Handicap question

[rogolf]

Golf Australia did an exhaustive statistical review prior to developing their new handicap system and implementing it on January 1, 2014. Here is a review after implementation

 

https://www.golf.org.au/2016-stat-review

 

A summary:

 

Five Key Findings from the Statistical Reports

* Since the introduction in January 2014 of Slope, DSR, and Stableford handicapping of Stroke rounds, there has not been any significant change to ‘equity in competition results patterns’. ‘Equity in competition results patterns’ means the balance of players at each handicap level who are winning competitions, and also who are finishing as placegetters in competitions.

* Overall across Australia, players of all handicap levels are winning close to their fair share of competitions, and players of all handicap levels are achieving close to their fair share of prizes in competitions. The profile of competition winners and prize winners does shift from season to season, from state to state, and even from club to club. There is a very small number of clubs that exhibit competition results patterns that are notably different to the national patterns – such outcomes can be the result for example of the unusual nature of a course, or of the profile of the membership (such clubs are always welcome to contact their state association to discuss potential solutions).

* Over the course of a year for women the current GA Handicap System produces a slight bias towards players with single-figure handicaps. This is the case when equity is measured by the chance of a golfer to finish as a placegetter in a competition, and also when equity is measured by the chance of a golfer to win the competition.

* Over the course of a year for men the current GA Handicap System produces a slight bias towards players with single-figure handicaps when equity is measured by the chance of a golfer to finish as a placegetter in a competition. The reverse is true when equity is measured by the chance of a golfer to win the competition – in this case the bias is slightly towards players with high handicaps. (Note: If GA was to fine-tune the handicap system to further enhance the chances of male single-figure players winning competitions, a direct side effect would be to create a severe bias towards the chance a single-figure player to finish as a placegetter in a competition.)

* The competiveness of players with single figure handicaps improves in winter whereas the competitiveness of players with high handicaps improves in summer. This effect is more notable in the southern states.

[davep043]

It's tricky to find the right balance. As a high handicapper, I believe that I am much more likely than a low handicapper of popping out a round several shots better than any of my previous twenty rounds. That said, I am also pretty likely to shoot one a few shots off the high end. For be, using fewer scores is going to (I think) put me a something of a disadvantage, since my actual scores are more likely to be more spread out. I have to confess to not reading the whole thread, but I got the feeling at the start that the issue seen by the OP might not be so much the actual numbers, but the definition, or application, of the word potential. And I could see that. I certainly have the potential to shoot 85, maybe lower. Now and then I get under 90, so it's certainly in there. But if my handicap was based on that, I would be out of luck, because it's probably not happening (at least it hasn't yet!) I know we'd like a perfect system, but what we have doesn't seem awful too me, We use a very similar (just not official) method in our summer golf league, and on any one day we find that pretty much any team can beat any other, and vice-versa. It's not perfect, but we have fun. And that's what I'm looking for!

To the two underlined phrases, I agree with each of them. I believe the OP was pointing at a disconnect between the idea of "potential" and the way the handicap is calculated. I also believe that by moving the handicap closer to the lower end of each player's spectrum, you're giving weighting the system to favor a more consistent player over a more erratic player. As higher handicaps are generally more erratic than better players, that same change weights the system towards a lower handicap, in general. Whether that is a good or bad thing is to make a value judgement over who deserves to have a better chance in a competition. I'd prefer to take the word "potential" out of the description rather that shift the balance toward any specific group of players.

[Fade]

The USGA says the purpose of a handicap is to identify a players potential ability so they can compete with players of differing abilities on an equitable basis.

 

The way I read the above premise doesn't seem to represent the USGA very well. You are putting too much emphasis on what the handicap number means or doesn't mean, IMO. I like their wording better.

 

The purpose of the USGA Handicap System is to make the game of golf more enjoyable by enabling players of differing abilities to compete on an equitable basis.

http://www.usga.org/...tml#!rule-14367

 

The handicap numbers used to accomplish that are primarily a means to that end.

 

I understand the goal is to allow everyone to compete on an equal basis.

 

USGA "A Handicap Index intends to reflect potential ability,...."

 

https://www.usga.org...r.asp?FAQidx=22

 

USGA "The USGA Handicap Index is widely recognized in America and elsewhere as a reliable measure of a player's potential ability."

 

http://www.usga.org/...e-23c19d10.html

 

Exactly! The other alternative is for the USGA to remove the words "potential ability" from the purpose of a handicap. Maybe replace it with "potential average ability".

 

They probably shouldn't be bragging about it like they did in the quote you partially underlined.

 

Then again, maybe the words "potential ability" are helpful to some golfers, provided they don't read too much into them.

 

FWIW, I would have no problem with your suggestion to remove these words. I suspect they'd have a terrible time trying to define what they mean exactly.

[nsxguy]

The USGA says the purpose of a handicap is to identify a players potential ability so they can compete with players of differing abilities on an equitable basis.

 

The way I read the above premise doesn't seem to represent the USGA very well. You are putting too much emphasis on what the handicap number means or doesn't mean, IMO. I like their wording better.

