or Connect
TheSandTrap.com › Golf Forum › The Clubhouse › Tour Talk › OWGR Biased Against PGA Tour Players
New Posts  All Forums:Forum Nav:

OWGR Biased Against PGA Tour Players

post #1 of 65
Thread Starter 

 

http://online.wsj.com/article/SB10001424052702304636404577299702668540234.html?mod=WSJ_hpp_MIDDLE_Video_Third

 

Quote:

At their presentation in Phoenix, summarizing research for a paper to be finished soon, they compared the world ranking of the top 200 players to a ranking of those same players' skill levels calculated using a statistical model they also devised, which simultaneously takes into account players' adjusted tournament scores and the difficulty of the courses. In every two-year period going back to 2003, the bias was stark. "For every given skill ranking, the official world golf ranking for PGA Tour players averaged 36 positions worse than for non-PGA Tour players," Rendleman said. At one point Pat Perez was ranked 95th in skill, according to their model, but 195th in the world rankings.

 

One of the main flaws in the current system involves strength-of-field ratings used to determine how many points players earn for good performance. Tournaments receive a minimum point value, depending on which of the world's tours is sponsoring it, regardless of how many quality players are in the field. Strength of field ratings based on the participating players' skill levels would be more equitable, they argue.

 

post #2 of 65

Isn't this part of the reason that some folks want to see a unified world tour? OWGR is a crazy beast.

post #3 of 65

The thing is, it's pretty tough to compare European Tour players with PGA Tour players when they aren't competing in the same tournament. Many will say the European Tour is much easier, but the reality is you are travelling to a different country most weeks, with completely different courses/culture/food/climate. The proof is in the pudding when some US players go over and can't make the cut or finish poorly. For example the Scottish Open last year, which many go to play the week before the Open. Gary Woodland MC, Brandt Snedeker MC, Ryan Palmer T58, Phil Mickelson T58, Matt Kuchar T10. I'm not saying the European is better than the PGA Tour, because the majority of the worlds best players play on the PGA Tour (where the money is), but it's not a cakewalk either. 

 

The other point with their system is, how do they decide what a player's skill level is? You cannot do that when players play completely different courses each week in different weather. It's not comparing apples with apples.

post #4 of 65
Quote:
Originally Posted by iacas View Post

"Tournaments receive a minimum point value, depending on which of the world's tours is sponsoring it, regardless of how many quality players are in the field. Strength of field ratings based on the participating players' skill levels would be more equitable, they argue."

Interesting article, but it sounds like a couple guys wanted to get their name in the paper, so they came up with their own system that nobody cares about. The minimum points thing is not noticeably unfair to US players --- minimum points on the Japanese, Asian, and Korean Tours are 16, 14, and 6, respectively. 14 is the same as the Nationwide Tour, and a PGA event has a guaranteed minimum of 24, but the only PGA events where the winner actually gets that few points are either Fall Finish, or alternate events played the same week as a WGC. Most regular season PGA events give at least 40 points; even Tiger's 18-man Chevron Challenge, with most of the top ten absent, got 44 points.
post #5 of 65
Quote:

Originally Posted by michaeljames92 View Post

 

Many will say the European Tour is much easier, but the reality is you are travelling to a different country most weeks, with completely different courses/culture/food/climate. The proof is in the pudding when some US players go over and can't make the cut or finish poorly. 


 

The Euro Tour regulars all face that, it doesn't make the tournament harder for a euro tour member to win.  The poor performance of top US players is likely because they are not used to the differences and have a harder time acclimating.  That's evidence of an additional challenge facing US players, not greater talent on the euro tour.

 


 

It sounds to me like they take the scores and weight them according to the course difficulty, just like a handicap system.  That sort of analysis seems logical and unbiased to me.

post #6 of 65

 

The OWGR is not terrible, but it could be improved.  I don`t know what criteria these guys used to determine "course difficulty" but unless it correctly includes weather, condition & firmness of the greens, pin positions, length of rough, etc. then it could have a bias as well.

 


Quote:
Originally Posted by dsc123 View Post


 

The Euro Tour regulars all face that, it doesn't make the tournament harder for a euro tour member to win.  The poor performance of top US players is likely because they are not used to the differences and have a harder time acclimating.  That's evidence of an additional challenge facing US players, not greater talent on the euro tour.

 


 

It sounds to me like they take the scores and weight them according to the course difficulty, just like a handicap system.  That sort of analysis seems logical and unbiased to me.

