What counts as a good serve score?
Short answer: for a club player, the 60s and 70s. But the total is the least useful number on the screen, and here is why.
What the number is measured against
FormCourt grades each measured checkpoint on a curve whose centre is the elite band — the range published serve-kinematics research finds in tour-level players.
That framing matters. 100 does not mean “no mistakes.” It means “on this measurement, indistinguishable from a professional serve.” Nobody reading this needs that, and chasing it on every checkpoint at once is how people end up rebuilding a serve that was mostly fine.
A clean, repeatable club serve with a real leg drive and a decent contact point typically lands in the 60s and 70s. Serves in the 40s are usually not “weak” — they are usually missing an entire phase, most often the load.
The shape matters more than the total
Two serves both scoring 68:
- Serve A: every checkpoint between 62 and 74. This is a coherent motion with no single failure. Improving it means raising the whole thing, slowly, and it is genuinely fine to leave alone for a season.
- Serve B: eight checkpoints in the 80s, two at 21. This is a good serve with a broken link, and fixing the two will move the ball speed far more than anything you could do to serve A.
Same total. Completely different instruction. The total is a headline; the bars underneath it are the actual product.
Why some serves have no total at all
A serve gets marked PARTIAL when one of the four scored categories came through with no grade, or when fewer than seven of the eleven measured checkpoints were readable.
That is not the app failing gracefully — it is the app refusing to hand you a comparable number built from half the evidence. The measurements it did get are all still shown. It just will not average them into something that looks like the score you got last week under better light.
If you are seeing PARTIAL a lot, it is almost always framing or light: how to film a serve properly.
What to actually do with your score
- Ignore the total for the first month. Look at which two bars sit twenty points below the rest.
- Work one of them. Not both. The plan picks it for you and schedules it around the days you already play.
- Re-film in two weeks under the same framing. The comparison is only meaningful if the camera setup was the same.
- Expect the total to dip once. Changing a real mechanic makes the rest of the motion temporarily worse. That is a sign it is working, not a sign to stop.
The score exists to point at something. Once it has pointed, the useful work happens off the screen.