 

The purpose of the USGA Handicap System is to make the game of golf more enjoyable by enabling players of differing abilities to compete on an equitable basis.

http://www.usga.org/...tml#!rule-14367

 

The handicap numbers used to accomplish that are primarily a means to that end.

 

I understand the goal is to allow everyone to compete on an equal basis.

 

USGA "A Handicap Index intends to reflect potential ability,...."

 

https://www.usga.org...r.asp?FAQidx=22

 

USGA "The USGA Handicap Index is widely recognized in America and elsewhere as a reliable measure of a player's potential ability."

 

http://www.usga.org/...e-23c19d10.html

 

Firstly I should say I've objected to the word "potential" for more years than I can count - but it is what it is.

 

But why do you focus on ONE single word, defining something that cannot POSSIBLY be measured ?

 

So if I caught lightning in a bottle and shot a minus 4 ONE TIME I should forever be expected to have a FAIR match with someone who's best was "only" a +3 by giving him 7 shots every time we play ? When both of our averages are around +7 or so ?

 

You posted yourself the PURPOSE of handicaps. Why get stuck on a single word ?

 

I know it's been a bit boring here lately but,,,,,,,,,,,,,,,,,,,,,,,,,,,, :rolleyes:

[HatsForBats]

Overall across Australia, players of all handicap levels are winning close to their fair share of competitions, and players of all handicap levels are achieving close to their fair share of prizes in competitions.

 

If I understand their figures correctly I am not sure they have shown that to be true. They appear to be treating a 2 way tie for 1st as half a win, a 3-way tie as 1/3rd of a win etc. This produces percentages for each player group where the overall total will be very close to 100%. With rounding it will sometimes be slightly more or less than 100%. I agree that is a nice & neat way to show the data but it doesn't tell me what I really want to know.. ie. what amount of the money does each player range win.

 

For example if there were 50 players that each put in $10 the prize pool is $500 with a payout as shown below (which is from a poker tournament payout schedule). If you count a 2 way tie as only half of a win then each player gets $80. In actuality they would each share half of the combined 1st and 2nd place prizes ($270) or $135 per player. It is a significant difference especially when you get to 3, 4 or more way ties.

 

1st = $160

2nd = $110

3rd = $82.50

4th = $62.50

5th = $45

6th = $40

 

This is what I am struggling with for my analysis of the USGA Appendix E data. I think in addition to the way the Australian report shows the data it also needs to be shown from multiple payout structures to get the true picture.

[ruca]

Overall across Australia, players of all handicap levels are winning close to their fair share of competitions, and players of all handicap levels are achieving close to their fair share of prizes in competitions.

 

If I understand their figures correctly I am not sure they have shown that to be true. They appear to be treating a 2 way tie for 1st as half a win, a 3-way tie as 1/3rd of a win etc. This produces percentages for each player group where the overall total will be very close to 100%. With rounding it will sometimes be slightly more or less than 100%. I agree that is a nice & neat way to show the data but it doesn't tell me what I really want to know.. ie. what amount of the money does each player range win.

 

For example if there were 50 players that each put in $10 the prize pool is $500 with a payout as shown below (which is from a poker tournament payout schedule). If you count a 2 way tie as only half of a win then each player gets $80. In actuality they would each share half of the combined 1st and 2nd place prizes ($270) or $135 per player. It is a significant difference especially when you get to 3, 4 or more way ties.

 

1st = $160

2nd = $110

3rd = $82.50

4th = $62.50

5th = $45

6th = $40

 

This is what I am struggling with for my analysis of the USGA Appendix E data. I think in addition to the way the Australian report shows the data it also needs to be shown from multiple payout structures to get the true picture.

The first bit was the odds of winning which slightly favoured higher hcps. Then they stated the odds of placing slightly favoured the lower hcps. What this amounts to overall depends on how any comp breaks down it's prizes which may vary.

I wonder if golf Australia has a table similar to USGA appendix E.

Hats, don't give up!. Your programming skills are better than most.

[rogolf]

Overall across Australia, players of all handicap levels are winning close to their fair share of competitions, and players of all handicap levels are achieving close to their fair share of prizes in competitions.

 

If I understand their figures correctly I am not sure they have shown that to be true. They appear to be treating a 2 way tie for 1st as half a win, a 3-way tie as 1/3rd of a win etc. This produces percentages for each player group where the overall total will be very close to 100%. With rounding it will sometimes be slightly more or less than 100%. I agree that is a nice & neat way to show the data but it doesn't tell me what I really want to know.. ie. what amount of the money does each player range win.

 

For example if there were 50 players that each put in $10 the prize pool is $500 with a payout as shown below (which is from a poker tournament payout schedule). If you count a 2 way tie as only half of a win then each player gets $80. In actuality they would each share half of the combined 1st and 2nd place prizes ($270) or $135 per player. It is a significant difference especially when you get to 3, 4 or more way ties.

 

1st = $160

2nd = $110

3rd = $82.50

4th = $62.50

5th = $45

6th = $40

 

This is what I am struggling with for my analysis of the USGA Appendix E data. I think in addition to the way the Australian report shows the data it also needs to be shown from multiple payout structures to get the true picture.

