How to Measure Hurling Striking Speed with Radar and IMU

A county coach clocks a sideline cut at 148 km/h on the radar gun and moves on satisfied, until the same player's ground strikes off the left side come in fifteen kilometres an hour slower for reasons nobody can point to. The radar reads sliotar speed off the bas beautifully, but it has no idea whether that speed came from a strong hip-shoulder turn arriving early, or a last-instant wrist break doing most of the work — two different faults with two different fixes.

Stick head speed at contact collapses two separate power sources into one figure: rotational drive built through the hips and torso from backswing into downswing, and the wrist break — the rapid uncocking of the bottom hand in the final tenth of a second before the bas meets the sliotar. A player can lose ground on one and quietly compensate with the other for weeks before a coach notices anything beyond noting the numbers are down a bit. This guide sets out a field protocol for measuring hurley stick head speed with a Doppler radar unit cross-checked against a three-point IMU array — pelvis, torso, and lower hand — that decomposes how much of that speed comes from body rotation versus wrist release, plus what the closest published research on hurling equipment and IMU rotational biomechanics says about reading the split.

Stick head speed at impact is worth tracking — a faster bas at contact generally means a faster sliotar off the strike. The problem is a single peak-speed reading, from a radar gun or a basic hurley-mounted accelerometer, averages two mechanically distinct contributions into one figure without saying which one did the work on a given strike.

A ground strike built mostly on hip-shoulder rotation and one built mostly on a late wrist snap can arrive at nearly the same stick head speed through completely different sequencing. A player whose rotational drive has quietly stalled — often from a lower-back or oblique niggle never reported as an injury — will mask it by breaking the wrist earlier and harder, and the radar will show a normal number right up until that compensation runs out. By the time overall speed actually drops, the rotational deficit has usually sat unflagged for weeks.

Stick head speed here is the linear velocity of the bas — the broad striking face of the hurley — the instant before it meets the sliotar. A radar unit reads this indirectly, clocking the sliotar's own exit speed after it leaves the stick, folding in the ball's coefficient of restitution alongside whatever the player did with the swing. An IMU mounted near the bas reads the stick's own angular velocity directly through the swing and converts it to linear head speed via the shaft-plus-forearm lever arm, before the ball's bounce enters at all — why this protocol treats IMU stick speed as the primary number and radar speed as a secondary cross-check.

Splitting stick speed by source works the way sports scientists separate rotational and release contributions in golf clubhead speed. A pelvis IMU and a torso IMU track how fast the hips and shoulders turn and how far the shoulders lead the hips — the rotational engine of the strike. A third IMU on the lower forearm, above the bottom hand's wrist, captures the stick's total angular velocity at that joint. The gap between what torso-pelvis rotation alone predicts the stick speed to be, scaled through the lever arm, and what the wrist sensor actually measures at impact is the wrist's own contribution — the late snap layered on top of whatever rotation had already been built.

Three IMUs are needed for the full split, plus a radar unit for the ball-exit cross-check. Strap one IMU at the belt line over the sacrum for pelvis rotation, one over the sternum for torso rotation, and one on the lower forearm above the bottom hand's wrist, tracking wrist flexion-extension and radial-ulnar deviation together. Position the radar roughly 3–4 metres behind the striker, angled along the ball flight path.

Calibration Sequence

1. Neutral stance hold (3 seconds): hurley still at address, setting the zero-reference orientation for all three sensors.
2. Torso-pelvis separation check: the player rotates the shoulders fully while keeping the hips still, then reverses — giving a reference range for how far this player's torso can lead the pelvis, the X-factor principle from golf.
3. Wrist-break calibration: three slow-motion swings with an exaggerated wrist break, confirming the IMU distinguishes a genuine break from ordinary grip movement.
4. Three full-speed calibration strikes: ground strikes off a tee, confirming impact detection on both the stick IMU and radar — re-strap after two consecutive misses.
5. Sampling check: confirm all three IMUs run at 400Hz or higher, above 50% battery, and the radar has a clear line to the strike zone.

Run the sensors through a mixed session of 20–30 strikes rather than one strike type — the split only means something with enough data to separate a real pattern from a single mistimed swing. Mix ground strikes, sideline cuts, and doubled (overhead) strikes in an order the player cannot fully anticipate, since strike type changes how much time the body has to build rotation before contact.

