The golf data revolution

General Discussion
#1

@LouStagner has taken a job at Arccos, and I’m excited for the continuing evolution in data in golf… Much like I fixed Ballnamic (haven’t gotten a thank you card) I’m going to lay out some suggestions for projects Lou should work on… I look forward to Lou sending me a thank you card.

  1. Suggest a tee box feature. This one is simple and close to Lou’s heart… should be able to use people’s data (both the players and the aggregate) to suggest the “optimal” tee box for players of varying skill… Look at handicap, driver distance and everything else, and point them towards the “correct” tees… Correct is obviously subjective here, but it might be an interesting feature to implement.

  2. On a similar vein, it would be cool if you could build out models for identifying courses players should play well or poorly on… From the simple idea of long hitters do well on big open courses to actually a data driven analysis of “like” players and where they play best… I’m not sure if Arccos has a big enough data set for that, but it would be cool to have a “scoring range” feature for various courses. It could be a useful feature from multiple angles, but also show people what types of holes they struggle on (and possibly why)

  3. This one is a little in the weeds, and based soley on Chasing Scratch, but you could likely optimize tee box selection based on handicap goals… there are some sand bagging implications here, and this ties back into #1… Basically identify “opportunities” in scoring average vs course rating… We won’t get into the ethics of this, and maybe this is only an internal feature… but it’s an interesting rabbit hole to me.

  4. I’m not sure if we can get this granular, but it would be interesting to see if you could show people which strokes gained would improve their score on a particular course… Basically an analysis of how many different types of shots you are hitting compared to the strokes gained… Basically my thought is: A really bad sand player could gain strokes getting better out of the sand OR hitting it into less sand traps… but the improved iron play would obviously help elsewhere.

  5. More of a big picture idea, but it would be cool if Arccos dug into practice suggestions and tracked peoples practice and improvements… give people different drills and see how much improvement they show (obviously you’d have to track their practice in Arccos)… this could drill in on what practice was effective and hopefully help guide people to more effective, actionable practice

Hope that helps! enjoy the new gig.

2 Likes
#2

Lou has already sent a thank you note… he’s a class act like that.

1 Like
#3

I’ve been thinking about this more… and I know Arccos touts it’s “AI driven” process, but it seems like with the right data and some computer power, you could brute force “best targets” on a golf course with Arccos…

It will have the player’s dispersion in the app, it knows the wind and temperature… so it can basically gaemplan every shot… Driver has a large dispersion circle and it can calculate the odds, figure out the optimal target, and then look at what the next shot will look like from every area on the circle…

Basically, this could be a hyper specific DECADE type strategy, where it not only knows WHERE you miss, it knows which misses punish you the most. i’m sure Lou is already on top of this, but it will be interesting to have a digital caddy who knows golf strategy.

1 Like
#4

It doesn’t take much to figure out where I’ll miss. Wherever is the worst spot to miss, that’s likely where I’ll leave it. No need to take into account wind, shot shape, et al.

1 Like
#5

It would be interesting indeed to get this kind of feedback, but I have a feeling that it will be a long time before the Rules will change to allow the use of electronic devices provide real-time game-planning.

1 Like
#6

I hope I’m not about to offend anyone. The “Golf Data Revolution” reminds me alot of Paul DePodesta who started the Analytics revolution in Baseball. When I was taught this game 38 years ago, I had mentioned that my first 2 lessons I never touched a club, it was mostly conversation. I was told to buy a pocket notebook and start making notes. I know many professional golfers did the same for years. Many people often wondered what caddies and players did back in the day, you bet it wasn’t just toting a bag. I would say until fairly recently, the avg player good or bad, really didn’t keep detailed notes like are available today, or could hear the conversations between caddies and players like you can hear today. If I said one of the most important and influential people over the last 50-60 years in this sport was Jack Grout. Because even the GOAT needed tweaks. I think it’s great all this “stuff” that people are swearing by…put it to good use…I’ve been using it all along…All things being equal though does one think Babe Ruth really cared about his OPS or Ben Hogan cared about his GIR? They knew what they were doing. If BH didn’t lose 7 years in his prime and maybe could improve his putting by .1…He would have won 25 majors…at the end of the day the player still must execute.