True, every player may face these same difficulties, but if it raises the average score for everyone by 1 stroke, then these guys would say that everyone is less talented (compared to players who play on a tour that does not face these challenges).  I agree that with jet lag, etc it is harder for US players when they first fly over to Europe, but some of the same issues face Euro and Asia based players when they fly over to the US for the Masters, US Open, PGA or WGC events that are help mostly in the States.  


 

 

I see problems with both the minimum points awarded and the strength of field determination of the OWGR, but generally think that the problem with these two tend to work in opposite directions, except with small field tournys like Tiger`s where the rich get richer by getting automatic points no matter where they finish (http://thesandtrap.com/t/54475/chevron-world-challenge-2011-thread/72#post_664686).   

 

I`ve looked at this a bit recently and think that instead of determining the strength of field by some predetermined point value based on a player`s OWGR (i.e. #1 brings 45 strength of field points, #2 34, #3 32, #16-30 11, #101-200 1 and #201+ 0) could be improved upon.   Seems kinda silly to me that the total number of players in a field is not given some weight.  IMO, if it is a small field, it is going to be easier to beat 1 #30 than 10 101 to 110, but the solo #30 brings more points.  I think they should do something based off the collective OWGR "average" points of each player in the field.  The gap between players is not always the same and, if set up properly, this would do a better job of determining the real strength of field without needing to have a point minimum.  If there is a bias against US Tour players, this would help eliminate it over time.   

 

Of course, it would also help if they correctly computed weighted average.  The OWGR are mathematically flawed with how they compute weighted average as they fail to weight adjust the denominator the way they do the numerator.  This leads to a bias against players who played a lot of tournaments towards the end of 2 year look back period.

http://thesandtrap.com/t/46168/2011-players-championship-qualification

http://www.mattelston.com/2011/04/27/world-golf-rankings-mathematically-flawed/

http://thesandtrap.com/t/46946/any-golfers-good-at-math

 

post #7 of 65

 

Quote:

At one point Pat Perez was ranked 95th in skill, according to their model, but 195th in the world rankings.

 

I`ve had dinner with Pat Perez and like the guy, but his performance in Majors and the Players hasn`t really set the world on fire.  This leads me to believe that there could be problems with their computation of "skill ranking" if he is one of the best examples that they can come up with.

 

http://en.wikipedia.org/wiki/Pat_Perez

 

Tournament 2002 2003 2004 2005 2006 2007 2008 2009
The Masters DNP T45 DNP DNP DNP DNP DNP CUT
U.S. Open CUT DNP T40 DNP DNP CUT T36 DNP
The Open Championship DNP DNP DNP T67 DNP T20 CUT DNP
PGA Championship 70 DNP DNP T6 CUT T18 T58 CUT
 

 

In the last 5 Players Championships, he has 3 missed cuts and a high finish of t42 (but did record a t3 in 2006)

 

 

 

 

 
post #8 of 65
Thread Starter 

 

Quote:
Originally Posted by MEfree View Post

I`ve had dinner with Pat Perez and like the guy, but his performance in Majors and the Players hasn`t really set the world on fire.  This leads me to believe that there could be problems with their computation of "skill ranking" if he is one of the best examples that they can come up with.


It's not like they claimed he was a top ten candidate. 195 to 95 or something, right?

post #9 of 65
Quote:
Originally Posted by iacas View Post

 


It's not like they claimed he was a top ten candidate. 195 to 95 or something, right?


Yes, but he was ranked in the top 50 for a few weeks in 2009 and has spent about 175 weeks in the top 100 (with an average rank of about 128 since 2003)...I would guess that most guys with those credentials have more than 1 top 10 and 3 top 25 in majors, but maybe I am wrong.  

 

Hard for me to find someone with a similar ranking as Perez, but Thomas Levet of France spent a bit more time in the top 50, but only 115 weeks in the top 100 with an average rank of about 190.  Here is his major record- 2 top 5 and 5 top 25  http://en.wikipedia.org/wiki/Thomas_Levet  He has won 6 times on the Euro tour compared to 1 PGA tour win for Perez.