Not a fully-qualified statistician, but I thought they talked about winning (presumably outright) and placegetting (top 10)?

[HatsForBats]

Overall across Australia, players of all handicap levels are winning close to their fair share of competitions, and players of all handicap levels are achieving close to their fair share of prizes in competitions.

 

If I understand their figures correctly I am not sure they have shown that to be true. They appear to be treating a 2 way tie for 1st as half a win, a 3-way tie as 1/3rd of a win etc. This produces percentages for each player group where the overall total will be very close to 100%. With rounding it will sometimes be slightly more or less than 100%. I agree that is a nice & neat way to show the data but it doesn't tell me what I really want to know.. ie. what amount of the money does each player range win.

 

For example if there were 50 players that each put in $10 the prize pool is $500 with a payout as shown below (which is from a poker tournament payout schedule). If you count a 2 way tie as only half of a win then each player gets $80. In actuality they would each share half of the combined 1st and 2nd place prizes ($270) or $135 per player. It is a significant difference especially when you get to 3, 4 or more way ties.

 

1st = $160

2nd = $110

3rd = $82.50

4th = $62.50

5th = $45

6th = $40

 

This is what I am struggling with for my analysis of the USGA Appendix E data. I think in addition to the way the Australian report shows the data it also needs to be shown from multiple payout structures to get the true picture.

Not a fully-qualified statistician, but I thought they talked about winning (presumably outright) and placegetting (top 10)?

 

I'm not even a partially qualified statistician but I like looking at numbers and trying to determine if they actually mean what they seem to show.

 

If the winning % in that report is. as you presume. outright wins then they would need to ignore a large portion of the dataset to get the percentages. If the Win % for each range shown in that report adds up to 100%, which the row I checked did, then they would either have to use only those events where there was an outright winner or use all the events but count a 2-way tie as 50% of a win, 3-way tie as 33% etc.

 

For example today I ran a test of 20 players using the USGA Appendix A figures with 20 players of various index ranges (0 to 4.9, 5.0 to 9.9, 10.0 to 14.9 up to 25.0 to 24.9). Over 500k events only 378,642 (or 75.7%) of the events had an outright winner. With 50+ players I presume the percentage of events that would have an outright winner would be even smaller. If I broke down the outright win % by player group then the percentage should only add up to 75.7% or if I ignored the events with ties for first the percentages would ad up to 100% but then I would be ignoring a large portion of the dataset.

 

The place getting was top 10% I believe (maybe something about 20% as well). So for a 50 player event top 5, at 55 it could be still 5 players or perhaps they round up to 6... we don't know. Maybe the details are in the full report which I haven't read.

[ruca]

I presume most tournaments will have some sort of countback to determine the overall winner when 2 or more players tie. Thats essentially the same as counting 0.5 when 2 players tie etc.

[nsxguy]

Overall across Australia, players of all handicap levels are winning close to their fair share of competitions, and players of all handicap levels are achieving close to their fair share of prizes in competitions.

 

If I understand their figures correctly I am not sure they have shown that to be true. They appear to be treating a 2 way tie for 1st as half a win, a 3-way tie as 1/3rd of a win etc. This produces percentages for each player group where the overall total will be very close to 100%. With rounding it will sometimes be slightly more or less than 100%. I agree that is a nice & neat way to show the data but it doesn't tell me what I really want to know.. ie. what amount of the money does each player range win.

 

For example if there were 50 players that each put in $10 the prize pool is $500 with a payout as shown below (which is from a poker tournament payout schedule). If you count a 2 way tie as only half of a win then each player gets $80. In actuality they would each share half of the combined 1st and 2nd place prizes ($270) or $135 per player. It is a significant difference especially when you get to 3, 4 or more way ties.

 

1st = $160

2nd = $110

3rd = $82.50

4th = $62.50

5th = $45

6th = $40

 

This is what I am struggling with for my analysis of the USGA Appendix E data. I think in addition to the way the Australian report shows the data it also needs to be shown from multiple payout structures to get the true picture.

Not a fully-qualified statistician, but I thought they talked about winning (presumably outright) and placegetting (top 10)?

 

I'm not even a partially qualified statistician but I like looking at numbers and trying to determine if they actually mean what they seem to show.

 

I'm not a partially qualified statistician either but the field of statistics is a science, i.e. NOT for the casual academic. So I'm not sure why you're going to such lengths to prove something when you really don't know whether your "formulas" or hypotheses are anywhere near correct.

 

As an exercise, or something that just interests you, fine. Everybody's got hobbies. But I don't think you'll actually "prove" anything.

 

It's kind of like the old statistics query about birthdays. If you took, say, the first 7 4-somes of a random day, 28 guys, to tee it up and asked them all to give their birthdays (Month and day, NOT year) and asked a person at random the odds for ANY 2 of those 28 guys to have the same birthday, people would give all sorts of different/crazy odds, probably none of them close to correct.

