Launched this week
PredictIQ Pro
Build smarter football apps with AI predictionsa
10 followers
Build smarter football apps with AI predictionsa
10 followers
Production-ready football data and ML predictions API powered by CatBoost, LightGBM, and Poisson goal models.
Stacking CatBoost, LightGBM, and a Poisson baseline is a sharp move, gives you both gradient power and real interpretability for the odds. Love that the API is already production-ready instead of stuck in jupyter hell.
@ekin1366265 Thanks! That was exactly the goal. I wanted the ensemble to balance predictive power with interpretability. CatBoost and LightGBM capture complex patterns, while the Poisson baseline keeps the goal probabilities grounded in football statistics. Everything is already production-ready behind the API, so users can integrate it directly instead of dealing with notebooks or model training.
Easy to drop into a side project, and the Poisson + boosting combo actually beats the naive baselines I tested against.
Soccer predictions that don't feel like guesswork, the Poisson and ML blend actually lines up with the stats I checked from last season. Solid stuff for anyone building a sports app.
Love that you leaned on battle-tested models like CatBoost and Poisson instead of tossing another black-box neural net at football data. The pragmatic stack here feels built for actual deployable accuracy, not hype.