 

 

Tournament 1998 1999
The Masters DNP DNP
U.S. Open DNP DNP
The Open Championship CUT T49
PGA Championship DNP DNP
Tournament 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
The Masters DNP DNP DNP CUT DNP T13 CUT DNP DNP DNP
U.S. Open DNP DNP T18 DNP CUT T52 DNP DNP CUT T45
The Open Championship DNP T66 T2 T22 T5 T34 DNP DNP DNP T38
PGA Championship DNP DNP 71 DNP CUT CUT DNP DNP DNP T51
Tournament 2010 2011
The Masters DNP DNP
U.S. Open DNP CUT
The Open Championship CUT DNP
PGA Championship DNP DNP
 

 

The article didn`t explain if they time weighted "skill" similar to how the OWGR are time weighted...this could effect things also.

 

post #10 of 65
Quote:
Originally Posted by MEfree View Post

Seems kinda silly to me that the total number of players in a field is not given some weight.

I agree, but the solution is simple. They can just cut the points in half for fields with less than 60 players.

Quote:
Of course, it would also help if they correctly computed weighted average.  The OWGR are mathematically flawed with how they compute weighted average as they fail to weight adjust the denominator the way they do the numerator.  This leads to a bias against players who played a lot of tournaments towards the end of 2 year look back period.

They have already taken steps to correct that by instituting a maximum divisor. But the real problem is that they go back two years, and depreciate the weight of events all the way down to zero. Because of that stupid policy, Phil Mickelson has a lower average now than he would have if he had skipped the 2010 Masters, instead of winning it.

The fix for that is to just go back 18 months, and leave the depreciation scale the same, i.e. the last week before an event doesn't count at all, its weight is about 30%. If they did that, a major win wouldn't lower anyone's average (except Tiger's during his absolute peak, when he had an average of over 30 points).

And as I'm sure you know, the minimum divisor is biased against players who play less frequently, but I agree with that policy, since it encourages more frequent play. I would probably want it lowered to 30 from the current 40, though.
post #11 of 65
Quote:
Originally Posted by brocks View Post



They have already taken steps to correct that by instituting a maximum divisor. But the real problem is that they go back two years, and depreciate the weight of events all the way down to zero. Because of that stupid policy, Phil Mickelson has a lower average now than he would have if he had skipped the 2010 Masters, instead of winning it
Crazy, isn`t it.  I have posted about numerous examples like this before.  It would be a real shame to see a guy ranked #51 in the world not be eligible for a major because he had a high finish in the same major 2 years earlier in a situation where his OWGR would actually be higher had he skipped the event 2 years previously.  I had discussions with OWGR members, but they didn`t want to admit that they had been doing it wrong.  
The fix for that is to just go back 18 months, and leave the depreciation scale the same, i.e. the last week before an event doesn't count at all, its weight is about 30%. If they did that, a major win wouldn't lower anyone's average (except Tiger's during his absolute peak, when he had an average of over 30 points).
The better fix, IMO would be to simply weight the divisor the same as the numerator- If the 100 points Mickelson won gets multiplied by .011 (or whatever it is), then only add .011 to the divisor, not 1...that gives a true weighted average.  Reducing the look back period to 18 months would mostly eliminate the situation where a win is worse than skipping an event, but with the top ranked players, you could still get a situation where a high finish that helped their "average" for a while ends up hurting their average later on...a properly computed weighted average does not suffer from this problem.

 


 

 

post #12 of 65
Quote:
Originally Posted by MEfree View Post

The better fix, IMO would be to simply weight the divisor the same as the numerator- If the 100 points Mickelson won gets multiplied by .011 (or whatever it is), then only add .011 to the divisor, not 1...that gives a true weighted average.

Well, no, it doesn't, not if the player hasn't played for a while. It gives the absolute average, not the weighted average. You might as well not depreciate for age at all.

To see this, forget about the minimum divisor for a second, and pretend that Phil blew out his knee after the 2010 Masters and hasn't played since, so that's the only event in his WGR window.

So according to your scheme, the weight of the event is .033, his points earned is 3.3, and his divisor is .033, so his official average is 100.

I agree that for a player who plays every week, it would be better than what we have now. But for a player like Tiger, who was out with injury for almost a year, it would be even worse than what we have now for propping up a #1 ranking long after he should have fallen back.
Edited by brocks - 3/29/12 at 11:21pm
post #13 of 65
Quote:
Originally Posted by brocks View Post


Well, no, it doesn't, not if the player hasn't played for a while. It gives the absolute average, not the weighted average. You might as well not depreciate for age at all.
To see this, forget about the minimum divisor for a second, and pretend that Phil blew out his knee after the 2010 Masters and hasn't played since, so that's the only event in his WGR window.
So according to your scheme, the weight of the event is .033, his points earned is 3.3, and his divisor is .033, so his official average is 100.
I agree that for a player who plays every week, it would be better than what we have now. But for a player like Tiger, who was out with injury for almost a year, it would be even worse than what we have now for propping up a #1 ranking long after he should have fallen back.


