We develop a novel nonparametric estimator of integrated variance by summing up the squared wick lengths of intraday candlesticks over a fixed time interval. The proposed wick-based estimator is robust to short-lived extreme price movements, such as gradual jumps and flash crashes. We investigate the asymptotic properties of the proposed estimator, and show that its asymptotic variance is about four times smaller than the state-of-the-art differenced-return volatility (DV) estimator. We also develop a Hausman-type test for the presence of both jumps and episodic extreme price movements. Monte Carlo simulations and empirical applications further validate the practical reliability of our proposed estimator.