OT - Do handicaps travel fairly?

[Wayne_Kozun]

I agree about 113 vs 100.  Here is something I found on the 113 as the "standard" or average slope rating  http://www.ongolfhandicaps.com/2014/07/why-is-113-in-slope-handicap-system.html

[Ken Moum]

I think your assumption is probably right.


But am also pretty convinced that there's not enough spread in bogey ratings, which means there's not enough difference in the slope between "hard" courses and "easy" ones.


As a result,  players at those easy courses will never get enough strokes from the ones who play hard courses.


Doesn't affect scratch golfers,  but two 15s...

It is not.


Let's take my home course, Arroyo Trabuco, vs Merion Golf Club. Rating is the same from the member's tees (71.5 vs 71.4) but the slope is 128 vs. 145, as compared to the 'standard bogey' slope of 113.


This means the 'bogey' rating is only 2.7 shots different at 92.4 [par 72] vs 93.1 [par 70]. If you assume ATGC's slope to be reasonable vs. the 'standard bogey' 113 (I would), then Merion's slope should really be around 160-165. It's not 2.7 shots harder for the bogey golfer, it's 8 or 9 shots harder.


I think some handicapping travel issues can be solved quickly by correcting the slope on the high end, thus expanding the whole ladder [and because I'm petty, return the baseline to 100 instead of 113 please]

That's exactly what I was talking about.


Small sample size, but I am currently in Fife and have played Leven Links and St. Andrews Strathtyrum in the last few days. 


My index is currently 18.7 so my course handicap on them was 11 on Strath and 23 on Leven. Twelve strokes difference sounds like a lot, but I actually shot 18 strokes more on Leven, and If I'd been playing medal competition and been required to finish every hole, it would have been more like 24 or 25 shots.


There's only one stroke difference in par, the ratings are 63.4 and 71.4, which seems reasonable enough.


But the slopes are 101 and 129.  So I agree that the change needs to happen at the top.


Another example is Jeff Brauer's Colbert Hills in Kansas.  I've played it a few times, with varying levels of players, and tbh the slope there doesn't begin to address how hard it is for bogey golfers.


In fact, when it was fairly new, one of my coworkers at GCSAA talked to a rater who said something like, "The slope is 152 and it should be higher."


My first thought was, then whyinhell isn't it?

[Thomas Dai]

Can't speak for other countries and how the WHS has been introduced but I'm not convinced of the quality/accuracy of course ratings as undertaken in the UK. Seem to be a few local oversights. Perhaps revisions will be made over time or new assessors introduced.
atb

[A.G._Crockett]

I think it's safe to say that this thread is hitting nerves on a bunch of my pet peeves.

Basing one's evaluation of the validity of the course rating and especially the slope rating on one or two plays of the course represents a HUGE misunderstanding not only of what a handicap index measures, but also the very purpose of the handicap system itself.  Not only is your index an expression of your potential, rather than any sort of a predictor of your score on a given day, but the purpose of the system is to allocate strokes for competition, whether formal or informal. 


So IF the course or slope ratings aren't accurate (a premise, btw, that I do NOT accept!) it doesn't really matter because ALL players are impacted on a given day are affected by the rating the same way.  If I am only allocated an away course handicap of 5 at Bushwood from the white tees instead of the 8 that I think I should be getting, it doesn't matter, since my opponents/fellow competitors are impacted exactly the same way.

An additional mathematical reality about golf is that scratch golfers and very low indexes tend to have their scores in a relatively tight range, which higher indexes shoot scores that vary MUCH more widely, which is the main reason that the slope system was devised in the first place.  So when a 15 or higher index (or really even a 10, for that matter) goes to a course and posts a good score relative to their course handicap, and then goes to a second course with a similar slope rating and the same course handicap and shoots a million, what has happened was mathematically completely predictable, and in NO way a commentary on the validity of the slope rating of either course.

When a golfer says, "I'm a 12, and my course handicap was a 10 at Bushwood, but I shot a 93; the slope rating there is MUCH too low!", he has set up a whole bunch of straw men and then knocked them down.  His real comment should be this:  "I shot a 93 at Bushwood with a course handicap of 10, so I played poorly.  However, if I play Bushwood from those same tees 19 more times, I'll probably have 4 (or so) good rounds where I actually play to the 10 course handicap.  For the other 15 rounds, I won't play to the 10 because that's just who I am as a golfer.  Which is to say, while I DO catch lightning in bottle every now and then, I'm not particularly good on my "average" day, and I really suck on my bad days."

