Freeze both teams at one instant. Where is the best place to pass the ball?

Pass Advisor is a small interactive app that answers this question live in your browser: open the demo — no install, no backend, one HTML file.

What it does

Given a static snapshot of both teams’ positions, the app estimates, for every point on the pitch:

1. Pass success probability P — can the ball get there?
2. Advantage gain ΔV = reward − λ·risk — how worthwhile is it once the pass succeeds?
3. Expected value EV = P × ΔV — the combined recommendation.

The surface is evaluated on an 84×54 grid; a white ★ marks the optimum and an arrow points to it from the passer.

Try it

• Drag any dot to move a player (touch works on mobile); pick from seven formation presets (4-3-3, 4-4-2, 5-3-2, …).
• Double-click a red player to make them the passer, or drag the ball to them.
• Switch between the three surfaces (P / ΔV / EV) and watch the recommended target move.
• Three sliders expose the model: λ (risk weight), interception sensitivity, control softness.

The model in one paragraph

Pass success combines a soft-Voronoi control probability (sigmoid of nearest-teammate vs nearest-defender distance to the target), lane openness (sigmoid of the minimum perpendicular defender distance to the passing lane), and a distance decay. Reward is control × proximity to goal, weighted toward the central channel; risk is a Gaussian density of defenders around the target. The framework follows SoccerMap (Fernández & Bornn 2020), the reward/risk EPV split of Overmeer et al. (2025), and the pitch-control tradition of Spearman (2017, 2018).

⚠️ The coefficients are a hand-tuned analytic approximation for teaching purposes — not fitted to real tracking data. Good for building intuition and playing with formations; serious analysis needs calibration on real data.

Links

• Live demo: https://littlecaps.github.io/pass-advisor/
• Code & full model write-up: https://github.com/LittleCaps/pass-advisor (MIT licensed, single-file HTML — view source is the documentation)

Next steps on the roadmap: a “pass to whom” mode that evaluates only the 11 teammates, player velocity vectors, and calibrating the coefficients on StatsBomb 360 freeze frames.