 

In reality the odds are 2-1 IN FAVOR of there being a match. If you took 75 random people it would be a virtual certainty of 2 of them matching.

 

Statistics is a very difficult field.

 

But if you enjoy running these models and toying with the outcomes, by all means, enjoy.

[nsxguy]

I presume most tournaments will have some sort of countback to determine the overall winner when 2 or more players tie. Thats essentially the same as counting 0.5 when 2 players tie etc.

 

I'm sorry as I haven't been following the examples closely, but by "counting 0.5" are you suggesting getting/giving partial strokes, to the "tenths" place to determine a net event ?

 

Or "when 2 players tie" ? Only using tenths when a tie needs to be broken ?

[ruca]

I presume most tournaments will have some sort of countback to determine the overall winner when 2 or more players tie. Thats essentially the same as counting 0.5 when 2 players tie etc.

 

I'm sorry as I haven't been following the examples closely, but by "counting 0.5" are you suggesting getting/giving partial strokes, to the "tenths" place to determine a net event ?

 

Or "when 2 players tie" ? Only using tenths when a tie needs to be broken ?

In reality when 2 or more players tie some sort of countback system determines the winner e.g. best back 9.

These are liable to be equally split over time between different types of players that tie, and so is mathematically equivalent to dividing the "win" by the number of players. There's no partial strokes.

[Roadking2003]

 

Firstly I should say I've objected to the word "potential" for more years than I can count - but it is what it is.

 

But why do you focus on ONE single word, defining something that cannot POSSIBLY be measured ?

 

Yes, it's one word, but it's a very important part of the "purpose of handicaps" message.

 

If we eliminate "potential", then the purpose would be "just find a method that makes the likelihood of winning close to even across all handicaps."

 

If we eliminate "potential", then there may be no reason to throw out your ten worst scores.

 

It would be interesting to see the impact of using all scores; just your average. How would that change from using your ten lowest?

[HatsForBats]

I'm not a partially qualified statistician either but the field of statistics is a science, i.e. NOT for the casual academic. So I'm not sure why you're going to such lengths to prove something when you really don't know whether your "formulas" or hypotheses are anywhere near correct.

 

I disagree that they are not for the casual academic when we are talking about using databases to produce random results where the formulas are already figured out (USGA Appendix E if you dig in far enough). To test the results I ran 25,000,000 random records for each group and then checked that the results would build the same frequency table as the USGA table. They do with only the least frequent Net Differentials being slightly off due to the sheer number of records needed to accurately account for something happening say once in 8,000 or 10,000 events.

 

As far as the Birthday question. Sure I couldn't answer it right there on the course but it took me all of 5 minutes to find the answer.

 

Number of pairs 378 = (28 * 27)/2 Chance of a unique pair 99.7260% = 364/365 Chance of 378 unique pairs 35.45% = (99.7260%)^378 Chance of some match 64.55% = 1 - 35.45%

 

That could easily be programmed into a database to get the same results and the check to make sure the formulas are correct is readily available. Or I could create 20,000,000 events of 28 players with a random number of 1 to 365 to assign them a birthday. Then could easily run a script to find the number of events where at least 2 players had the same birthday. Seems kind of silly for the birthday question but not so silly when talking about Net Differentials between varying numbers of players and varying index ranges. Seems kind of silly for the Birthday question but not for Net Differentials in golf matches of varying sizes and index ranges.

 

I'm sure if you asked about a specific event such as what percent of the time will a 0 to 4.9 index lose when playing 1 player from each of the other index ranges then someone could develop a formula to answer that. Then start asking them questions like how often will the 0 to 4.9 index have a streak of 20 to 29 events where they don't win? 30 to 39 events? 50 to 75 events? 100+ events? How much would they win on average per event if the payout were $40 for 1st, $25 for second and $15 for third? What about a different payout scheme? etc. A database can answer so much more than a single formula or even a few formulas.

[rogolf]

Firstly I should say I've objected to the word "potential" for more years than I can count - but it is what it is.

 

But why do you focus on ONE single word, defining something that cannot POSSIBLY be measured ?

 

Yes, it's one word, but it's a very important part of the "purpose of handicaps" message.

 

If we eliminate "potential", then the purpose would be "just find a method that makes the likelihood of winning close to even across all handicaps."

 

If we eliminate "potential", then there may be no reason to throw out your ten worst scores.

 

It would be interesting to see the impact of using all scores; just your average. How would that change from using your ten lowest?

 

Fly at it and let us know your findings. In the meantime, I will accept the work that has been done by the USGA and other national golf associations. Also, I'm not worried about the wording in the handicap description or manuals. The authors are doing good work.

[nsxguy]

I'm not a partially qualified statistician either but the field of statistics is a science, i.e. NOT for the casual academic. So I'm not sure why you're going to such lengths to prove something when you really don't know whether your "formulas" or hypotheses are anywhere near correct.

 

I disagree that they are not for the casual academic when we are talking about using databases to produce random results where the formulas are already figured out (USGA Appendix E if you dig in far enough). To test the results I ran 25,000,000 random records for each group and then checked that the results would build the same frequency table as the USGA table. They do with only the least frequent Net Differentials being slightly off due to the sheer number of records needed to accurately account for something happening say once in 8,000 or 10,000 events.

