Practising With a Launch Monitor Instead of Just Watching Numbers
Walk into any covered bay on a wet Tuesday and you’ll see the same thing. Ball down, swing, head turns to the tablet, small nod or small grimace, next ball. Sixty balls, sixty readings, and nothing that could honestly be called practice. The monitor was doing its job the whole time. The golfer was just narrating.
Here is the thing nobody tells you when you unbox a Garmin R10 or an MLM2PRO: a launch monitor does not create practice. It creates data, and data only becomes practice when you decide, before the ball, what counts as a good one. That’s the whole method. If you want to know how to practise with a launch monitor rather than commentate on one, you set a target window per shot, you call the result yourself, and you keep a score. The screen becomes a referee instead of a mirror.
Reading the screen after every ball makes you worse
A mid-handicapper hitting 7-irons off a mat will have a carry spread of roughly 8 to 14 yards on shots they’d call “decent”, and considerably more once you include the thin ones. Strike location drives most of that. Half a groove low on the face costs you 5 to 8 yards of carry and a couple hundred rpm without any change to your swing at all.
So what happens when you glance at 148 after a stated 157-yard 7-iron? You make a correction. You swing a bit harder, or you try to “get through it”, and the next ball goes 163. Now you correct the other way. That’s not feedback, it’s a feedback loop with no damping, and by ball twelve you’re steering a swing you didn’t have a problem with. You were reacting to one sample from a distribution you never bothered to measure.
Scoring breaks the loop. If your window is 152 to 162 yards and you hit 148, that’s one miss on a card of ten. It doesn’t earn a swing thought. It earns a tally mark.
Build the window out of three numbers, never more
A usable window has at most three constraints, and at least one of them has to be a start line or a dispersion figure. Distance-only windows produce golfers who are brilliant at hitting a net.
For a 15-handicapper with a stock 7-iron carry of 157:
- Carry: 152 to 162 yards
- Offline: within 15 yards of the target line at carry
- Ball speed floor: 104 mph
That third one is a strike proxy. Ball speed correlates with centredness far better than anything else your unit reports, and unlike smash factor it doesn’t get muddied by a bad club speed reading. All three have to land for the ball to count as a hit. Two out of three is a miss. Be strict, because a generous window teaches you nothing.
A worked session: ten 7-irons, and what it looks like on paper
Fifteen minutes, ten balls, a full minute between each. Here’s a genuine-looking card from a session on a Rapsodo MLM2PRO, outdoors, premium ball:
7-IRON | Window: carry 152-162 | offline <15y | ball speed >104
# Carry Offline BallSpd Called Actual
1 149 6L 102.4 miss MISS
2 158 9R 106.1 hit HIT
3 161 4L 107.0 hit HIT
4 144 18L 100.8 miss MISS
5 156 11R 105.3 hit HIT
6 163 7L 106.8 hit MISS (carry +1 over)
7 151 3R 104.2 hit MISS (carry -1 under)
8 157 12L 105.9 hit HIT
9 153 21L 104.6 hit MISS (offline)
10 159 8R 106.4 hit HIT
SCORE: 5/10 CALL ACCURACY: 7/10Two useful things fall out of that card. The score is 5, which is your baseline for the next four weeks. And the call accuracy is 7, which tells you your hands know less about where the ball went than you assumed, particularly on the left misses.
Call it before you read it
This is the single change that does the most work. After each ball, before your eyes go anywhere near the screen, say out loud whether it was a hit or a miss and why. “Hit.” “Miss, thin.” “Miss, started left.”
Then look.
Over a few sessions your call accuracy climbs, and that matters more than your score does, because on the course you have no monitor. You have a ball flight and a feeling. Golfers who practise this way stop arriving at their ball surprised. If your call accuracy sits below 6/10 on irons, you are not ready to be working on swing changes: you can’t yet tell what you just did.
Which AI tools actually read your phone footage correctly
You will want video at some point, and this is where a lot of money gets wasted on apps that produce confident nonsense. A short, honest table:
| Tool | What it genuinely does | Where it falls over |
|---|---|---|
| Sportsbox 3D Golf | Builds a 3D body model from a single phone camera. Pelvis sway/lift/thrust in inches, chest bend and rotation in degrees, hand depth. Gross body measurements are reliable and repeatable. | Needs correct setup: 60fps minimum (120 better), camera at hip height, 8 to 10 feet away, whole body and club in frame at address and at the top. Club and shaft data is far weaker than body data. Roughly £15 to £20 a month. |
| Onform | Capture, auto-trim, frame-by-frame scrub, drawing tools, voiceover you can send to your pro. Excellent as a video notebook. | It isn’t measuring anything. Treat the “analysis” as your own eyes, not the app’s. |
| Rapsodo Impact Vision | Short impact-interval clip synced to the shot’s numbers. Genuinely useful for tying a bad number to a specific look. | Fixed camera position, so it answers one question only. Combine with foot spray on the face. |
| Hudl Technique / V1 Golf | Free or cheap slow-motion with lines and a comparison split screen. | Same caveat as Onform. No measurement, and the auto-detection of positions is unreliable on a mat with a busy background. |
| Generic “AI swing coach” chat apps | Produce fluent paragraphs. | Most are pose estimation plus a language model writing plausible sentences. Feed one the same video twice and check whether the advice matches. Often it doesn’t. |
One technical point that decides everything above. At 100 mph of clubhead speed, the head travels about 29 inches between frames at 60fps. At 240fps it travels about 7 inches. Any tool claiming to measure impact conditions from 30 or 60fps footage is interpolating, and your iPhone’s 240fps slow-motion mode in 1080p is the cheapest real upgrade available to you. Film down-the-line one week, face-on the next, not both in a rush.
Use video only when the scorecard says something is broken. Score drops from 6/10 to 2/10 with the same window? Now film. Score holding steady? Leave the phone in your bag and hit balls.
Two more windows worth stealing
Wedge ladder. Fifteen balls, five each at 50, 65 and 80 yards carry, alternating rather than in blocks. Window is ±4 yards, nothing else. Most 12-handicappers score 4 or 5 out of 15 the first time and are genuinely shocked. Get to 9 and you’ve saved more shots than any driver work will give you.
Driver, for a player who fights the left side. Window: start line 0 to 8 yards right of target, spin under 3000 rpm, ball speed above your own floor (say 142 mph). Ten balls. Notice that direction comes first and distance is a floor rather than a target, which stops the session becoming a speed contest.
Your unit’s numbers are the only ones that count
Set windows from your own device’s history, never from a chart of tour averages. A Garmin R10 estimates spin from a model rather than measuring it directly, so its carry figures indoors can drift several yards from reality. A Mevo+ wants metallic dots on the ball for true spin. Range balls fly 5 to 12 yards shorter than a premium ball on a 7-iron, and hitting off a mat flatters your strike. None of this matters if you’re consistent, because you’re scoring against yourself.
Log the score, the window and the date. Four sessions in, you’ll have a trend line, and a trend line is the raw material for an actual training block rather than a series of pleasant afternoons. If you want the structure for turning those cards into a six-week plan with priorities and a review point, that’s covered properly in Turning Analysis Into a Practice Plan.
Next time you get in the bay, write the window on the back of your hand before you pull a club. Ten balls, call each one, tally it, and leave. If the number on the card is lower than you expected, you’ve just learned something that sixty unscored balls would have hidden.