S-MAS Sprint

The set-up decides whether the footage can be scored at all. Set it once, measure it, write the dimensions down and repeat it exactly at every retest — only then are you comparing the athlete rather than the camera position.

• Level: floor and wall lines in the frame are horizontal (turn on the grid and level in the camera app).
• The camera stays still during the whole trial — no panning, no zooming. The app does the zooming.
• The whole body stays in frame at every step, from the top of the head to the feet.

Do a test pass at a jog and watch it straight away — 30 seconds that can save the whole session.

Sprint mechanics depend on the warm-up, fatigue and motivation. A warm-up that is always the same matters more than a "perfect" one.

• Analyse the trial with the best flying 10 m time or — without timing — a typical trial with no stumble. Pick the rule once and stick to it.
• You score the leg closer to the camera. Score the other side from trials run in the opposite direction (the camera stays put).
• After each trial note the file name, the time and the athlete's comments (pain, RPE).

Open sprint.oleksysport.com in the browser on a phone or computer. The analysis view is a six-step wizard — you cannot move on until the previous step is complete. Videos and scores stay on the device.

• Athlete: "Athletes" tab → "+ Add". First name and initial, birth year, position and height. Always enter the height — it calibrates the image without extra measurements.
• Step 1 · Video: load the file from the gallery. The "Frame rate" field must show the RECORDING frame rate (120 or 240), not the playback rate. Tick the protocol conditions and set the running direction.
• Step 2 · Athlete and scale: choose the athlete and the calibration — "Athlete height" or "Two points" (tap both cones and enter 2.00 m). Without a scale you still get times and the score, but not step length or speed.
• Step 3 · AI scan: set the A–B range to one cycle (from toe-off to the next ground contact of the same foot), then "AI: range → phases". With a busy background use "Athlete box".

AI suggests — it does not score. The scan proposes five phase markers and computes contact, flight and step times. You confirm every marker by hand and score every item yourself.

S-MAS does not score "the whole run" but five specific frames of one cycle. The most common source of disagreement between raters is not a different opinion about mechanics — it is different frames.

• Tap a marker, check frames −2…+2 with the −1/+1 buttons and confirm the right one.
• Watch the whole cycle at 0.1× or 0.25× — the order must be CTO → MVP → LSW → TD → MST.
• If no frame fits perfectly, pick the one closest to the definition and add a note.

Score each item in the frame of its phase, not from your impression of the whole run: 0 — sound mechanics, 1 — the feature is present. The total runs from 0 to 12 — lower is better.

The measurement error of the scale is about 1 point. A one-point difference between assessments sits within the noise — a change of 2 points or more is the signal to act on.

• Score "blind", without looking at the athlete's previous result.
• Now and then re-score the same video after 48 h or hand it to a colleague. A difference above 1 point usually means different phase frames.

The assessment ends with a report and a plan. The athlete and the coach need a picture, one number with its interpretation and a short list of what to do next.

• Put at most 1–3 priorities in the report, most important first — not every 1 at once. Start with items linked to ground contact and load acceptance (7, 9, 10, 12), then swing (2, 4, 5) and trunk (3, 6).
• Retest after 4–6 weeks of corrective work, under identical conditions: venue, camera, run-up, warm-up, footwear, trial-selection rule.