 

As far as the Birthday question. Sure I couldn't answer it right there on the course but it took me all of 5 minutes to find the answer.

 

Number of pairs 378 = (28 * 27)/2 Chance of a unique pair 99.7260% = 364/365 Chance of 378 unique pairs 35.45% = (99.7260%)^378 Chance of some match 64.55% = 1 - 35.45%

 

 

Did you find the answer" via Google ? Or did you work out the solution for yourself ?

[nsxguy]

Firstly I should say I've objected to the word "potential" for more years than I can count - but it is what it is.

 

But why do you focus on ONE single word, defining something that cannot POSSIBLY be measured ?

 

Yes, it's one word, but it's a very important part of the "purpose of handicaps" message.

 

If we eliminate "potential", then the purpose would be "just find a method that makes the likelihood of winning close to even across all handicaps."

 

If we eliminate "potential", then there may be no reason to throw out your ten worst scores.

 

It would be interesting to see the impact of using all scores; just your average. How would that change from using your ten lowest?

 

No, it's NOT a "very important part of the message"

 

What ? Now you're making up your own stuff ? You don't like "The purpose of the USGA Handicap System is to make the game of golf more enjoyable by enabling players of differing abilities to compete on an equitable basis." ???

 

That is the "message".

 

Do you really think that averaging all of the last 20 would be better ? Or are you "just curious" ? Why not the last 50 ? Why not ALL your historical scores ?

 

It's just incredible that here we have an organization that is supposed to have the game's best interest as their primary focus and who's crunched 10s of millions of rounds and countless scenarios and yet some people think they can do it better. That they know something the USGA does not.

 

I'd say you can't make this stuff up,,,,,,,,,,,,,, but obviously some can,,,,,,,,,,,,

[Roadking2003]

Firstly I should say I've objected to the word "potential" for more years than I can count - but it is what it is.

 

But why do you focus on ONE single word, defining something that cannot POSSIBLY be measured ?

 

Yes, it's one word, but it's a very important part of the "purpose of handicaps" message.

 

If we eliminate "potential", then the purpose would be "just find a method that makes the likelihood of winning close to even across all handicaps."

 

If we eliminate "potential", then there may be no reason to throw out your ten worst scores.

 

It would be interesting to see the impact of using all scores; just your average. How would that change from using your ten lowest?

 

No, it's NOT a "very important part of the message"

 

What ? Now you're making up your own stuff ? You don't like "The purpose of the USGA Handicap System is to make the game of golf more enjoyable by enabling players of differing abilities to compete on an equitable basis." ???

 

That is the "message".

 

Do you really think that averaging all of the last 20 would be better ? Or are you "just curious" ? Why not the last 50 ? Why not ALL your historical scores ?

 

It's just incredible that here we have an organization that is supposed to have the game's best interest as their primary focus and who's crunched 10s of millions of rounds and countless scenarios and yet some people think they can do it better. That they know something the USGA does not.

 

I'd say you can't make this stuff up,,,,,,,,,,,,,, but obviously some can,,,,,,,,,,,,

 

I agree.... some can make stuff up.... like your post..... incredible......

[ruca]

Firstly I should say I've objected to the word "potential" for more years than I can count - but it is what it is.

 

But why do you focus on ONE single word, defining something that cannot POSSIBLY be measured ?

 

Yes, it's one word, but it's a very important part of the "purpose of handicaps" message.

 

If we eliminate "potential", then the purpose would be "just find a method that makes the likelihood of winning close to even across all handicaps."

 

If we eliminate "potential", then there may be no reason to throw out your ten worst scores.

 

It would be interesting to see the impact of using all scores; just your average. How would that change from using your ten lowest?

 

No, it's NOT a "very important part of the message"

 

What ? Now you're making up your own stuff ? You don't like "The purpose of the USGA Handicap System is to make the game of golf more enjoyable by enabling players of differing abilities to compete on an equitable basis." ???

 

That is the "message".

 

Do you really think that averaging all of the last 20 would be better ? Or are you "just curious" ? Why not the last 50 ? Why not ALL your historical scores ?

 

It's just incredible that here we have an organization that is supposed to have the game's best interest as their primary focus and who's crunched 10s of millions of rounds and countless scenarios and yet some people think they can do it better. That they know something the USGA does not.

 

I'd say you can't make this stuff up,,,,,,,,,,,,,, but obviously some can,,,,,,,,,,,,

 

If you critically look at peer reviewed articles involving statistics in a variety of fields where the authors used a statistician, you'd be amazed at the errors that (a)they can make, and (b)the reviewers don't notice.

 

The hcp system is something that's evolved over time, and is very different now to what it was say 50yrs ago. People inherently resist change, thus when flaws in the system become apparent, it takes time for a change.

 

The USGA have never even submitted their data for peer review (such as the slope system), so we have no idea of how good their methods are.

