Shot Tracking Apps for UK Club Golfers, Compared

Type “golf shot tracking app uk” into Google and you get about forty results that all claim to use AI, and roughly four of them will work with what you actually own: an iPhone or Android handset, possibly a Garmin R10 sitting in a cupboard, and a membership at a parkland course that’s soft from October to April. The gap between the marketing and the reality is mostly a technical one. Some of these tools read sensor data, some read radar returns, and some read video frames, and those three jobs fail in completely different ways.

This page is about which ones read your inputs correctly, and what you should do at the range once they do. For the wider framework of how strokes gained works and why amateurs misread their own numbers, the overview at /strokes-gained/ sets the context this page assumes.

Three different things called “AI” in golf apps

Shot detection is the first. Arccos, Shot Scope and Garmin’s CT10 system all put a tag or sensor in the butt of the grip, log a timestamp and a GPS coordinate, and then run a classifier over the round to decide what was a shot and what was you rummaging in the bag. Arccos quotes detection accuracy in the mid-90s percent. In practice, on a GB&I course with mature trees and a clubhouse that blocks satellites, you will edit two or three shots per round.

Ball-flight measurement is the second. A Doppler radar unit like the Garmin Approach R10 or FlightScope Mevo reads frequency shift off a moving ball and club. It measures nothing at all about your body. Photometric units like SkyTrak+ do it with cameras and strobes instead.

Swing analysis is the third, and it’s where most club golfers get burned. Sportsbox AI 3DGolf, V1 Golf, Hudl Technique and a dozen smaller apps run pose estimation over video frames to infer joint positions. The model is only ever as good as the pixels you hand it, and in November, on a British range, those pixels are usually poor.

Why your phone footage breaks the video tools

Do the arithmetic before you blame the app. A 14-handicap swinging a driver at 95 mph is moving the clubhead at 42.5 metres per second. Record at 60 fps, the default on most phones in standard video mode, and the frames are 16.7 milliseconds apart. The club travels 71 cm between one frame and the next. Your entire downswing, roughly 0.25 seconds from the top, occupies about 15 usable frames.

At 240 fps, the same downswing gets 60 frames and the club moves 17.7 cm per frame. That’s the difference between a pose model tracking a shaft and a pose model guessing. Impact itself lasts about 0.45 milliseconds and you will never capture it at any consumer frame rate, which is why every app that claims to show you “impact position” is showing you the nearest frame, up to 2 ms either side.

Motion blur is the second killer, and it’s worse in the UK than anywhere the software was tested. Phone auto-exposure under a grey Midlands sky at 4pm in January will happily choose 1/60 s. During that exposure the clubhead smears across 71 cm of sensor. To freeze it you want roughly 1/2000 s, which drops movement during exposure to about 2 cm, and which needs far more light than an overcast afternoon provides. The practical answer is to film outdoors before 3pm between March and September, or use a range bay with bright artificial light and lock the shutter manually.

Rolling shutter is the third. CMOS phone sensors scan top to bottom rather than capturing the whole frame at once, so a fast shaft appears bent. Pose estimators handle this badly and will report shaft lean that isn’t there.

Recording settings that actually work
  Frame rate ........ 240 fps (iPhone "Slo-mo", Android "Super Slow")
  Shutter ........... 1/2000 s or faster, locked (Filmic Pro, Open Camera)
  Camera height ..... hands height, face-on; belt height, down-the-line
  Distance .......... 2.5 to 3.5 m, tripod, phone in landscape
  Framing ........... full body plus 1 club-length of headroom
  Light ............. behind the camera, never behind the golfer

Get those right and Sportsbox AI’s 2D-to-3D reconstruction produces pelvis sway, thorax bend and hand-path numbers that are consistent session to session. Get them wrong and it will tell you your hip turn changed by 11 degrees when all that changed was the cloud cover.

