the theory is easy, but how is it done in practice? i ask because from everything i read about it you buy an option with a low vol, and expect to capture profit by trading the delta around the option position. but how often do you trade? do you set a vol level that you trade at? ex: if you bought at 20 vol, do you calc levels to trade based on 25 vol? i'm in options marketing, but the traders next to me are vol traders. they buy cheap vol, sell rich vol, and then leave the delta hedging to another guy whose job it is to delta hedge everything. how does he know when to trade? and what about when you have more than 1 option position in a name: options far up the skew for example? how do you average out the target vol? thanks in advance for help. if there's anything else i should i know pls post it.

Theoritically, you use 'real vol', adjusted for *many* things. However, since the delta is computed from black scholes, this is much more art than science. Black scholes is theory far from practice. Among factors to consider: - Transaction costs, and variance of P/L you can support. This is an important cost/risk utility problem; - Vol of vol, mean reversion of vol, and empirical vol bounds. If you can support some P/L variance, you can use a moving average on the vol to reduce your rebalancing; - Frequency of vol. Wont talk much about this, find for yourself; To my belief, lowest-variance rebalancing is based on delta bounds, not time intervals. Discretization of black scholes PDE could prove this. There are many papers on the subject. Google them out. Good luck! cosine

Actually, thinking of it, it's quite funny that the vol traders at your place dont do their own hedge. Seems like inefficient separation of tasks... cosine

Most of the dealers do it this way - eliminates the redundancies in hedges and decrease transaction costs. For example, if one guy is short delta and another is long delta on the same bucket, it would be netted out.

How does having a single person doing this prevents further transaction costs? It's the same desk, two traders can trade one with the other. It's a bank after all. I wouldn't want my gamma rebalancing P/L to be in the hands of someone else and see my bonus vaporized if implied vol gets desynched with realized vol. Also removes a whole lot of possible strategies in hand, both for the vol trader and the hedger.... Just a thought, I dont know much about dealer environment... cosine

Actually, you don't have to worry about your bonus. The hedger is hedging the overall firm's exposure, it has nothing to do with your exact position.

we have a 2 person team, with a 3rd guy hedging the vol of all the options. i can see how you'd be worried he doesn't do a good job. from what i gather, it's not as easy as theory to delta hedge. let's say the trader thinks 20 vol is low, he isn't just going to hedge at 23 if that's what he thinks it should be; he will watch the mkt and take delta risk at certain points. regarding the idea of hedging to realized vol; let's say there is 6mos left until expiration, would you hedge on this day at the past 6mos RV? and then tomorrow on the 6mos-1d RV, etc...? i'm trying to grasp 'volatility trading'. vol can be cheap or rich like anything else, but from what i understand from all the papers i read is: vol is captured thru delta hedging. if this is the case then for the guys here who are 'vol traders', delta hedging would be important and cosine would be correct: why leave it in the hands of someone else? just trying to ties this all together...thanks for the good replies.

Quote from straddle_me: from what i gather, it's not as easy as theory to delta hedge. let's say the trader thinks 20 vol is low, he isn't just going to hedge at 23 if that's what he thinks it should be; he will watch the mkt and take delta risk at certain points. That's pretty much it. regarding the idea of hedging to realized vol; let's say there is 6mos left until expiration, would you hedge on this day at the past 6mos RV? and then tomorrow on the 6mos-1d RV, etc...? 6 months in the past and 6 months in the future are different things. The way you estimate future volatility for hedging is much more related to how you expect volatility to move in the future. Statistically, you can use a garch model. Practically, traders should know vol clustering, boundaries, and how they expect it to evolve for the next 6 months. I said one who buys vol should rebalance, because the trade was made on his insight of the vol to be seen in the future, so he should also be the best placed to hedge the option. i'm trying to grasp 'volatility trading'. vol can be cheap or rich like anything else, but from what i understand from all the papers i read is: vol is captured thru delta hedging. if this is the case then for the guys here who are 'vol traders', delta hedging would be important and cosine would be correct: why leave it in the hands of someone else? Actually I also hear quite often the vol capture through delta hedging thing. This is not totally true. Vol is bought when you buy the option. Theoritically, frequency of delta hedge affects your expected P/L only through transaction costs, and its variance. Practically, high frequency rebalancing can be expected to be more profitable, but for reasons not related to gamma.