CASE NO. 04 · SPORTS TECH / WEARABLES · CLIENT: 4 SECOND FOOTBALL

A football that coaches the quarterback.

A quarterback's decision speed is one of the hardest things in the sport to put a number on. 4 Second Football embeds motion sensors in the ball itself, measuring release time and decision speed. The app turns every rep into coaching.

4secondfootball.com →

What we delivered

Embedded motion sensorsBLE streamingEdge ML (catch-to-throw timing)Mobile appCoach's dashboardDrill analytics

The shape of it

A sensor in the ball, intelligence in the pocket.

The client

4 Second Football makes a quarterback training system: a smart football paired with an app that gives real-time feedback on release timing and mechanics. It is used by individual quarterbacks, by high-school and college programs, and by coaches running camps.

The friction

Quarterback decision-making is one of the hardest things in the sport to measure. Practice never runs at game speed. And nothing on the field times the window between catch and throw, so every rep leaves the field as an opinion instead of a number.

The constraint

The instrument is the ball itself.

  • Sensors must take being thrown and caught
  • Sub-second timing precision
  • BLE from a moving ball
  • Zero perceptible weight difference

Engineering highlights.

01

Sensors in the ball

A gyroscope and accelerometer sit inside the football, streaming motion data out over BLE. The instrument rides inside the thing it measures.

02

Catch-to-throw on the edge

The catch-to-throw timing model runs on-chip and on the phone. There is no round-trip to a server. The window that matters gets measured where it happens.

03

The coach's dashboard

The dashboard follows a quarterback across sessions. Individual reps add up into progress a coach can point to.

Inside the build.

SMART BALL(GYRO+ACCEL)BLEQB APPCLOUDCOACH DASHBOARDCATCH-TO-THROWTIMER
Fig. 01·The four seconds that matter are measured before the cloud ever sees them.

Finding the throw in the noise.

Sensor data from a football in play is violently noisy. The model has to pull a clean start and a clean stop for the catch-to-throw window out of raw gyroscope and accelerometer streams. It runs on the device. There is no server round-trip, and the timing has to land fast enough that the feedback feels instant on the field.

Have hardware that needs to feel like magic?