With a proper weighted average, I would still advocate a minimum divisor (of around 22.5) so as to avoid the examples you give.  This would lead to a player who hasn`t played for a while to run into a problem with the minimum divisor sooner as his older events start to decay in value, while a player who has played a lot recently can get up above the minimum sooner.

 

post #14 of 65

I think the whole OWGR is a bit silly.  Look at how they have been bouncing around since Tiger lost the spot.  I guess I just think there is too much stock put into the #1 being "the best player in the world" when I don't think the system really reflects that.  Perhaps its how I view "the best", though. 

post #15 of 65
Thread Starter 

 

Quote:
Originally Posted by Paradox View Post

I think the whole OWGR is a bit silly.  Look at how they have been bouncing around since Tiger lost the spot.  I guess I just think there is too much stock put into the #1 being "the best player in the world" when I don't think the system really reflects that.  Perhaps its how I view "the best", though. 


I think that illustrates how close things are at the top in the absence of a dominant Tiger.

 

I think the OWGR does a pretty good job overall of ranking people within about 5-10 spots of where people would put them (if they were aware of all the facts). And that's the main point. The top ranked player is not really the main point for the OWGR at all. That player is afforded exactly zero extra benefits over the others (except seeding in the Match Play). The harsher line comes at #50 or #64 or whatever numbers are used to determine eligibility for certain events.

post #16 of 65

Going back to the OP, there was further explanation of the system Broadie and Rendleman used to calculate skill in http://www.golfdigest.com/golf-tours-news/2012-04/gwar-stachura-world-golf-ranking which was briefly discussed in a recent thread http://thesandtrap.com/t/57602/luke-donald-back-to-1#post_706070

 

After reading the Golf Digest article, I have more of an understanding of the methodology used (basically if A beats B by a stroke in Tourny 1 and C beats B by 2 strokes in Tourny 2, then C is 1 stroke better than A).  While this makes more sense than what I originally thought (which was that they basically gave each course a course rating based on distance, etc. and then used that to compare scores between players who didn`t compete against each other), I have a number of questions and see a number of problems:

 

1.  Is there any differentiation made based on tournament finish?  i.e. Say A wins Tourny 1 and beats B by 4 shots.  In Tourny 2 B finishes 20th and beats A by 4 shots.  Are these two players considered equal (same total strokes) or is A considered better because he performed better in a more meaningful situation?

 

2.  How are Playoffs handled?

 

3.  How are Match Play Tournys handled?

 

4.  Some courses may have a tendency to spread out players more than others, Is this considered?

 

5.  How reliant are comparisons between Tours on a limited number of cross over players in a limited number of tournaments?  i.e.  If a Japanese Tour regular happens to have a good showing in the US, does this boost the rankings of other Japanese Tour players by a lot or a little?

 

6.  Are certain tournys given more weight than others?  i.e.  If A beat B in the US Open by 4 shots and then B beat A in a Nationwide event by 4 shots, are they considered equal, or is A considered better based on tourny importance?

 

7.  Does the system use any time weighting like the OWGR?

 

8.  How are MCs, DQs, WDs, etc. handled?

 

I think that Broadie and Rendleman may be right that there could be a bias in favor of certain tours and that there is certainly some politics/arbitrariness in how OWGR Event Points are awarded.  However, without knowing all the facts, I tend to think the system they propose is probably better suited for handicapping or scoring average comparisons rather than as a method for determining world rankings.  I can think of several methods to improve the awarding of event points that can utilize much of the current OWGR framework without introducing a bunch of potential new problems that I see with this system.  

 

 

 

post #17 of 65
Thread Starter 

I just read (again) the article at http://www.golfdigest.com/golf-tours-news/2012-04/gwar-stachura-world-golf-ranking and think that, pending the favorable outcome of a peer review phase, it'd be really nice to have this type of thing be put into place. You could even phase it in over the course of a year or two, so as not to immediately disturb the rankings. For example: 

 

Week 1 of 2013: Old OWGR given weight of 51/52, new OWGR given weight of 1/52.

Week 2 of 2013: Old OWGR given weight of 50/52, new OWGR given weight of 2/52.

