Video Analysis for Basketball Coaches: What to Actually Look At.
Watch shooting film in a fixed order: base and balance, then lower-to-upper timing, then release point and follow-through. Compare a player's makes against their own misses rather than against a model shooter. Film side-on and from the front on separate sets, since each angle hides what the other reveals.
Film the right angles.
One angle is never enough, and two angles at once is usually impractical, so shoot separate sets.
| Angle | Reveals | Hides |
|---|---|---|
| Side-on, hip height, 3-5 m | Arc, release point height, knee and hip timing, forward drift | Left-right alignment |
| Front-on, hip height | Elbow flare, wrist deviation, base symmetry, shoulder tilt | Arc and release depth |
| Behind, slightly elevated | Ball path relative to midline, drift left or right | Almost everything about timing |
Hip height matters. Filming from standing eye level distorts arc badly enough to make a good shot look flat.
The order to watch in.
Watch the same shot four times, looking for one thing each pass. Watching for everything at once is how coaches end up saying "something looks off."
- Base and balance. Feet width, stance consistency, and where the player lands relative to takeoff. Most left-right misses are decided here, before the ball moves.
- Timing. Do the legs and the release still connect? The classic fatigue signature is the arms firing after the legs have already peaked.
- Release point. Height and repeatability. Consistency matters more than height.
- Follow-through. Held or snatched back, and whether the wrist finishes in the same place on makes and misses.
Then, and only then, watch the ball.
Compare makes to misses, not form to a textbook.
The single most useful film exercise is to pull ten makes and ten misses from the same session and play them side by side. You are not asking "is this good form?" You are asking "what is different?"
That reframing kills the two biggest film-session failures: chasing cosmetic deviations that do not affect outcomes, and correcting idiosyncrasies that are load-bearing for that particular player. Plenty of elite shooters break textbook rules consistently. Consistency is the trait; the textbook is not.
Frame rate and slow motion.
Standard 30 fps is too coarse for release timing. Film at 60 fps or higher and step through the release frame by frame. Slow-motion playback at normal frame rate looks smooth but interpolates away the detail you are looking for.
Mistakes that waste film sessions.
- Too many corrections. One cue. A second cue is a different week.
- Filming only makes. The misses contain the information.
- Inconsistent setup. Different distance, height, or angle each week means you cannot compare anything.
- Public playback for young players. Show the player their own clips privately, or you are coaching their embarrassment rather than their shot.
- Analysis without a re-test. If you never film the same set again, you never learn whether the correction worked.
Where automated analysis helps.
Frame-by-frame film review takes a coach 15 to 30 minutes per player, which is why it happens twice a season instead of weekly. Pose-estimation tools produce joint angles, release height, arc, and variance automatically, which turns "I think his elbow drifts when he's tired" into a number you can re-check next week.
Kima does this on every rep during the session rather than afterwards, comparing each shot to that player's own makes — the same makes-versus-misses logic as a good film session, run continuously. Film review remains better for decision-making, spacing, and context; automated mechanics analysis is better for the repetitive measurement no coach has time to do by hand.
FAQ.
What frame rate should I film shooting at?
60 fps minimum, 120 fps if your phone supports it. At 30 fps the release occupies too few frames to judge timing.
How far should the camera be from the shooter?
Three to five metres, at hip height, with the full body and the ball's early flight in frame. Closer loses the arc; further loses the joint detail.
Should I show players their own film?
Yes, privately and briefly. Two clips — one make, one miss — with one thing to look at beats a ten-minute breakdown. Players remember the last thing they saw, so end on a good rep.
How often is film review worth doing?
Meaningful mechanical change takes two to three weeks, so a re-test cadence of every two weeks matches the biology. More often than weekly and you are measuring noise.
Can I analyse shooting from game film?
Only partially. Game footage is usually too far away and at the wrong angle for mechanics, and shots are contested and varied. Use game film for shot selection and workload; use dedicated clips for mechanics.