In-Session Steps

1. Complete calibration, then run the mixed session. The app logs stick head speed, peak torso and pelvic rotational velocity, and wrist angular velocity per strike, alongside radar ball-exit speed.
2. Confirm strike-type classification after each ball — signatures learned from calibration.
3. Calculate rotation-predicted stick speed from torso-pelvis kinematics and the lever arm, subtract it from the measured stick speed; the remainder, as a percentage, is the wrist-contribution share.
4. Group strikes by type and compute mean stick speed, mean wrist share, and SD per type.
5. Log both figures against the player's own baseline from at least three prior clean sessions per strike type.

A player's own baseline share matters more than any fixed target. Two sound hurlers can show shares 10–15 points apart on the same strike type and both be fine — a coach watches whether a player's own share drifts over a block of sessions, which usually flags a rotational issue before it shows up as a raw speed drop.

These bands come from practical field session data across club-to-inter-county players rather than a single controlled study — no published research has yet run this exact wrist-versus-rotation decomposition on a hurling strike, so treat the ranges as starting reference points to compare a player against their own history, not a fixed grading scale.

The pattern holds even where exact numbers vary: strikes giving the body more time to load a hip-shoulder turn — the ground strike, most of all — lean more on rotation, while a doubled strike on the volley leans harder on a late wrist snap because there isn't time to build the same rotational base. Hurling equipment research has reported sliotar exit speeds as high as roughly 160 km/h (44 m/s), well above the stick head speeds here — the sliotar's own rebound amplifies speed on its own, which is why radar and IMU numbers are complementary, not the same measurement twice.

Two published studies sit closest to what this protocol measures, though neither tested a live player's wrist-versus-rotation split directly.

Alsakarneh, Bryan, Cotterell, and Barrett (2012), in Sports Engineering, fired a sliotar into a hurley from an air-cannon rig at controlled speeds, filming 32 impacts at 12,500fps across sliotar brands, hurley brands, and impact locations. Sliotar brand was the single most significant factor in peak impact force, with up to a 27% difference between brands at the same inbound speed. The limitation: the rig struck a stationary hurley at a fixed, machine speed — it says nothing about how a live player generates stick speed through rotation or wrist release. What it establishes is that ball-exit radar speed measures equipment variation as much as swing quality, which is why this protocol treats stick-head IMU data, not radar speed, as the primary outcome.

Kim and colleagues (2023), in Sensors, validated a two-IMU setup — T1 and L4 vertebrae — against 3D motion capture for golf rotational kinematics in 36 golfers: near-perfect agreement for upper torso rotation (ICC 1.00) and pelvic rotational velocity (ICC 0.98), somewhat lower for pelvis-obliquity (ICC 0.91). This is the evidence base for using body-mounted IMUs, rather than a motion-capture lab, for the torso-pelvis sequencing this protocol borrows. The limitations: golfers on a different implement and plane, a sample of 36 rather than hurlers, and the authors' own note that agreement narrowed for unusual swing styles. The wrist-contribution percentages here are a within-player trend to track, not a validated absolute figure — no study has yet run this decomposition on a hurling strike.

Key References

• Alsakarneh, A., Bryan, K., Cotterell, M., & Barrett, J. (2012). The influence of equipment variations on sliotar–hurley impact in the Irish game of hurling. Sports Engineering, 15(4), 177–188.
• Kim, W., et al. (2023). Validation of Inertial Measurement Units for Analyzing Golf Swing Rotational Biomechanics. Sensors, 23(20), 8433.

The wrist-versus-rotation split earns its place on a fixed schedule, not as a one-off diagnostic after a run of poor striking in matches.

• Every striking session: full protocol across mixed strike types, updating each player's baseline stick speed and wrist-contribution share.
• Pre-season block: three to five sessions purely for baseline-building — a share measured on too few sessions moves around more than the player's actual technique does.
• After a lower-back or oblique complaint: re-test before full volume returns — a share still above baseline signals the rotational engine hasn't fully returned, even if the player feels fine.
• In a run of inconsistent match striking: pull the last few sessions' figures by strike type before changing anything technically.