How Honest Jump Measures a Jump

A measurement takes about fifteen seconds and needs nothing but a phone and somewhere to stand. What happens during it, how to hold the phone, and where each number in the report comes from.

You hold a phone against your chest and jump. Fifteen seconds later there is a force curve on the screen. This page says what happens in between.

What you need

• Your phone.
• A flat, firm floor. Not a mat, not grass.
• Room to land where you took off.
• Sound on. The app talks you through it.

What happens

1. Press Start. Then put the phone where it is going to stay.
2. Hold the phone flat against your chest, with both hands. Not in a pocket, not in one hand at your side.
3. Stand still. The app is working out which way is up.
4. Wait for the count. Five down to one, then "jump".
5. Jump. Down, then up, in one movement, as high as you can.
6. Stand still for two seconds. Walking away early cuts the recording off in your landing.
7. Read the report.

Holding the phone

Firmly, flat, against the middle of your chest, both hands, through the landing.

This is the part that matters most. The app reads the phone's acceleration and takes it to be yours, so a phone that slides is a measurement of the phone. Both hands also keep your arms out of the jump, so every measurement is made the same way.

What the phone can sense

A phone has an accelerometer and a gyroscope. It senses how it is moving and which way it is tilted. It does not sense force, which is what a force plate measures by having a body push down on it.

What the paper found

The force curve is the shape of the whole jump: it starts at your body weight, dips as you drop, rises through the push, and ends as your feet leave the floor. Almost everything a force plate reports is read off it.

The paper behind this app predicts that curve from the phone's motion, using a neural network built for signals that run over time.

Nineteen adults each did ten jumps holding a phone at chest level, the position this app asks for, with two force plates recording the true force alongside. The model was then tested on jumps it had never seen.

The predicted curve stayed close to the measured one, and for most of the measures read off it the error came out under five per cent ¹. An erratum to this study has also been published ².

What Honest Jump does

Peak force is the highest point of that predicted curve.

The curves and the phases in your report are read off the same curve.

Jump height and take-off velocity come from how long you were in the air instead. mRSI divides that jump height by the time from movement onset to take-off. What the numbers mean sets out those methods.

What the web app validation found

In internal comparisons of this web app with force-plate measurements, mean errors were about 6% for the force waveform and discrete variables. These web app results are separate from the published study above.

Where a reading goes wrong

The published study and the web app checks used laboratory measurements, on a hard floor, with the phone held the way this page asks. Outside that:

• A loose grip. The biggest one, and the only one you control.
• A very small jump. Nothing under 350 milliseconds of flight is reported at all, which is about 15 centimetres. Below that a jump and a shuffle look the same.
• A soft floor. It gives at the take-off and again at the landing.
• Landing somewhere else. A step forward is movement the app has to read through.
• Stopping early. Those two seconds are part of the measurement.
• One jump. Three and their average is a steadier number.

Read next

• The countermovement jump is the test itself, and what it is good for.
• What the numbers mean is the report itself, one number at a time.

References

1. Kim H, Kong T, Han H, Ha D, Kipp K (2026). Smartphone IMU-Based Prediction of Vertical Ground Reaction Force During Countermovement Jumps: A Deep Learning Approach. Journal of Strength and Conditioning Research. Advance online publication. doi:10.1519/JSC.0000000000005597
2. Kim H, Kong T, Han H, Ha D, Kipp K (2026). Erratum: Smartphone IMU-Based Prediction of Vertical Ground Reaction Force During Countermovement Jumps: A Deep Learning Approach. Journal of Strength and Conditioning Research. Advance online publication. doi:10.1519/JSC.0000000000005699