Week 26 of 2013: Old OWGR given weight of 26/52, new OWGR given weight of 26/52.

 

Etc.

 

Or they could do it over two years.

 

I think they should do it. If I understand things correctly, it'd help people who win by bigger margins, too, and slightly diminish a guy who wins in a four-way playoff (because his skill would be almost equal to those who lost the playoff). And that makes sense to me - three guys who T2 because they lost a playoff get a pretty low percentage of points these days despite, for all intents and purposes, tying the winner.

post #18 of 65
Quote:
Originally Posted by iacas View Post

I just read (again) the article at http://www.golfdigest.com/golf-tours-news/2012-04/gwar-stachura-world-golf-ranking and think that, pending the favorable outcome of a peer review phase, it'd be really nice to have this type of thing be put into place. You could even phase it in over the course of a year or two, so as not to immediately disturb the rankings. For example: 

 

Week 1 of 2013: Old OWGR given weight of 51/52, new OWGR given weight of 1/52.

Week 2 of 2013: Old OWGR given weight of 50/52, new OWGR given weight of 2/52.

Week 26 of 2013: Old OWGR given weight of 26/52, new OWGR given weight of 26/52.

 

Etc.

 

Or they could do it over two years.

 

I think they should do it. If I understand things correctly, it'd help people who win by bigger margins, too, and slightly diminish a guy who wins in a four-way playoff (because his skill would be almost equal to those who lost the playoff). And that makes sense to me - three guys who T2 because they lost a playoff get a pretty low percentage of points these days despite, for all intents and purposes, tying the winner.

I don`t see that happening even if the peer reviews are favorable.  I had a number of questions above that I still have, but my guess is the biggest problem with their system is it doesn`t give extra weight to majors and some of the other more important tournaments.  I am only guessing, but I imagine that if B beats A by 4 at Memphis with both finishing well back and A wins the US Open beating B by 3, that B would be ranked ahead of A (based on those two results) because he beat him by 1 shot overall.  My guess is the public would like to see the US Open winner (A) ranked higher even if B has a slightly better scoring ave in the same events.  

 

In my mind, their system is similar to the stroke differential against the field stat (but with adjustments to make it relevant across tours and to compensate for strength of field).  Maybe I am wrong about how their system would work and if so, please direct me to the system details that explain how this would not be the case.   

 

FWIW, Tiger is currently ranked #1 in that stat for this year- sure he has had a good season and may still be the most skilled player in the world, but I am sure that he wished one of his wins came in a major

 

 

2012 PGA TOUR STROKE DIFFERENTIAL FIELD AVERAGE
Select Year: ----201220112010200920082007200620052004200320022001200019991998199719961995199419931992199119901989198819871986198519841983
Y-T-D statistics through: U.S. Open Jun 17, 2012
RANK
THIS
WEEK
RANK
LAST
WEEK
PLAYER ROUNDS STROKE DIFFERENTIAL PLAYER'S SCORE FIELD AVERAGE SCORE
1 1 Tiger Woods 35 +2.18 70.30 72.48
T2 6 Jim Furyk 47 +1.91 70.39 72.30
T2 T2 Justin Rose 45 +1.91 70.30 72.21
4 4 Matt Kuchar 52 +1.88 69.98 71.86
5 7 Lee Westwood  titleist_38x11.gif 34 +1.83 70.36 72.19
6 5 Bubba Watson  titleist_38x11.gif 38 +1.79 70.19 71.98
7 T2 Rory McIlroy  titleist_38x11.gif 32 +1.75 70.42 72.17
8 T9 Jason Dufner  titleist_38x11.gif 55 +1.72 69.93 71.65
9 8 Luke Donald  titleist_38x11.gif 35 +1.58 70.68 72.26
10 11 Zach Johnson  titleist_38x11.gif 55 +1.53 70.24 71.77
11 T9 Phil Mickelson 51 +1.48 70.41 71.89
12 12 Dustin Johnson 34 +1.33 70.55 71.88
13 T26 Graeme McDowell 33 +1.29 70.84 72.13
14 14 Hunter Mahan  titleist_38x11.gif 52 +1.24 71.07 72.31
T15 25 Padraig Harrington  titleist_38x11.gif 39 +1.23 70.67 71.90
T15 T26 Webb Simpson  titleist_38x11.gif 49 +1.23 70.50 71.73
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Tour Talk
TheSandTrap.com › Golf Forum › The Clubhouse › Tour Talk › OWGR Biased Against PGA Tour Players