Simulation / Sports analytics

Live

CricSim

Ball-by-ball T20 cricket simulator with a custom stochastic engine.

CricSim Pick Your XI screen for RCB against GT at M. A. Chidambaram Stadium, with an AI matchup read noting the pitch favours spin and the opponent's likely attack, and a ranked list of players with batting and wicket ratings. — Pick-your-XI screen: the AI matchup read weighs Chepauk's spin-friendly pitch (92/100) against the opponent's likely attack.

[01] problem

Simulate a T20 match one delivery at a time, so that team selection and bowling plans actually change the outcome.

[02] approach

01Each delivery samples from an outcome distribution built from batting power/control, bowler economy and wicket threat, pitch state and pace/spin matchups.
02A heuristic XI selector weighs pitch conditions against the opponent's bowling attack and recommends per-over bowlers and a batting order.
03A FastAPI REST layer serves a React 19 + Vite frontend with Recharts stat radars over a 250-player ratings dataset.

[03] what's interesting

Outcome distributions are computed per delivery, so pitch state and pace/spin matchups feed into every ball.
PostgreSQL (Neon) handles auth, Google Sign-In and resumable tournament saves.

[04] stack

FastAPI
React 19
Vite
PostgreSQL (Neon)
Recharts