#7

This is something I actually went back and forth with Arccos a couple of years ago. I didn’t think that showing the wind would be legal. Their response was the wind reading is set prior to the round and doesn’t change during so therefore it was legal. Honestly I didn’t take the time to research it and just turned it off during tournament rounds. So my question becomes if the pin positions/tee locations are set prior to the round in the app and then the data is used to select the best targets would that be legal? As long as the data is not adjusted during the round.

#8

Check 4.3a(3). As you describe it, I think it would be legal

(3) Information Gathered Before or During Round.

  • Allowed.
    • Using information that was gathered before the round (such as playing information from previous rounds, swing tips or club recommendations), or
    • Recording (for use after the round) playing or physiological information from the round (such as club distance, playing statistics or heart rate).
  • Not Allowed.
    • Processing or interpreting playing information from the round (such as club recommendations based on current round distances), or
    • Using any physiological information recorded during the round.
2 Likes
#9

You are 100% correct… the best players figure this out on their own. Tiger is a perfect example. Scott went back and looked at Tiger’s strategy and you can basically conclude he does exactly what the math / DECADE preaches. It’s called DECADE because it’s decades of experience in an app…

The interesting area to me is that we can start to find the trends in our own games and separate them from feel and belief. I played with a three handicap today who said he hated to leave 40 yard shots into greens and would rather have 100 yard full swings in… He’s a smart guy and a decent stick… but he’s behind on the data.

The funny part is, he might even be right (or at least not totally wrong) but the conclusion isn’t to hit it to 100 yards, it’s to take some time and learn to his a short wedge shot. If someone said they had better tempo on 20 foot putts and would rather be 20 feet than 10 feet, we’d look at them like they were crazy… but someone saying they’d rather be 180 feet farther from the hole is a constant on the golf course.

I think of it like cooking a steak… a good chef doesn’t need a thermometer. They will know from the feel, look, smell when the steak is done. it’s an incredible skill and one that can be learned with patience and practice… or a home cook can use a thermometer, and know what temperature they want to pull the steak out. Sure, you lose some artistry, and the great chefs will likely still make a better steak… but the amateur chef can make a SIGNIFICANTLY better steak with less errors because they now have access to information they did not before.

I think the big missing piece, and what Jon is working towards and what both Lou and Scott could do better is helping people make better decisions based on the data.

Lou loves to post “Manage your expectations” on twitter with average scores from certain distances or situations… it’s great information and awesome to see… It take a scratch golfer 3 strokes on average to hole out from 100 yards. What does that mean for me, a 5 handicap? It probably means I should try to make a par in this situation and make sure I’m on the green and in a position to 2 putt.

This all skews back to the concept of aiming for the middle of the green… Which is great advice and something I wish I remembered more often… but at the end of the day, the data shows us our inconsistency and can hopefully be used to improve our play… Some people do it better than others naturally… but I sure don’t! I wonder how much better I would have been as a young golfer if I had had the data to improve my game.

1 Like
#10

How? The wind for the nearest publicly reporting weather station (which is where I’d guess Arccos would be getting its data from) can be different than that along the intended flight path of the ball. Winds aloft magnitude, I’m guessing, is going to be quite different than even what you’d measure at ground level with something like a Kestrel anemometer. Even at ‘ground level’, winds can twitch unpredictably, which is why long range rifle competition ranges have wind flags all along both sides of the range, from the firing line to the targets.

1 Like
#11

As I understand it, rangefinders are kosher in competition, but ones that also calculate slope aren’t, right?