That, of course, is NOT how we like to talk about ourselves, so we construct false narratives about handicaps not traveling well, or course/slope ratings being wrong, and so forth.   

Think your handicap doesn't "travel well"?  Travel more and get an index that reflects that.  Think the slope rating MUST be wrong because you didn't play to your course handicap?  Instead of blaming the slope rating and your course handicap, play that course more (like 19 more times!) to get an accurate picture. 

[Sean_A]

AG, I understand and sympathise with your PoV. It's easy to blame the system, but I know some players have been playing these courses for years. Remember, the system has only just been introduced in GB&I and it's been a bit of a mess to be honest. I too am convinced that some of ratings are well wrong. Those ratings were previously balanced by the standard scratch score changing depending on conditions. It was very common for the sss to go up at least one shot. Now, it is very rare for that equivalent number to budge at all. Which means, the ratings need to be adjusted. I suspect this will occur over time.

Ciao

[Tony_Muldoon]

Many UK clubs play Matchplay amongst themselves.


Home wins occur perhaps 75% of the time.


That doesn't fully reflect how each clubs handicaps travel to a variety of courses, but it's a hint that home advantage is definitely worth a shot or two.

[Thomas Dai]

Remember, the system has only just been introduced in GB&I and it's been a bit of a mess to be honest. I too am convinced that some of ratings are well wrong.
Ciao

+1
atb

[A.G._Crockett]

Many UK clubs play Matchplay amongst themselves.


Home wins occur perhaps 75% of the time.


That doesn't fully reflect how each clubs handicaps travel to a variety of courses, but it's a hint that home advantage is definitely worth a shot or two.

Speaking as an interclub captain myself with a huge data set to draw on from the Carolinas Golf Association, I’d guess that 75% is probably a little bit high, but not much.  Throwing out ties, the home team wins right at 2/3rds of the matches among over 300 teams in three divisions of CGA interclub competition.  My own team went 3-0 at home, but we were 0-2-1 in our away matches.  Home course advantage is so significant that the CGA interclub playoff matches, consisting of four net four ball matches, are played with two matches going on simultaneously at both courses, and the semifinals and finals are played at a neutral site.  (In the Carolinas, we’re quite fortunate in this regard; the “final four” interclub matches are played on back to back days at Pine Needles and Mid Pines, adjacent to CGA headquarters.)


Home course advantage is real, and we all know that, which why we play practice rounds at away courses before tournaments and even Interclub matches.  This is maybe even a bigger deal now even than it used to be, since some courses (like mine) still have bent grass greens, while others have converted to one of the Bermuda varieties.


But it is a far, far different thing to acknowledge that home course advantage is real vs this thing about how well handicaps “travel”.  During the regular season, I get as many guys in the lineup as often as I can, but I keep track of net differentials for each of my players all season.  Then in the playoffs, the lineup, as much as possible, is the guys who play closest to their handicaps, whatever those might be.  It’s a net competition, so I don’t care if a guy is scratch or a 15 if he can play close to that under the gun.




Wanna guess what those guys have in common?  They tend to play lots of away rounds, and they tend to play a lot of competitive golf fully under The Rules of Golf.  Both of those things make their indexes at least marginally higher, but more importantly it makes it more likely that they’ll perform well under competitive pressure. 




Again: Handicaps that “travel well” (or don’t…) have ZERO to do with what club you belong to, or how hard or easy your home course is; it is completely individual.  And the (slight) possibility that a given course might be somehow incorrectly rated is a complete red herring, since the ratings apply equally to ALL players in the competition.

[Padraig Dooley]

The skill of green reading is a big factor in this, the vast majority of golfers are poor at reading greens, at their home course they can and do putt on memory, take them to a different course they will struggle to read greens and the score will go up as a result.

[Wayne_Kozun]

Agreed, although having a caddie or member giving you advice at the away course can help offset that.

[archie_struthers]

 


Great topic and lots of good replies. Some courses play harder for higher handicappers based on the design. For instance Stone Harbor here at the Jersey Shore is a lot harder for high handicaps because the hazards impact the really bad shot more. If you can drive it over 250 many of the landing areas are bigger. It's kind of reverse engineered to favor the strong player. There aren't many subtleties in the chipping and putting game to test the expert, just severe carries and severe hazards flanking the holes. I have found a 10-15 there tends to be really good relative to other 10-15's in the same geographic area. They just make way fewer doubles away than at home. 

[Ken Moum]

Archie, that's what I've tried to say.


If it's true, then the slope system doesn't work as well as I think it should.