 

They state the purpose is to make golf equitable for players of differing ability.

It's obvious to some of us that it would be more accurate if there was a correction for the number of players. This correction is often larger than slope . As none of the hcp bodies even acknowledge the issue, it is a cause to question them.

It's even more obvious that if the system doesn't protect against the small number of cheats then it can break down (this is my biggest concern with the incoming system; CONGU/EGA despite other flaws, at least addressed this inherently).

 

Appendix E is like a treasure trove of hidden information, what Hats is doing is not complicated from a maths perspective, but his programming skills are bringing out some interesting observations.

By providing a variety of tables for different numbers of players and different pay out structures, you can see how to set up a more equitable competition for your particular group.

Isn't that what you want?

[nsxguy]

Firstly I should say I've objected to the word "potential" for more years than I can count - but it is what it is.

 

But why do you focus on ONE single word, defining something that cannot POSSIBLY be measured ?

 

Yes, it's one word, but it's a very important part of the "purpose of handicaps" message.

 

If we eliminate "potential", then the purpose would be "just find a method that makes the likelihood of winning close to even across all handicaps."

 

If we eliminate "potential", then there may be no reason to throw out your ten worst scores.

 

It would be interesting to see the impact of using all scores; just your average. How would that change from using your ten lowest?

 

No, it's NOT a "very important part of the message"

 

What ? Now you're making up your own stuff ? You don't like "The purpose of the USGA Handicap System is to make the game of golf more enjoyable by enabling players of differing abilities to compete on an equitable basis." ???

 

That is the "message".

 

Do you really think that averaging all of the last 20 would be better ? Or are you "just curious" ? Why not the last 50 ? Why not ALL your historical scores ?

 

It's just incredible that here we have an organization that is supposed to have the game's best interest as their primary focus and who's crunched 10s of millions of rounds and countless scenarios and yet some people think they can do it better. That they know something the USGA does not.

 

I'd say you can't make this stuff up,,,,,,,,,,,,,, but obviously some can,,,,,,,,,,,,

 

If you critically look at peer reviewed articles involving statistics in a variety of fields where the authors used a statistician, you'd be amazed at the errors that (a)they can make, and (b)the reviewers don't notice.

 

The hcp system is something that's evolved over time, and is very different now to what it was say 50yrs ago. People inherently resist change, thus when flaws in the system become apparent, it takes time for a change.

 

The USGA have never even submitted their data for peer review (such as the slope system), so we have no idea of how good their methods are.

 

They state the purpose is to make golf equitable for players of differing ability.

It's obvious to some of us that it would be more accurate if there was a correction for the number of players.

 

Appendix E is like a treasure trove of hidden information, what Hats is doing is not complicated from a maths perspective, but his programming skills are bringing out some interesting observations.

 

By providing a variety of tables for different numbers of players and different pay out structures, you can see how to set up a more equitable competition for your particular group.

Isn't that what you want?

 

As you say the USGA hasn't given up their data for peer review. So you're saying that their "peers" could've/would've/should've done it better ? Better still, how do you know they didn't have their data reviewed ? Who are their peers ?

 

I've already said Hats' has a noble idea but it appears as though, since there is little commentary on his outcomes (other than yourself) you are the only one who pretends to understand what he is coming up with. And therein lies the rub. You need to have a system that is at least somewhat understandable to the players who are going to use it.

 

The current handicap system is pretty transparent and easy enough to understand for anybody who tries to. And the "exceptions", such as playing from different tee sets and the odds of exceptional scores are also fairly easy to understand (if one tries).

 

But you think that a player should only enter a tournament if he's a 7.3 and there is a specifica mix of higher and lower handicaps and a certain number of players in the field and the payouts are a, b, and c and then determine if it's worth his while to enter ?

 

Not to mention there may be more entries after he enters that would then skew the parameters again so that now he has less of a chance and now wants to withdraw ? Madness. Sheer madness.

 

Assuming Hats could even come up with a formula for ANY number of players in an event and adjust the handicaps and payouts accordingly WHO ON EARTH would understand it ? You don't think that these formulae would cause more uproar than the current sandbagging complaints, usually by those who can't be bothered to understand the handicap system or the odds of an exceptional score ?

 

Sandbaggers are usually obvious. It's the organizers who don't want to "upset the apple cart". Then they lose participants and wonder why - or they don't lose people because the ones who would leave shrug their shoulders and say "It is what it is. Maybe I'll have a great day this time around" instead of confronting the organizers and/or the sandbaggers themselves - after all, who really wants (more) confrontation in their life ?

 

And you have Hats' admission that he's no statistician but you're "relying" on his result over numbers that big and his methodology ?

 

Q. "Hey I won last week against 23 guys and got paid less than the winner this week with only 18 guys in the field. What's upi with that ?"

 

A. "Well, you see, that's what the tables say"

 

Good luck with that.

 

And good luck in your (and Hats' of course) endeavors. Oh, and the definitions that upset RK too,,,,,,,,,,,,,

[nsxguy]

Firstly I should say I've objected to the word "potential" for more years than I can count - but it is what it is.

