward. If there is an equation that will take all that into account, I will use it. But I don't want to be in the business of deciding what type of player someone is and therefore which scores should be kept or thrown out based on my personal knowledge and feelings.

Harv and Rich missed out by 2 pts. from qualifying as a team and they would not have any advantage over the others.

They certainly weren't alone. But I don't understand why you mention it. You think they should've played? Or you think Rich was one of those G-Masters with too low of a rating? Did they want to play? Because all they had to do was ask. Rich only had 6 rounds.

How about using your powers to assign players into pools based on all factors, K-rating, personal knowledge, division played, and feedback. Starting at 970 up would give you about 20 top players. They could only pair with someone in the 920ish and lower pool. Whether to have 1 or 2 middle pools could be decided, but they could mix in any combo. The TD would have say on all teams.

This is somewhat similar to Russ's suggestion. The advantage being that unrated players would need to be placed in a range of ratings instead of assigned a specific, accurate rating. What other advantages does it have? One downside would be that someone like a Converse (or even an EMac if he got a wild hair) could have a similarly skilled partner as a Ben Maass or even a Harvey Barger. Also, everybody near a pool division (folks near 920 or 970 etc) would feel they were really 2 points worse than their rating and really outa be in the lower pool. If I ask Schoenny about an unknown player's league results and he concludes the player is about a 920, which pool am I supposed to put him in? I see the opportunity for even more people to be unhappy with my decisions. I don't want those kinds of decisions. I really only had to make one decision this time around. It was a decision I thought was obvious and there was still a huge stink.

I'm curious if anyone knows enough about math/statistics to point me (or Adkins??) to an equation that would remove statistical outliers like Smitty suggested. I know they do this for the PDGA ratings but I don't know how to implement it. I know there were some bizarre scores especially from a tourney like Cherry Street where you play fewer holes/round and there is a bunch of O.B. But I'm not going to go through by hand and pitch scores that don't fit. Just getting the numbers there was enough work.

[Swede]

ah yeah, Mike! Standard Deviation, Yo. Mean, median and mode, Yo.

Seriously, though, Standard Deviation and Range are what Buddy and I are talking about, I think.

Mike, I actually already reduce the non-PDGA rounds some because they were so much higher than the PDGA rounds.

[Schoen-hopper]

Though some handicap systems throw out the high and the low, the PDGA system doesn't throw out any high rounds. Low rounds are thrown out only if they fall below the specific players range, which gives them a little more wiggle room if they tend to be erratic.

In my opinion, the 980+900 team has the slight advantage on the 940+940 team in standard best shot doubles. They are often able to produce a better drive. 2 putts at birdie are better than zero, and with a great putter on the team, you also have a better chance of making birds. But the other formats strongly favor the 940 duo. I think these other formats should be used more sparingly and selectively for 1880. A divisional doubles tournament would be a different story. Doesn't matter what kind of curveballs you throw; everyone knows what they signed up for.

[Schoen-hopper]

http://www.pdga.com/discussion/showthread.php?t=24143

Check out the beastmasters rating predictor file. I believe it can be written over. Just plug in the rounds and round ratings. Follow a few steps and you should get a PDGA rating replica. More difficult that the K-ratings method, but this should produce a more accurate rating.

[smitty]

My last comment on the ratings and format for this event.

The event was Awesome and all there had a great time. The formats were fun and different.

Things that could be changed might include: The holes that different formats were played on or alt. tee pad for players under a specific K-rating.

Someplace mentioned above there was discussion about raising the Combined Rating (I think??). This would be a bad idea IMO. If anything the rating could be lowered thus including more lower rated players in the event.

I will play this event no matter what formats are used and how it is changed.


Last thing:

Double Birdie Doubles would be fun to add also.

A and B both drive on a hole and are playing best shot. If both can birdie from their own drives they can go for a double birdie or 1(Much like switch+). The difference comes in the fact that if the first person to putt (near/far doesn't matter) would happen to miss, the partner can putt from the same lie and try to make the 2 (playing best shot instead of DBD). If however the second putter misses there is nobody to cover the miss and the team will take a 3 or worse.

Best Disc is a good format also. Where both people play the hole solo and take the lower score. This would be a good use for alt. tees also.

[b]

Sorry for not being clear on my thoughts. A lot of late night rambling. I do want to be clear that I do not think JD and Casey were wrong or unfair in any way. They played great and were within the TD's guidelines. I do stand by my opinion on JD's current playing level. Give him the Most Improved award.