[Garland Bayley]

I don't know that you need AI/ML to do this.  Assuming that you had the data you should be able to do this using more traditional stats measures like regressions.  If you have hundreds/thousands of scores from a course with golfers of varying abilities then you can calculate what the slope and stroke rating should be.  Presumably the stroke rating is the average that scratch golfers shoot on the course for the given tee.  And then look at how scores go up with a golfer's handicap.
I wouldn't be surprised if the USGA already does this.  Once you build the code it should be pretty easy to run for each course, assuming that you have access to the database of golf scores and golfers.

Problem is that is just a one dimensional average that doesn't account for differing styles of playing.

For example a wild Willy (Knuth's term) will set a lower index playing a longer than normal set of tees while steady Eddy will set a lower index playing a shorter than normal set of tees.

Simply calculating ratings from masses of data won't change that.

Moral of the story, get Willy to play against you on the forward tees, and Eddy to play against you on the back tees.

[Matt Schoolfield]

Archie, that's what I've tried to say.


If it's true, then the slope system doesn't work as well as I think it should.

Yea, the slope system is linear slope. A more accurate system would take convexity into account. I think calculating a curved handicap with three points, including a double bogey rating should facilitate that where it is most needed. I think the biggest variable that’s ignored is variance (I’m presumed that a normal distribution is assumed), but that gets into needlessly complex math.

[Ken Moum]

For example a wild Willy (Knuth's term) will set a lower index playing a longer than normal set of tees while steady Eddy will set a lower index playing a shorter than normal set of tees.

Simply calculating ratings from masses of data won't change that.

Moral of the story, get Willy to play against you on the forward tees, and Eddy to play against you on the back tees.

This is totally true.  I am VERY sensitive to length, but I have always been a short hitter with a good short game.


I don't think it's a big factor in whether there's enough slope in the Slope System.

[A.G._Crockett]

Ok...

To (re)state what should be the very obvious, the purpose of course and slope ratings is to allocate strokes in net competitions, be they formal or informal.  As such, it doesn't really matter whether you think a rating is "fair" or not, or too high or too low, since it impacts all players in the competition exactly the same.  Whether or not a given player can play to his course handicap, home or away, is an individual matter that has nothing to do with any comparison between his course and the away course.  GHIN takes care of that, and IMO does so remarkably well.


Further, there seems to be a premise, at least by some here, that rating teams are much more subjective, or hit and miss, in their work than is actually the case.  Rating teams are not only made up of multiple persons, but they are generally across the spectrum of golfing ability.  They also work off of very carefully devised worksheets into which they plug numbers to make an inherently subjective topic (How easy/hard is this course?) into the most objective result possible.  On top of that, most courses are re-rated in every 5 year period, but AT LEAST every 10 years, so a mistake made in one rating isn't likely to persist.

Elephants leave tracks, so if you make the very large claim that lots of courses are incorrectly rated, or that handicaps from "hard" courses travel better than handicaps from "easy" courses, provide data and links so we can all read about that.

[John Kavanaugh]

I am so much older and fatter and my handicap is consistently 2 strokes, 30% to 50%, lower because I moved to an easier course. I get that Victoria National is an abomination but that doesn’t make you any less not correct.

[Garland Bayley]

... it doesn't really matter whether you think a rating is "fair" or not, or too high or too low, since it impacts all players in the competition exactly the same.  ...

Nope

As the Pope of Slope has indicated, you move a wild Willy back he gains advantage, forward is a disadvantage. Steady Eddy forward advantage, back disadvantage.

[A.G._Crockett]

... it doesn't really matter whether you think a rating is "fair" or not, or too high or too low, since it impacts all players in the competition exactly the same.  ...

Nope

As the Pope of Slope has indicated, you move a wild Willy back he gains advantage, forward is a disadvantage. Steady Eddy forward advantage, back disadvantage.

I look forward to a link, so that I can read Dean Knuth explaining how the system HE devised does not work properly.  Thanks in advance.

[Matt Schoolfield]

I look forward to a link, so that I can read Dean Knuth explaining how the system HE devised does not work properly.  Thanks in advance.

