Swing plane
On which plane do you swing, and do you hold it from the backswing to impact? The plane is one of those values that today needs a camera system set up in the room.
Today DeepGolf measures the rhythm: backswing, downswing, the ratio between them. That is the beginning. The same sensors carry more than that, and not only in golf.
In the app
The app measures the timing of a swing: how long the backswing takes, how long the downswing, and how the two relate. Measured at 100 Hz at the wrist, computed on the device. That part is established, and everything else builds on it.
The same signals we use to compute the rhythm today also carry the geometry of the movement: where the hands travel, on which plane the club swings, how evenly that holds across ten balls. These are the questions we are working on. Whatever reaches the app will reach it once it is established.
On which plane do you swing, and do you hold it from the backswing to impact? The plane is one of those values that today needs a camera system set up in the room.
How far back does the swing go, and is it the same distance every time? The way back is part of what decides how much time is left for the way down.
How fast do the hands travel, and when does it get fast? The same total time can come from an even pull or from a late jerk.
What happens in the moment the swing changes direction? It lasts fractions of a second and is too short to feel.
In which direction does the motion set off first after the turn? If the hands come too far forward there, the rest of the swing is already laid out.
Not one good swing, but ten similar ones. How much spread is normal, and when does it become a sign of something else?
A measured value tells you what happened. It does not tell you what to practise tomorrow. That is why our work does not stop at measuring: out of the values of many shots, a plan should emerge that fits your swing.
Goal
What should come out is not a measurement log, but one sentence: this is what to practise next.
The plan grows out of your shots, not out of a general teaching doctrine. Two players with the same number can need different exercises, because for each of them the number comes from a different movement.
Swing plane and hand path can be measured today: in a lab, with camera systems, with a coach standing next to you. That is precisely the problem. Anyone who wants to see these values needs an appointment, a facility and usually someone to read the numbers.
We want to get them from the sensors already sitting on your wrist.
Not more accurate than the lab, but available: after every swing, on every course.
For most golfers that is the difference between once a year and every round.
Research
A model that infers the course of a movement from wrist sensor signals is not tied to golf. Golf is our first case: a movement that repeats, that people want to improve, and for which there is a clear yardstick.
The same approach carries over to other sports, and to movement analysis in medicine, which today likewise requires a specialised lab. That is research, not a product. Should anything come of it, it will also have to be judged differently from a golf app.
If you work with movement data, in sport or in research, write to us. We would rather talk with people who measure than about them.