I can see DG’s data and expertise doing some quite interesting stuff here, both for purists who just want to know more and serious punters who probably struggle to balance skill with frequency. As an example, how many 2nds, 5ths 10ths accumulate to earn more than a winner. Come to think of it, l’d guess that more than a few players inform their choices with considerations like expected field strength and purse. (Tangentially, any idea as to the ‘quality’ of winners, top 5s, etc. relative to purse (or even prestige). An analysis of explanatory variables like field strength, tornie choices in previous seasons updated for present season. Could there be explanatory information in that history like # played, DG and FedEx rankings, purses, intra-seasonal preferences like # played as season continues, courses, locations, weather. (I’m just freewheeling here - the Courchenes could find many more. And even if a projection capability is too intractable for reliable outcomes, it would sure be interesting to see the pure historical breakdowns by a bunch of factors.
Just some thoughts for the resident brains trust!
I think Data Golf does a Fedex Cup projection on the last event before the playoffs and also during the playoffs
I know about that. But I’m talking about something much more extensive that, for example, could provide projections before the season begins.
First, before I forget, we have a blog post looking at exactly the tangential point you mention!
To your main point: we like the idea, for both of the reasons you mention, too. We actually tried to build something around this last fall. The original motivation was more for the purists (and also for the professionals themselves – we had a couple players / caddies reach out inquiring about the possibility of this).
Our plan was to have a weekly-updating page that would have live season-long FedExCup (and earnings, and any other season-long variable you’d want) projections for 2021, but we hit a couple big roadblocks when we were working on it in the fall that sunk the project. The approach we wanted to take was to actually simulate the remainder of the season (i.e. simulate every event). The hardest part about this simulation is predicting who will play in which events. A lot of the things you mention are related to this: e.g. which tournaments a guy has played in past seasons. I did some preliminary work collecting data to build a model like this (the PGA Tour priority rankings each week for the last few years was one important piece that was hard to get). There are many intricacies with the OWGR rankings and PGA/KFT priority rankings that make it difficult to come up with a good model, and then you also have player-specific variables that we don’t have great data on, like the location of their home base, injuries, etc, that make it harder to predict a player’s starts. Obviously to simulate things from the start of a season, a rough model would suffice. But we got it in our heads that we wanted this page to be able to update throughout the year (in a reasonably automated way), and as you get closer to the end of the season, your predictions of who starts in each event become very important. Looking back I think we we became too concerned with minor details, and should maybe have just plowed ahead with a rough model. We will most likely revisit this at the end of this year.
A rough model at the start of the season would be fun to look at.
With weekly updates when we reach Korn Ferry Tour and Fedex Cup playoffs
You could end up with an amazing model for a “One and Done” pool. The pool I am in every year allocates points equal to that weeks winnings but pays 1.5x for majors and -$250k for a missed cut. Side pots for certain parts of the year and just the majors.
The hardest part of planning each year’s picks on the front end is figuring out who will play in each tournament. Purse size and strength of field are huge as well.