Why Your Bay Numbers and Your Course Numbers Disagree
Six mechanisms sit between an indoor carry figure and the ball flight you get on a Saturday. Only two of them are the machine's fault, which is why "get a better launch monitor" solves less of this than people assume.
Every winter somebody comes in holding a printout from a session elsewhere and asks why the 262-yard driver on the page became a 238-yard driver in April. The honest answer is that both numbers are probably correct and they were describing different events. What follows is the full list of what changes between the two, ranked roughly by how much it typically matters in central Indiana.
1. Carry indoors is simulated, always
This is the one that surprises people most and it is not a criticism of any particular device. In an enclosed bay, no instrument observes the ball landing. Flight is terminated by a screen a few metres away. Everything past that point — peak height, descent angle, carry, total — comes from a flight model fed by the measured launch conditions.
The models are good. They are also models. They assume a specific air density, a specific ball, a specific drag and lift behaviour, and no wind. Change any of those assumptions and the output moves without any change in how you swung the club.
The practical implication is that ball speed, launch angle, spin rate and club delivery are the trustworthy indoor numbers. Distance is a derived figure, and it should be read as "what this launch condition would produce under the model's assumptions", not as a measurement.
2. Air temperature and density
A heated bay in February sits around 68°F. The round you are preparing for might be played at 48°F. Cold air is denser, which increases drag, which shortens carry. As a rough working figure, a driver loses in the region of two yards for every ten degrees of cooling, though the effect compounds with ball speed and is larger for high-spin players.
Barometric pressure and humidity move in the same direction and by smaller amounts. Humid air is very slightly less dense than dry air, which is counterintuitive and worth roughly a yard on a driver between a dry winter day and a muggy August afternoon.
None of this is mysterious and all of it is calculable. What causes the disagreement is failing to write down the bay conditions in the first place, at which point the adjustment cannot be applied even in principle.
3. Wind, which is not zero on a golf course
Indoor flight models assume still air. Central Indiana averages something in the region of nine miles per hour of surface wind, and an exposed course magnifies that. A two-club headwind on a par 3 makes any indoor carry figure irrelevant for that shot.
The more subtle version is that wind interacts with spin. A high-spin driver loses far more into a headwind than a low-spin one, which means the correct indoor recommendation for a golfer who plays an exposed course is not the same as for someone who plays a sheltered one, even with identical swing data. We ask where you play for this reason.
4. The hitting surface
A thick indoor mat is forgiving in a specific and misleading way: the club can bounce into the ball after contacting the mat behind it, producing an acceptable strike from a delivery that would have been fat on grass. The captured numbers look fine. The equivalent shot in April is a chunk.
This matters most for irons, wedges and fairway woods, and it matters least for a teed driver. It is the reason we run a natural turf tray alongside the mat and record which surface was in use for each block. A fitting done entirely on a thick mat has systematically under-sampled your bad strikes.
5. The ball
A range ball that has spent a season in a bucket has a compressed core and a scuffed cover. It reads lower spin and lower ball speed than a fresh premium urethane ball, sometimes by hundreds of rpm and several miles per hour. Fitting a shaft against under-read spin points you toward a lower-spinning profile than you actually need, and you discover the error on a windy day in October.
This one is entirely avoidable and it is the cheapest fix on the list. Use a consistent premium reference ball, replace it on a fixed schedule, and record the model on the sheet.
6. You, indoors
The least discussed mechanism and not a trivial one. A bay removes consequence. There is no out of bounds, no water on the right, and no group waiting behind you. Most golfers deliver the club differently under those conditions — typically a little more aggressively and a little more freely.
That is not a reason to distrust indoor fitting. It is a reason to build a session around dispersion and the shape of the misses rather than around the best strike, because the best strike is exactly the thing that a consequence-free environment inflates.
So what should an indoor session actually do?
Log the room before the first recorded swing: temperature, pressure, humidity, ball model, surface. Apply a stated adjustment to the modelled distance figures and print both the raw and the adjusted values so the adjustment is auditable. Build the recommendation on the measured values — ball speed, launch, spin, path, face, strike — rather than on the simulated ones. Use dispersion and descent angle as the primary criteria. Then, in season, go outside and check.
Done that way, an indoor fitting is not merely acceptable, it is better than an outdoor one for most purposes, because the variables hold still long enough for a comparison between two shafts to mean something. A windy Tuesday at a range is more representative and almost useless for isolating a single variable.
The failure mode is not indoor fitting. It is undocumented indoor fitting, where a modelled number is handed over as though it were an observation and nobody can reconstruct what it was made of.
What we do about it
Every session here starts with the room logged. The adjusted and raw columns both appear on your sheet. From May through September you can book a 45-minute outdoor confirmation block at a partner range where the same clubs get re-run and the adjusted figures are checked against what actually happened. We keep every one of those comparisons, because they are the only way to find out whether our own model is drifting.
If it ever drifts far enough to matter, that will be written up here.