How'd they figure this out? [Archive] - Two Plus Two Older Archives

BottlesOf

06-09-2005, 11:57 PM

Consecutive holes in one odds (http://sportsillustrated.cnn.com/2005/golf/06/09/consecutive.aces/index.html?cnn=yes)

This article cites "Golf Digest" which says the odds of a foursome hitting 2 holes in one on a hole is 17 million to one. Where did they come up with that?

RacersEdge

06-10-2005, 12:01 AM

SWAG..

somethingstupid

06-10-2005, 01:50 AM

yes

blaze666

06-10-2005, 01:54 AM

that is a very silly statistic, it's impossibe to cluculate odds of something like that.

EliteNinja

06-10-2005, 05:52 AM

[ QUOTE ]
that is a very silly statistic, it's impossibe to cluculate odds of something like that.

[/ QUOTE ]

Maybe someone counted all the attempts.

Frequitude

06-10-2005, 07:03 AM

If they could estimate a mean shot accuracy (average distance from the hole) for the "average player", you would have a start.

Maybe assume that "average distance from the hole" is the mean in a normal distribution, and getting a hole-in-one would correlate to the very edge of the distribution.

Then again, I don't know if a normal distribution can be used for something like mean radial distance.

Thoughts?

maryfield48

06-10-2005, 08:47 AM

Couldn't they use the total number of hole-in-ones over the total number of par three holes played to work out a frequency?

jakethebake

06-10-2005, 08:49 AM

[ QUOTE ]
Couldn't they use the total number of hole-in-ones over the total number of par three holes played to work out a frequency?

[/ QUOTE ]

Which is probably what they did...which is also why they're wrong.

PokerGoblin

06-10-2005, 11:12 AM

Common knowledge in golf is that the probability of a hole in one on an average par 3 hole is roughly 1 in 13,000.

That figure was probably used to come to that conclusion.

Later

PG

jakethebake

06-10-2005, 11:15 AM

[ QUOTE ]
Common knowledge in golf is that the probability of a hole in one on an average par 3 hole is roughly 1 in 13,000.

[/ QUOTE ]

LOL...This doesn't answer the question at all. Common knowledge based on what? That's the whole question!

RacersEdge

06-10-2005, 11:34 AM

If you do the math, they figured the probability of a single golfer getting a hole in one at about 10,000-1. Seems high to me.

But maybe I'm biased because I have one. /images/graemlins/grin.gif

drewjustdrew

06-10-2005, 11:40 AM

They can't do this based on the average golfer. They need to at least figure out the handicaps of the golfers in question. Even then, some handicaps are based on better putting or driving than others, so it is all just BS

CallMeIshmael

06-10-2005, 11:45 AM

[ QUOTE ]
LOL...This doesn't answer the question at all. Common knowledge based on what? That's the whole question!

[/ QUOTE ]

At my course, they give out a plaque whenever you get a hole-in-one. You also get your name on this board. Thus, the total number of holes-in-one is recorded.

Im not sure about how it works at public clubs, but at private courses, word on the hole-in-one gets around very quickly. I've had one, and the clubhouse knew about it before I finished playing (and, it was on the 16th hole, so it wasnt like they had a long time to hear about it).

Finding out a reasonable hole-in-one frequency is nowhere near as hard as you seem to think it is.

CallMeIshmael

06-10-2005, 11:48 AM

[ QUOTE ]
They can't do this based on the average golfer. They need to at least figure out the handicaps of the golfers in question. Even then, some handicaps are based on better putting or driving than others, so it is all just BS

[/ QUOTE ]

Not necessarily.

17 million - 1, is the number that is produced assuming no prior knowledge of the golfers.

Sure, if one group is composed of 4 scratch players, and another all 20+ers, the numbers would change. But, that still doesnt mean you cant calculate it assuming no prior knowledge.

jakethebake

06-10-2005, 11:48 AM

[ QUOTE ]
Finding out a reasonable hole-in-one frequency is nowhere near as hard as you seem to think it is.

[/ QUOTE ]

I'm sure this is true. But a historical frequency has nothing to do with the odds of doing it.

