Analyzing My Basketball Shot with AI: Turning AI’s Strength into Human Strength
Illustration · not a service recordUsing Opus 5.5 for sports science analysis, to turn AI’s strength into human strength.
The stronger an AI model gets, the more the way to feel its strength is to try it on something that concerns you personally, but where AI could only help a little before.
My main impression since Opus 5.5 came out is that anything to do with images and video has leapt forward. Earlier I made animations with it; last night I asked the AI to analyze my recent basketball shooting. It produced a website, which I have put in the comments, and it is really incredibly detailed!
It is quite interesting. For example, it gave very clear feedback on what to focus on in training:
1
Pull the landing spot back
Drill
Tape a box one shoe-length deep in front of your toes. Shoot from 4.5–5.5 meters first; a make counts only if both feet land inside the box. Do five sets of ten shots, then move back to the three-point line under the same rule.
Retest target
Forward drift reduced from 0.7–0.9 meters to within 0.3 meters.
2
Find the shallowest effective squat depth
Drill
Distance ladder. Start at 4 meters with only a quarter squat and a relaxed rhythm; when you make four of five, step back half a meter. The point where your squat starts getting deeper is your current effective range. Twice a week, add squat jumps, pogo hops and medicine-ball chest throws.
Retest target
Hips drop 15–20 cm rather than 30 cm, with the entry angle held at 48°.
3
Ball and hips rise together
Drill
Dip and go. The ball starts at the waist and sinks together with the knee bend, and it must pass the forehead while the legs are still pushing. Start with one-handed shots from 2–3 meters.
Retest target
In 240 fps video, there is no pause at all between the dip and the release.
____________
Good. Now I have clear training goals.
- Original: Ray Lu’s personal Facebook post (published 2026-10-04)
Course titles follow the source. Check each course page for its teaching language.
Introduction to Claude Cowork
Delegate multi-step work with workspaces, Skills, Plugins, and safety reviews. Free course; hands-on use needs paid Claude. The course retains the Cowork name as interfaces merge into Claude.
Applied AI Foundations
Free learning with a ChatGPT account. Turn recurring work into reusable steps with human review; AI Foundations is recommended first. Complete the course and pass its assessment at 80% for a badge.