 

But why do you focus on ONE single word, defining something that cannot POSSIBLY be measured ?

 

Yes, it's one word, but it's a very important part of the "purpose of handicaps" message.

 

If we eliminate "potential", then the purpose would be "just find a method that makes the likelihood of winning close to even across all handicaps."

 

If we eliminate "potential", then there may be no reason to throw out your ten worst scores.

 

It would be interesting to see the impact of using all scores; just your average. How would that change from using your ten lowest?

 

No, it's NOT a "very important part of the message"

 

What ? Now you're making up your own stuff ? You don't like "The purpose of the USGA Handicap System is to make the game of golf more enjoyable by enabling players of differing abilities to compete on an equitable basis." ???

 

That is the "message".

 

Do you really think that averaging all of the last 20 would be better ? Or are you "just curious" ? Why not the last 50 ? Why not ALL your historical scores ?

 

It's just incredible that here we have an organization that is supposed to have the game's best interest as their primary focus and who's crunched 10s of millions of rounds and countless scenarios and yet some people think they can do it better. That they know something the USGA does not.

 

I'd say you can't make this stuff up,,,,,,,,,,,,,, but obviously some can,,,,,,,,,,,,

 

I agree.... some can make stuff up.... like your post..... incredible......

 

I don't think my post was "incredible". Actually it was kinda sorta obvious. But thanks anyway. :D

 

Good luck on your shouting at the rain. :good:

 

[ruca]

Nsxguy

By peer review, I mean publishing their data in appropriate journals so we can see their methodology. If it was there, it should be easy to find.

By way of analogy, suppose the fda says, "here's a drug for condition x, we've decided this is what everyone has to take". No data, studies or info in the public domain. Do you just blindly trust, or try to find out more?

 

All Hats is doing is running simulations to see how equitable different situations of player numbers and hcps actually are. Nothing he's done so far is at all complicated (to a numbers person)

 

This is nothing to do with sandbagging; which is a whole separate issue.

 

 

 

 

[nsxguy]

Nsxguy

By peer review, I mean publishing their data in appropriate journals so we can see their methodology. If it was there, it should be easy to find.

By way of analogy, suppose the fda says, "here's a drug for condition x, we've decided this is what everyone has to take". No data, studies or info in the public domain. Do you just blindly trust, or try to find out more?

 

All Hats is doing is running simulations to see how equitable different situations of player numbers and hcps actually are. Nothing he's done so far is at all complicated (to a numbers person)

 

This is nothing to do with sandbagging; which is a whole separate issue.

 

If you or I spent all our time pouring over peer reviews and journals and trials and tests and, and, and, and, we'd die from whatever was ailing us in the first place before we could make a decision to take, or not take, said drug.

 

That is what the CDC is for - to do all that for us. Surely there's a body in the UK that is tasked with the same processes, no ? Do YOU do all that research to "peer review" their findings ? That is also what doctors are for - to do all that for us.

 

Experts in their field do all that legwork. If you don't trust their motives, that's fine. Do what ya gotta do.

 

In this case the USGA, while not perfect, is one of the "guardians" of the game. I see no motivation for them to do anything that's not in the game's best interest. And there is such a thing as "internal peer review".

 

Hats is involved in an exercise that he finds interesting. Good for him. Really.

[HatsForBats]

All Hats is doing is running simulations to see how equitable different situations of player numbers and hcps actually are. Nothing he's done so far is at all complicated (to a numbers person)

 

Correct. I think NSX is taking exception with a post in another thread where I said 'I can't help but think some mathematical formula could even things out more". Which I still can't help but think but not for one minute do I think I could come up with one using statistics. More results may squash that belief or they may reinforce it. All I'm doing is generating the random events and looking over at the results. I will definitely try different 'fixes' for some groups just to see how it changes things. That's just curiosity. If it got to the point where some other than me thought a statistician could do something with it then great.

 

The kicker is this all changes come next year, right? When we will be using only 8 of the last 20 scores rather than 10 of 20. That is not a deterrent to my curiosity. It's a nice winter project.

[rogolf]

All Hats is doing is running simulations to see how equitable different situations of player numbers and hcps actually are. Nothing he's done so far is at all complicated (to a numbers person)

 

Correct. I think NSX is taking exception with a post in another thread where I said 'I can't help but think some mathematical formula could even things out more". Which I still can't help but think but not for one minute do I think I could come up with one using statistics. More results may squash that belief or they may reinforce it. All I'm doing is generating the random events and looking over at the results. I will definitely try different 'fixes' for some groups just to see how it changes things. That's just curiosity. If it got to the point where some other than me thought a statistician could do something with it then great.

 

The kicker is this all changes come next year, right? When we will be using only 8 of the last 20 scores rather than 10 of 20. That is not a deterrent to my curiosity. It's a nice winter project.

Changes to handicapping aren't coming until 2020.

Can you refresh my memory on what the goal is with changes?

