PGA FORUM

When: Nov 14-17

Where: Playa Del Carmen, Mexico

Course: El Camaleon GC | A Par 71 playing at 7039 yards

Last Year’s Winner: Matt Kuchar at -22 in “Caddygate”

Last Year’s Cut: -4

No Shot-tracker

Coverage: Golf Channel

Google Doc

RG Golf Forum League

Happy Meal Standings

Admin Note: The mod team is working to keep these daily threads more on topic – Golf strategy talk for this week’s contest. Post referring to last week’s contest will be moved to the correct thread. Any off topic posts or posts containing full lineups will be deleted..

The PGA Tour has hit it’s bye week. Don’t know why they couldn’t have had the Bermuda this week instead of last.

Agreed – a week off sucks

Finally got 6/6 through at Bermuda!
And now a week off.
What a letdown.

Since we are on an off-week, I’ll post this article and shake my head that a pro doesn’t know this rule

https://golfweek.com/2019/11/03/golf-two-players-penalized-lpga-q-series-violating-advice-rule/

Or the looper “Pro Jock”!?

Looks like I need to make a thread for the Turkish Airlines Open..

Happy Meal Cup standings have been updated. eruthruff takes over 1st place.

Turkish Airlines is up now on DK finally

Question for you data science guys. I’m finally building my own database by painstakingly copy/pasting data from the PGATour Media Guide site. Working in Excel for now, but will eventually start importing and playing with it in Python. Anyhow, for the dependent variable, I think I’m going to use the total strokes for the tournament, as opposed to finishing position. Ideally, I’d like to use DK points, but I would need to manually calculate that (pars, bogies, birdies, etc.) as historical DK point data isn’t available. My question on total strokes is, what’s the best way to handle missed cuts? I don’t really want to ignore them all together, and I don’t want to make them all the same, so for right now, I’m just multiplying them by 2, giving an indication of what the golfer may have scored if he played all 4 rounds. Of course, in some cases, it actually means that the missed cuts end-up with a better score than some of the made cuts, since some made cuts go on to shoot terrible rounds on Sat/Sun. But, even with that, it still seems like a good indication of their performance at a course. Any other suggestions on how to handle missed cuts?

I was hoping for a little more fanfare in that announcement…

I think it depends on how you’ll use the info. I use per round scores for most analysis and adjust it based on how others scored in that round/course and sometimes morning vs. afternoon. This is the same technique used at datagolf and other places like tournumbers (a fixed effects regression or similar). For other stuff where I’m just looking at how difficult a tournament is or not, would just take the average of all scores for each player (so yeah it kind of effectively assumes missed cut folks performed at their round 1/2 level for duration).

And I think one of those sites has some of that raw scoring data available to download to save you some copy and paste time.

I applaud the effort to go and get the data. I was 1 1/2 years into my database with espn data when I realized the SG data from media guide site would be better. I didnt have the umph to change all my stuff up.

I use excel and have 2+ years of DK pts per round and all bonuses accounted for. (some holes exist). you would need to sort and format to get what you want.

You confuse me by needing to calculate X2 on MC guys. With the variance in golf per front and back 9 and different days or many other reasons, why cant you take the data given and go with it. X2 isnt going to make anything helpful. LMAO an MDF would F you up.

Because I’m using the total strokes for the tournament. If I didn’t x2 the MCs, they’d be the winners! I could use average per round, but we’d basically have the same thing, as ilayup was saying.

I haven’t found one that lets you download the raw data. They keep close hold on that. PM me if you have some other differing info.

this has the last 2 years for free here (are we allowed to provide links here? guess I’ll find out)… i swear they’ve had earlier data for a fee before… worth checking…tournumbers.com/data-room

Sorry I am not a very good data guy. I guess dependent on strokes per tournament and all the variables (ie par 70,71,72 winner scores -25 and -8) lots of things are in the way of a perfect laboratory test.

Sorry I jumped in. Im clueless

It’s all good. I probably confused the matter with lack of details. Just looking at one course at a time (Mayakoba classic now), so the par won’t matter. Fitting types of golfers to courses. Interestingly enough, I don’t think there has been an MDF here, so don’t have to deal with that now. I did delete some WDs and DQs, though.

Thanks, I’ll take a look. Yea, this might get deleted.

Thanks for this site. I was looking it over and bookmarked it for future possibilities.

no more MDF’s ever but I laughed when I typed it because it was relevant to past events.

FWIW, my initial tests of a few models found random forest to be the best predictor. The two most important features leading up to the Mayakoba were Driving Accuracy and Bogey Avoidance. We’ll see how that plays out.

And, for clarity, this doesn’t necessarily mean that those with high DA and BA during the tournament will perform the best; rather, it’s those with high DA and BA leading up to the tournament will perform the best. So, to test, we would need to look at those who perform well, then go back and check where they stood before the tournament.

What timing! Might be able to scrape more of what we need from the PGA site now:

https://twitter.com/RobBoltonGolf/status/1192240628723310592?s=09

I was just going to post that all you Data Scrapers are going to be happy