Round trackers: the UK comparison that matters

Shot Scope is worth flagging for UK players specifically. The company is based in Edinburgh, the course library covers obscure GB&I nine-holers that American-built apps miss, and there is no subscription, which over five years is a difference of roughly £600 against Arccos. Arccos gives you better in-round club recommendation and a genuinely good wind model, but it needs your phone awake in your pocket for eighteen holes, and a four-hour round in the rain with the screen off still costs you 30 to 40% of an iPhone battery.

None of these post to your WHS record in Great Britain and Ireland. You still enter scores through ClubV1, intelligentgolf or the MyEG app, so treat the tracker as a parallel performance record rather than a handicap system.

Clippd deserves a mention if you’re at the 6 to 10 handicap end and playing county or university golf. It’s UK-built, it ingests Arccos and Shot Scope data alongside manually logged practice, and it rolls everything into a single Player Quality score per shot. It’s overkill for a 20-handicap and very good for someone with a coach.

Launch monitors on a soggy British range

A Garmin R10 at around £450 is the default budget choice, and its weakness is exactly the thing UK ranges force on you. Doppler radar needs to watch ball flight to compute a trajectory. Outdoors with an open field it has plenty to work with. Inside a netted bay with 3 metres of flight, it is extrapolating from a very short observation window, and carry numbers drift.

Set-up distances are not negotiable. The R10 wants roughly 1.8 to 2.4 m behind the ball and at least 2.4 m of flight. FlightScope’s Mevo asks for about 2.4 m behind and the same again to the net, and it will not give you true spin without the metallic dot stickers. Rapsodo’s MLM2PRO reads genuine spin only with its own RPT marked balls, at around £40 for a dozen, and estimates it otherwise.

Then there’s the ball. Beaten-up UK range stock carries 5 to 10% less than a premium ball, and limited-flight range balls can be 15% down. If your R10 says your 7-iron carries 148 yards on range balls, your on-course carry with a Pro V1 is closer to 158, and if you set your yardages from the range you will come up short all summer.

Same 7-iron, same swing, two set-ups
                        Bay with net (3m)     Open field (60m+)
  Ball speed              109.4 mph             110.1 mph
  Launch angle            17.2 deg              18.6 deg
  Spin                    6,100 rpm (est.)      6,740 rpm
  Carry                   161 yds               153 yds

The ball speed agrees. Everything downstream of it, on the constrained set-up, is a model rather than a measurement.

Worked example: twenty rounds, one 14-handicapper

A member at a Surrey heathland club logged 20 rounds on Shot Scope over a season. Scoring average 88.4, greens in regulation 21%, 33.1 putts per round but only 1.6 three-putts. His strokes gained split against scratch came out roughly: approach -6.2, off the tee -2.8, around the green -3.9, putting -2.1.

The putting number is the trap. Thirty-three putts feels like the problem, and it isn’t. He takes 33 putts because he hits 4 greens a round and leaves himself 74 feet average proximity from 100 to 150 yards, against about 50 feet for a scratch player. Fix the approach and the putting number falls out on its own.

Digging into the approach data, 68% of his misses from that band finished short of the flag. His carry gapping explained it: pitching wedge 118 yards, gap wedge 104, sand wedge 88, 9-iron 133. Anything between 119 and 132 was a full pitching wedge he could not reach or a decelerating 9-iron. That’s a 14-yard dead zone he faced about five times a round.

His practice changed to two things. Thirty balls at a 122-yard target with a three-quarter 9-iron, counting how many carried within ±8 yards. Baseline for a mid-handicapper on that test is around 55 to 60% and he started at 43%. Second, every approach session finished with ten shots where the only rule was “pin-high or long”, to break the short miss habit.

What to do next week

Pick one tracker and commit for a full season, because eight rounds of data is noise and thirty rounds is a signal. Film two swings a month, not two a week, and film them under the settings above so the comparison means something. Put your true carry numbers into your phone notes with the ball you actually play, then add three yards to every club in November when the air is dense and the ground is holding.