CallMeIshmael

06-10-2005, 11:51 AM

[ QUOTE ]
I'm sure this is true. But a historical frequency has nothing to do with the odds of doing it.

[/ QUOTE ]

Given the VAST amount of data that is available on the subject, this statement is not true.

jakethebake

06-10-2005, 11:53 AM

[ QUOTE ]
[ QUOTE ]
I'm sure this is true. But a historical frequency has nothing to do with the odds of doing it.

[/ QUOTE ]

Given the VAST amount of data that is available on the subject, this statement is not true.

[/ QUOTE ]

Yea. I think it is. Vastness of data doesn't get you there.

CallMeIshmael

06-10-2005, 11:54 AM

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
I'm sure this is true. But a historical frequency has nothing to do with the odds of doing it.

[/ QUOTE ]

Given the VAST amount of data that is available on the subject, this statement is not true.

[/ QUOTE ]

Yea. I think it is. Vastness of data doesn't get you there.

[/ QUOTE ]

Do you not believe in statistics?

jakethebake

06-10-2005, 11:56 AM

[ QUOTE ]
Do you not believe in statistics?

[/ QUOTE ]

I believe in statistics. I also believe they're too often misused.

CallMeIshmael

06-10-2005, 12:04 PM

[ QUOTE ]
I believe in statistics. I also believe they're too often misused.

[/ QUOTE ]

lol.

Given we are talking about a binomial distribution with an N that is just absurdly large (like ten or hundreds of millions) we can draw very reasonable conclusions, even on data with a very large q/p ratio.

I cant even believe this is being debated.

tbach24

06-10-2005, 12:05 PM

[ QUOTE ]
[ QUOTE ]
I believe in statistics. I also believe they're too often misused.

[/ QUOTE ]

lol.

Given we are talking about a binomial distribution with an N that is just absurdly large (like ten or hundreds of millions) we can draw very reasonable conclusions, even on data with a very large q/p ratio.

I cant even believe this is being debated.

[/ QUOTE ]

NERD /images/graemlins/tongue.gif

jakethebake

06-10-2005, 12:08 PM

BreakfastBurrito

06-10-2005, 12:15 PM

The methods they use to come up with the number are faulty, but they're still in the ballpark. I think the entire point is just to give a general idea of how ridiculously unlikely this kind of thing is to someone who isn't that in to golf.

CallMeIshmael

06-10-2005, 12:16 PM

[ QUOTE ]
I have no idea what the distribution would look like, but you think if you were going to make a bet on those two women making back to back HiOs that 17-million is the correct odds?

[/ QUOTE ]

A) You have to be aware that there are different sets of information.

17m-1, assumes no prior infomation. But, now we know the foursome includes two people who shot x and y (forgot), so, no, 17m-1 isnt correct, because you would have to bring Beyes into it, and find a HiO frequency.

B) 17m-1 isnt going to be exact, but given the huge amount of data (one course, over 10 years: 40 foursomes (close?) * 100 days/year * 10 years = 160,000 pieces of data / 6 (# of par threes, again, an estimate) = 26,667) we are going to be able to draw some very good conclusions. Yes.

maryfield48

06-10-2005, 02:09 PM

They may well have used historic information for all recorded scores, not just those for this particular hole, or this particular course.

pshreck

06-10-2005, 02:12 PM

Jake is offering the correct perspective here people. The 1 in 17 million was clearly offered up as more crap to boost this story with the media and the public. It is nothing more than a guesstimation.

edtost

06-10-2005, 09:21 PM

[ QUOTE ]

B) 17m-1 isnt going to be exact, but given the huge amount of data (one course, over 10 years: 40 foursomes (close?) * 100 days/year * 10 years = 160,000 pieces of data / 6 (# of par threes, again, an estimate) = 26,667) we are going to be able to draw some very good conclusions. Yes.

[/ QUOTE ]

as someone who does recordkeeping at a golf club - we get between 10 and 15 thousand rounds/year (in the northeast - southern or west coast clubs probably get closer to 25). mulitply that by the standard 4 par 3's per course, so we have about 50,000 par 3 holes played per year; most years feature 2-3 HiO's, i think. if data was actually collected over 10 years, that is definitely statistically significant.