My thoughts on G-masters ratings were to support the pool system. The ratings reflect performance on only those courses one played. In Harv's case that was Mulvane and Mac. PDGA events are usually played on much harder courses; easier to get lower rated rounds. Some players only play a few local events at known courses, resulting in inflated ratings. While the K-ratings are great tool for those who play regularly, the TD needs input from local knowledge to establish unknown ratings, as was the case this year. I believe you should have the power to OK any proposed team and not be restricted by a number. Instead use it as a guideline. I know you don't want that authority, so make it a joint decision with those who know. You trusted Smitty and Schoen., and there are others in the state who could chime in on their areas.

As for Harv and Rich, I don't think they were available, but would you have made an exception to your own rule if they had asked. With only a few rounds on the year, Rich dropped 5 pts at Mulvane, but Harv gained a few. Knowing that they could have tanked a few strokes and qualified for 1880 is not an ideal situation for the integrity of the game, but would have been possible under strict guidelines. I say grouping players not only by a rating but known abilities into pools and calling them equal is as fair a penalizing a few because of a couple of pts. I don't think any 2 G-masters teamed up would have an advantage, but the ratings would have prevented several possibilities. There were probably players in other division in similar situations.

[pironix]

I'm curious if anyone knows enough about math/statistics to point me (or Adkins??) to an equation that would remove statistical outliers like Smitty suggested. I know they do this for the PDGA ratings but I don't know how to implement it. I know there were some bizarre scores especially from a tourney like Cherry Street where you play fewer holes/round and there is a bunch of O.B. But I'm not going to go through by hand and pitch scores that don't fit. Just getting the numbers there was enough work.

So... below is the suggestion I think you are talking about from Smitty.

The best thing to do IMO would be to drop rounds that were several rating points below your rating. Drop that round were the person knows they are out of the cash, so they are trying to ace every hole or throw in that sport center highlight putt.

Those rounds aren't really an indicator of the persons skill level.

Don't know what the magic number is? I had a PDGA round that was 90 points lower then my current rating and it didn't count. Also, you most current rounds are the most accurate judge of what your rating really is. Maybe only use your last 10 or 12 rated rounds. ????

We would probably have to do like the PDGA does and have set dates where the rating is set in order to make this work. Otherwise it becomes kind of circular (although doable). Calculating an average including the outliers and then recalculating to exclude those off of the original average.. not really sure if what would get us the results we want. It's something we could try out though.

For the hell of it.. I threw together the calculations when we dropped a person's lowest round. If they shot more than 1 round at this rating, then all of these rounds were dropped. If the person has no rounds above that rating (only 2 rounds and both are at that rating), then I just took an average of what they had...

Below is the # of rounds available, old rating, and new rating (with lowest dropped) per player... You can see the spreadsheet https://docs.google.com/spreadsheet/ccc ... HZiWFR2YVE, on average people jumped up 13 points.... median was down around 9 so it just inflated the ratings as was expected.

[Swede]

Adkins, I'm getting errors on all the formulas.

Keep up the fiddling though. You clearly have a better grasp on this stuff. I'm a fish out of water with a spreadsheet. Brittany was fielding a lot of phone calls while I tried to set that thing up.

As for everything else...

I won't lose any sleep if the the K-Ratings stay just as they are. They are more than accurate enough to handicap 2 tourneys a year. They do the job for 85%+ that will play these events. As for the folks without one, I'm glad you are coming but please understand I'd rather error on the side of caution than give you an advantage over the folks that play the tournaments.

There won't be pools.

There will be leniency for those with few rounds (as there was this year had anybody asked).

There will continue to be no leniency for healthy folks with a bunch of rounds. It is a cap, not ballpark.

If the ratings cap gets raised, it will be accompanied by a second division-- 1910 and 1820 for example.

I'm leaning though toward giving one more year to the 1880 while trying some of the things Smitty and Schoenny suggested to get the top players competitive. (alt tees on certain holes for sub 900 players, different hole/format pairings, less Alternating shot, maybe a bit more Best Shot)

[Ruder]

Good Work Swede.

[pironix]

Adkins, I'm getting errors on all the formulas.

Lame... apparently google docs doesn't support the countifs and averageifs functions. Sheet 3 has the values though so you can compare them there...

[Swede]

Adkins, I'm getting errors on all the formulas.

Lame... apparently google docs doesn't support the countifs and averageifs functions. Sheet 3 has the values though so you can compare them there...

try looking for extra parenthesis. for some reason converting Excel to Google gives extra "order of operations" functions. Removing one extra set of parenthesis might fix all the errors at once.

[Schoen-hopper]

I'm able to view it okay with Firefox. Explorer is no good for Google Docs. I need an explanation on the columns though. Can you hide the unneeded columns for easier viewing?