About this question
Medium · time_series · Quant Researcher interview question · realized_variance, bipower_variation, jump_detection, high_frequency
Realized bipower variation (BPV) separates the continuous diffusive component of quadratic variation from price jumps in high-frequency return data. It is a critical input to jump-robust volatility forecasting models used across execution desks and statistical arbitrage pipelines. Task Implement the function realized_bipower_variation(returns: list) -> list that computes three key volatility statistics from a series of n high-frequency log returns, r_t. The function should calculate the followi