#12

Absolutely correct. You can read the regulations on uses of “other” equipment in Rule 4.3. Wind and weather data, in response to your previous post, is discussed in 4.3a(2). Oh, in general I don’t reference “in competition”, many people prefer to play within the Rules at all times, not only in competitions. I know that doesn’t apply to everyone, but I’d rather not get used to the “crutch” of a slope measurement and then find out that I really want it (or need it) when I need to play by the Rules.

#13

Yeah, you are correct… but Arccos can know what the wind is approximately and how it will change your dispersion pattern… I don’t think they can say “hey, there is a left to right wind blowing at 5 mph, so aim 4.6 yards left of your target” but I do think they can look at data from thousands of rounds and figure out how much the wind and weather change someones dispersion patterns.

To any earlier point, you still have to hit the shot… and “safest” target is much different from a safe target…

I don’t know! I don’t know what data is being collected. I don’t know what parts of it are actionable… When MoneyBall was written, the A’s were the only team signing guys who weren’t getting OUT… now teams are figuring out how to add more spin to player’s pitches because they have the data showing it’s harder to hit them.

Golf is a unique sport in that their are millions of amateurs who play every weekend and maybe want to get better, but don’t necessarily know how or where to put in the work… or don’t have time to put in the work… There are a ton of variables out there, and it’s different for every player… Arccos has a unique position in the market in that it collects a ton of accurate data for thousands of golfers every day…

How they use that data is yet to be seen… but I’m excited to see what Lou does with it. I think we are about to move into a more data driven game and people are going to know and understand their own stats… or those who want to get better are.

1 Like
#14

I’m pretty (or completely) ignorant about shooting events. I know that in golf, you are not allowed to use any kind of device to measure the actual wind at the course or use anything artificial (like power) to gauge the wind. Is that the same in shooting?

1 Like
#15

See im one of those guys that PREFER to lay up to a yardage on a par 5 depending or if I cannot hit into a bunker because I hit a wayward tee shot and my choice is get to 100 or possibly put myself into position to make 6 or worse. That’s just managing my game. That DOES NOT mean I hit a drive and have 170 left that I chip down to 100, of course I take a swing with my 5 or 6 iron. If I can’t get there in 2, I prefer 95-105. Not 40. Y? Because I cannot spin the ball and control my shot comfortably. How do I know? Well using data that I have compiled after hitting thousands of shots. I end up 5-7 ft closer with a full SW or GW in than I do with something I have to shall we say manipulate my swing. That’s what the data tells me. Not because I say so, the data says so. For your 3hdcp player…did you get into his head as to why? Like I said in one of my posts…ZJ won the Masters laying up to 100 yds on every par 5 and birdied 11 out of 16. Never even took a whiff of going for a green in 2. He could have, but at that time he was a pretty good wedge man. Still is. BUT Y? He sat with his coaches and caddy, game planned his approach to the week, the data told him…take all trouble out of play, the full wedge game and putter were looking strong…for those 4 days…he was the best in the world.

#16

Nope, in fact “doping wind” is an essential part of the sport. Every competitor knows how fast their particular bullet leaves the barrel, and what X MPH of 90 degree wind means for wind drift at Y 100 yds or meters. It’s not as much as you’d think, but the X ring (tiny ring inside the 10-ring, useful for breaking ties) is awfully small. Competitors know that a range flag at this angle, means about a 5 MPH wind, while a billowing flag means a lot more. Dial in 12 clicks left, note the adjustment in your range book, regain your position, and shoot before the wind changes…

In hunting, you have to use other cues: smoke rising, trees bending, grass waving, etc… Most of which is vague enough to induce the hunter to get closer… .Mil, I’m reasonably sure, has tech that can interpret Doppler shift on a return laser bounce, or read actual dust mote relative motion along the beam path. At least, that’s how I’d do it. Weather radar can give differential speeds on closely positioned blocks of air, which is how we ‘see’ incipient tornadoes. I don’t see why the same concept can’t apply closer.

2 Likes