Realized Volatility

Computing the Actual Move

Realized volatility is backward-looking: take a series of past returns — typically daily, over a trailing window such as 20 or 60 trading days — and annualize their standard deviation. It answers a simple, factual question: how much did this security actually move, on average, over the period measured? Unlike implied volatility, which is extracted from current option prices and reflects what the market expects to happen going forward, realized volatility is pure historical fact, computed after the fact from a price series that has already printed. It carries no forecast and no opinion embedded in it — it is simply a measurement of what already happened, which is exactly why it is the natural benchmark implied volatility gets compared against.

The IV-RV Spread

Implied volatility tends to run above subsequent realized volatility, on average and over long samples — the market generally prices in more movement than actually shows up once the period has played out. That persistent gap is called the volatility risk premium, and it is the structural reason option-selling strategies carry a positive expected value over long samples: the seller collects a premium priced for more movement than typically arrives, pocketing the difference between the fear priced in and the calm that usually follows. The spread is not constant, though — it compresses in calm, low-volatility regimes and can invert sharply around known event risk, when realized volatility briefly spikes above what the options market had priced in ahead of the event.

Why It Matters for Sizing

Comparing IV to RV, name by name, is the starting input for deciding whether a premium-selling structure is actually being paid a fair price for the risk it is taking on. A wide IV-RV spread on a given security suggests the options market may be overpricing near-term movement relative to how the stock has actually been behaving recently; a narrow or negative spread suggests the premium on offer may not adequately compensate for the risk being carried into the position. Closelook's VIX level and single-name skew readings are both downstream of this same realized-versus-implied comparison, applied at the index and single-stock level respectively, and both move alongside each other more often than not.