Reconciling Laser/Doppler Sprint Speed With Timing Gates

A sprint coach pulls up the laser printout from Tuesday's session: 9.62 m/s peak velocity, recorded 34m into a 40m sprint. The same session's timing gates, set at 30m and 40m, show a 30-40m split of 1.09 seconds, which converts to a 9.17 m/s average for that segment. Same athlete, same sprint, same afternoon, and the two devices disagree by 0.45 m/s, enough to shuffle a squad's speed rankings by half a position. The usual instinct is to assume one unit drifted out of calibration, and the next week gets spent re-leveling gate beams or re-flashing laser firmware chasing an error that was never there.

Almost always, both devices are doing exactly what they were built to do; they are just answering two different questions. The laser reports the single fastest instant anywhere in its continuous velocity trace. The timing gate reports the average velocity across the entire distance between two fixed beams, and an average across a 10m window can never exceed the fastest instant that occurred somewhere inside it. Once you know which question each device is actually answering, the mismatch stops being a mystery and becomes an arithmetic fact you can predict before you run the test.

A Doppler radar gun or laser system samples continuously, somewhere between 35Hz on an older radar unit and up to 100Hz on a tracking laser, building a full velocity-time curve for the sprint. Peak speed on that printout is a single point lifted off the curve, whatever the highest instantaneous reading happened to be. A timing gate pair cannot do this. It has exactly two events to work with, a beam break at entry and one at exit, and the only number that falls out of two timestamps and a known distance is displacement divided by elapsed time, an average velocity by definition, and it stays an average no matter how tightly the gates are spaced.

This matters because average velocity across any interval is capped by the fastest instant somewhere inside it. If an athlete is still accelerating at the entry beam and starts to level off before the exit beam, the true peak happened in the middle of the split, and the split average has no way to report a number as high as that peak, even with a flawless timing system. The gap is not noise waiting to be calibrated away; it is what division does to a curve that is not flat.

How large the gap gets depends on how much velocity changes across the split and where the true peak sits inside it. A 5m window straddling the peak, near-flat on either side, might show an average within a percent or two of the laser's number. A 10m window catching the last stretch of acceleration on one side and the first stretch of deceleration on the other routinely shows a 3-6% gap on team-sport athletes. Stretch the window to 20m, capturing a meaningful chunk of deceleration for most non-specialists, and the gap can push past 10% from arithmetic alone, before any device error gets added on top.

Haugen and Buchheit (2016), in their methodological review published in Sports Medicine, walk through this exact distinction across the sprint-timing technologies used in team sports. Their central point is that continuous devices such as radar and laser output a genuine instantaneous maximum, while photocell and timing-gate systems output split times that convert to an average across the gate-to-gate distance, and treating the two as the same variable produces precisely the unexplained gap that sends staff back to re-calibrate equipment that was never wrong. They describe the gap as systematic rather than random: it shrinks toward zero as split distance shrinks and grows as split distance grows, which is why a flying 10m split and a flying 20m split from the same run will not show the same-size discrepancy. Their review synthesizes findings across multiple timing setups rather than running one controlled comparison of its own, so it does not hand practitioners a single correction factor; the actual gap on a given set of gates depends on gate spacing, the athlete's acceleration profile, and where in the run the gates sit.

Simperingham, Cronin, and Ross (2016), reviewing sprint-speed testing methodology for team-sport athletes in the same journal, add the equipment side. Laser devices sampling up to 100Hz and radar units in the roughly 35-46Hz range both build a fine-grained velocity-time curve, giving typical peak-velocity measures a trial-to-trial coefficient of variation in the low single digits, while timing-gate-derived average velocities carry a separate reliability profile tied to gate height, beam width, and trigger sensitivity rather than sampling rate, since a gate pair is not sampling a curve at all. Their review flags that pooling laser, radar, and timing-gate studies into one universal agreement figure is not appropriate, since the underlying systems are not measuring the same construct, the same conclusion the arithmetic above reaches independently. The limitation worth carrying forward: most underlying studies used a single brand of each device type, so an exact disagreement percentage from one lab's comparison will not transfer intact to a different manufacturer's units without a local check.

You need one continuous-output device to establish exactly where the true peak occurs, a set of timing gates configured at more than one split length, and distance marking precise enough to know which gate pair actually bracketed that peak.

Set every gate at the same beam height across every athlete and session, roughly hip height for adults. Raising or lowering the beam changes which body segment breaks it first, shifting the effective split by a few centimeters that never shows up in the recorded distance.

1. Warm-up (10-12 minutes): Standard dynamic preparation finishing with two build-up strides near 90% effort.
2. Instrument the full run: Log the entire sprint, typically 30-40m, with timing gates at every 10m along the same course.
3. Run three maximal trials: Full recovery between reps, at least 3 minutes, with gate height and radar or laser position identical across all three.
4. Locate the true peak: From the trace, read off the exact distance at which velocity peaked in each trial, not just the peak value.
5. Identify the bracketing split: Note which 10m segment contains that distance. This is the only split fairly compared to the continuous peak.
6. Compute the predicted gap: Estimate the average the trace implies across that same segment and compare it to what the gates recorded. A close match confirms the gates work correctly; a larger mismatch points to a hardware or placement issue.
7. Repeat with a second split length: Recompute using 5m and 20m segments from the same trace to watch the gap move as the math predicts, separating a real device fault from the expected effect of split length.
8. Discard criteria: Drop any trial with a false start, a beam broken by an arm swing ahead of the trailing leg, or a mid-trace dropout.

Total time runs about 30 minutes for three trials plus setup, most of it recovery between maximal sprints.

Once the true peak is located and the bracketing split is identified, the pattern of the gap points to a specific conclusion.

Neither device is correct here, and the fix is not picking a favorite. Use the continuous trace, laser, radar, or IMU, whenever the number that matters is a true instantaneous maximum: force-velocity profiling, sprint-mechanics work, anything asking exactly how fast the athlete moved at their fastest instant. Use timing-gate splits when the number that matters is an average across a defined distance, which is most of the time in practice: a standard 30m sprint test, a repeated-sprint protocol, a return-to-play benchmark over a fixed course. Both answer different questions, and reporting one without naming which is what turns a predictable gap into a false alarm.

Log the split length or sampling method next to any velocity number, the same way you would log wind or surface. A peak speed with no split length attached is not comparable six months from now if gate spacing changes, and a coach reading last season's file cannot tell whether 9.1 m/s was a true peak or a 10m average with the real peak sitting a stride inside it.