It’s on his website, note that HRT means Handicap Research Team:

8. Future Research: The Player Variability Problem
The HRT has shown that 80% of all bogey golfers fit the model within one stroke. However, two distinct types of golfers fall outside of these limits. These types have been labeled "Steady Eddy" and "Wild Willy." Steady Eddy represents 12% of all golfers and he is a very straight, but short ball striker who has an outstanding short game. When taken from a short course to a long course, his score increase8 greater than the model would show, thus he is under-handicapped at a high Slope course. Conversely, Wild Willy is a long-hitter, but is inaccurate. Representing 8% of bogey golfers, this type can be over- handicapped on a long open course, but under-handicapped on any very tight and punitive course.
The HRT is considering a solution of adopting a normal model handicap formula which would mesh a two dimensional handicap to the Slope System. The solution could result in a Steady Eddy receiving more strokes on a high Slope Rated course than a Wild Willy of equal Handicap Index would receive.

http://popeofslope.com/courserating/twoparameter.html




This is exactly the type of problem a collaborative filtering algorithm can solve. Simply because it can group Steady Eddys and Wild Willy’s into their own categories.

[Ken Moum]

AG,  I'm not suggesting that rating teams are making errors on lots of courses, apart from my example of one actually saying the slope at one course "should have been higher."


My contention is that the SYSTEM doesn't have enough variation from high slopes to low slopes.


A few uears ago I had a pleasant conversation with a former member of the USGA Handicap Committee who admitted that players from low slope courses are generally disadvantaged in competition against players from high slope courses.


It doesn't show up with slopes I commonly see in the 125-135~ range.

[John Kavanaugh]

Steady Eddie, Wild Willie, Pope of Slope…I wish they didn’t talk to us like children.

[A.G._Crockett]

I look forward to a link, so that I can read Dean Knuth explaining how the system HE devised does not work properly.  Thanks in advance.

It’s on his website, note that HRT means Handicap Research Team:

8. Future Research: The Player Variability Problem
The HRT has shown that 80% of all bogey golfers fit the model within one stroke. However, two distinct types of golfers fall outside of these limits. These types have been labeled "Steady Eddy" and "Wild Willy." Steady Eddy represents 12% of all golfers and he is a very straight, but short ball striker who has an outstanding short game. When taken from a short course to a long course, his score increase8 greater than the model would show, thus he is under-handicapped at a high Slope course. Conversely, Wild Willy is a long-hitter, but is inaccurate. Representing 8% of bogey golfers, this type can be over- handicapped on a long open course, but under-handicapped on any very tight and punitive course.
The HRT is considering a solution of adopting a normal model handicap formula which would mesh a two dimensional handicap to the Slope System. The solution could result in a Steady Eddy receiving more strokes on a high Slope Rated course than a Wild Willy of equal Handicap Index would receive.

http://popeofslope.com/courserating/twoparameter.html




This is exactly the type of problem a collaborative filtering algorithm can solve. Simply because it can group Steady Eddys and Wild Willy’s into their own categories.

Thanks for the link.  I’ll freely admit that the term “collaborative filtering algorithm” is WAY past my math knowledge base.  I have no idea how it would be workable, but there is much I don’t know.




But this is another question entirely from the question of whether or not handicaps “travel fairly”.  I don’t think anyone would dispute the fact that not all courses favor all players equally, but that in no way speaks to the efficacy of either slope ratings or whether or not handicaps in general travel well.




I think it’s time for me to stop on this thread since I’m just repeating myself now.  One last time: In every case that I’ve known of, guys who complain about the handicaps from their club not traveling well either don’t play enough away rounds, and/or don’t play full under the Rules.  In either case, the result is a lower handicap, and one that they struggle to play to, even once out of every 5 rounds.

[A.G._Crockett]

AG,  I'm not suggesting that rating teams are making errors on lots of courses, apart from my example of one actually saying the slope at one course "should have been higher."


My contention is that the SYSTEM doesn't have enough variation from high slopes to low slopes.


A few uears ago I had a pleasant conversation with a former member of the USGA Handicap Committee who admitted that players from low slope courses are generally disadvantaged in competition against players from high slope courses.


It doesn't show up with slopes I commonly see in the 125-135~ range.

Even if this were so, and I see NO evidence of that, so what?  The purpose of slope is to more accurately allocate strokes, and if the rating on a tough course is too low, it’s too low for every player. 




If your point is that individual players from low slope courses are somehow at a disadvantage at high slope courses, I don’t accept that, simply because they get more strokes than they get at home. 

[Ken Moum]

My argument is apparently going over your head, I'm probably not being clear.


I'll restate it.


I'm suggesting that, for instance,  a 15 index player from a low slope course is a much less skilled player than a 15 index player from a high slope course.


Regardless of where they play,  they'll have the same course handicaps.


As has been suggested here, the problem is that the rating template underestimates how hard some golf courses are for bogey golfers.


Of course the guy from the easy course gets more shots on the hard one.


But if he played there all the time, I guarantee his index would be higher.


That should not be the case.