Gamma Scalping (Long Options, Delta Hedged)

Overview

Almost every strategy on this list is short volatility. This one is the other side, and it is worth understanding for exactly that reason.

You buy a straddle: a call and a put at the same strike. That position makes money if the underlying moves a lot, in either direction. But you do not want to bet on direction, so you hedge out the delta by trading the underlying. Now you are flat directionally and long the movement itself.

Here is where it gets interesting. Because you own options, your position has gamma, which means your delta changes as the market moves. As the market goes up, you get long. As it goes down, you get short. Every time you re-hedge back to flat, you are mechanically forced to sell after a rise and buy after a fall. You are systematically buying low and selling high, and the wiggle of the market pays you for it.

That is gamma scalping. The market's noise is your revenue.

The catch, stated plainly

You pay for this privilege every single day. Options decay. That decay is theta, and it arrives whether the market moves or not.

So the whole trade is a race:

• Gamma earnings depend on how much the market actually moves. More movement, more scalping profit.
• Theta cost depends on what you paid for the options, which is set by implied volatility.

You win if realized volatility comes in above the implied volatility you paid. You lose if it does not. Everything else is detail.

This makes the trade conceptually the mirror image of variance swap carry. And because the variance risk premium is usually positive, meaning implied is usually above realized, long gamma is usually a losing trade. You are fighting a structural headwind. You need a genuine reason to think implied is too cheap right now, not just a hope.

Strategy logic

• Signal: Forecast realized volatility over the option's life. Compare it to at-the-money implied. Only enter when your forecast is meaningfully above implied, with enough of a buffer to cover hedging costs.
• Structure: Buy an at-the-money straddle. At the money maximises gamma per unit of premium, which is what you want.
• Hedge: Re-hedge the delta on a defined rule. Every close, or whenever delta drifts outside a band, or whenever the underlying moves a set amount.
• Exit: Close when implied vol rises to your forecast, when the theta budget is spent, or at expiry.

Parameters (knobs)

• Hedging frequency: This is the parameter that decides your fate, and it is far more subtle than it looks. Hedge too often and transaction costs eat you alive. Hedge too rarely and you leave gamma profits on the table and take on real directional risk between hedges. There is no universally right answer, and the "optimal" frequency found in a backtest is one of the easiest things in all of quant finance to overfit.
• Tenor: Short-dated options have enormous gamma, which is great, but their theta is brutal. Longer-dated options bleed slowly but scalp slowly too.
• Entry threshold: How far below your forecast implied has to be before you commit. Given the structural headwind, be demanding.
• Volatility forecast model: A simple exponentially weighted moving average of squared returns captures the key insight that volatility clusters. Fancier models add little for this purpose.

Where it works and where it hurts

It works when volatility is about to expand and the option market has not priced it yet. Practically, that means the periods just before a regime shift: a complacent market with implied vol scraped down to multi-year lows, right before something breaks. The classic winners were long gamma going into late 2007 and going into February 2020.

It hurts, relentlessly, in calm markets. Implied vol is already above realized, you pay theta daily, the market does not move enough to scalp your way out, and you bleed. And crucially: the bleed is boring and the payoff is rare. Most people who run long gamma quit after the fifth consecutive losing month, which is usually one month before the move.

Backtest design checklist

• The hedging cost is the whole game. If you backtest with zero transaction costs, long gamma looks fine. Put realistic costs on every re-hedge and a large part of the edge evaporates. This is not a minor adjustment.
• Use real option prices. You need the actual implied vol you would have paid, including the spread you cross to get in.
• Attribute the P&L. Split it into gamma gains, theta paid, vega mark to market, and hedging slippage. If you cannot produce that breakdown you do not understand your own results.
• Test hedge frequency across a wide grid and look at whether the results are smooth or whether one frequency happens to win. If only one wins, you have found noise.
• Do not let the forecast peek. Your realized volatility forecast can only use data available at the time. Volatility forecasting is where lookahead bias creeps in most easily.

Common failure modes

• Buying cheap vol because it is cheap. Low implied volatility is often correct. Markets are genuinely calm sometimes. Cheapness alone is not a signal.
• Over-hedging. Chasing every small move, paying the spread each time, and turning a positive-expectancy position into a slow loss through costs.
• Under-hedging. Letting delta run because the market is trending in your favour. You have now abandoned the strategy and are making a directional bet.
• Quitting at the wrong time. The payoff distribution is heavily skewed toward a few large winners. Cutting the strategy during the inevitable dry spell means you keep the losses and miss the gains.

Our notes and suggestions

We would treat gamma scalping less as a standalone money maker and more as an education and as a hedge. Running it teaches you, in a way that no amount of reading can, exactly what the short-volatility sellers are being paid to bear. And a modest long gamma position is a natural offset if the rest of your book is short volatility, which for most systematic traders it quietly is.

Be realistic about the base rate. Implied is above realized most of the time, which means this trade is swimming against the current most of the time. Do not run it always on. Run it when implied vol is genuinely, unusually cheap, when you have a defensible reason to think movement is coming, and when you have budgeted for how much you can afford to bleed before you are wrong.

What would change our mind: a persistent regime where realized volatility exceeds implied, which would flip the structural headwind into a tailwind and make this the strategy to run rather than the one to respect.