A financial time series consists of continuous price movements with intermittent turning points. Traders call these reversals — local highs and lows where the market direction shifts.

This competition challenges you to classify swing points in anonymized stock data. Using both raw numerical features (momentum and ratios) and boolean Signal Descriptors, you’ll build models that map reversal points into one of the four classical swing categories:

H: Local or Global High

L: Local or Global Low

None: Not a swing point

These labels define market structure and are widely used in trend analysis, chart pattern recognition, and algorithmic trading.

Participants will train on a provided dataset (train.csv) and submit models that Kaggle will score against a hidden validation set.