[HatsForBats]

All Hats is doing is running simulations to see how equitable different situations of player numbers and hcps actually are. Nothing he's done so far is at all complicated (to a numbers person)

Correct. I think NSX is taking exception with a post in another thread where I said 'I can't help but think some mathematical formula could even things out more". Which I still can't help but think but not for one minute do I think I could come up with one using statistics. More results may squash that belief or they may reinforce it. All I'm doing is generating the random events and looking over at the results. I will definitely try different 'fixes' for some groups just to see how it changes things. That's just curiosity. If it got to the point where some other than me thought a statistician could do something with it then great. The kicker is this all changes come next year, right? When we will be using only 8 of the last 20 scores rather than 10 of 20. That is not a deterrent to my curiosity. It's a nice winter project.

Changes to handicapping aren't coming until 2020. Can you refresh my memory on what the goal is with changes?

 

Right, I'm in 2019 mode already.

 

For the changes the payout schedule... Right now the goal for me is just to see how different payout schemes change the final numbers and to see if it would boost up those falling behind. For the current file reducing the payout for 1st and increasing the other payouts helps the Low Single catch up but hurts the high handicappers.

[rogolf]

All Hats is doing is running simulations to see how equitable different situations of player numbers and hcps actually are. Nothing he's done so far is at all complicated (to a numbers person)

Correct. I think NSX is taking exception with a post in another thread where I said 'I can't help but think some mathematical formula could even things out more". Which I still can't help but think but not for one minute do I think I could come up with one using statistics. More results may squash that belief or they may reinforce it. All I'm doing is generating the random events and looking over at the results. I will definitely try different 'fixes' for some groups just to see how it changes things. That's just curiosity. If it got to the point where some other than me thought a statistician could do something with it then great. The kicker is this all changes come next year, right? When we will be using only 8 of the last 20 scores rather than 10 of 20. That is not a deterrent to my curiosity. It's a nice winter project.

Changes to handicapping aren't coming until 2020. Can you refresh my memory on what the goal is with changes?

 

Right, I'm in 2019 mode already.

 

For the changes the payout schedule... Right now the goal for me is just to see how different payout schemes change the final numbers and to see if it would boost up those falling behind. For the current file reducing the payout for 1st and increasing the other payouts helps the Low Single catch up but hurts the high handicappers.

As I alluded to in a previous post, I think that "payout schedules" are irrelevant to the apparent issue. I see the issue as each grouping of handicaps (low single, high mid etc) having a relatively equivalent chance of finishing first, top three, top ten etc. Whether they earn any "payout" or not is strictly dependent on how, and how many, the Committee decides to award prizes. Remember that amateurs (and only amateurs are eligible to have handicaps) are not permitted to play for cash; so, imo, the prize value is an irrelevant issue. The real issue is merely order of finish.

[Roadking2003]

All Hats is doing is running simulations to see how equitable different situations of player numbers and hcps actually are. Nothing he's done so far is at all complicated (to a numbers person)

 

Correct. I think NSX is taking exception with a post in another thread where I said 'I can't help but think some mathematical formula could even things out more". Which I still can't help but think but not for one minute do I think I could come up with one using statistics. More results may squash that belief or they may reinforce it. All I'm doing is generating the random events and looking over at the results. I will definitely try different 'fixes' for some groups just to see how it changes things. That's just curiosity. If it got to the point where some other than me thought a statistician could do something with it then great.

 

The kicker is this all changes come next year, right? When we will be using only 8 of the last 20 scores rather than 10 of 20. That is not a deterrent to my curiosity. It's a nice winter project.

Changes to handicapping aren't coming until 2020.

Can you refresh my memory on what the goal is with changes?

 

That is a GREAT question and one I have been asking since I heard about the world handicap initiative.

[Roadking2003]

Remember that amateurs (and only amateurs are eligible to have handicaps) are not permitted to play for cash; so, imo, the prize value is an irrelevant issue. The real issue is merely order of finish.

 

I've seen the handicap list at Whisper Rock and it includes Phil, McCord and several others.

 

Also, I've won lots of cash in amateur tournaments. They just have to be incidental to the main event.

[rogolf]

All Hats is doing is running simulations to see how equitable different situations of player numbers and hcps actually are. Nothing he's done so far is at all complicated (to a numbers person)

 

Correct. I think NSX is taking exception with a post in another thread where I said 'I can't help but think some mathematical formula could even things out more". Which I still can't help but think but not for one minute do I think I could come up with one using statistics. More results may squash that belief or they may reinforce it. All I'm doing is generating the random events and looking over at the results. I will definitely try different 'fixes' for some groups just to see how it changes things. That's just curiosity. If it got to the point where some other than me thought a statistician could do something with it then great.

 

The kicker is this all changes come next year, right? When we will be using only 8 of the last 20 scores rather than 10 of 20. That is not a deterrent to my curiosity. It's a nice winter project.

Changes to handicapping aren't coming until 2020.

Can you refresh my memory on what the goal is with changes?

 

That is a GREAT question and one I have been asking since I heard about the world handicap initiative.

I wasn't asking about the goal of the WHS - they've been previously stated. I'm looking to be reminded of the goal(s) of Hats' simulation work.