Yes, on the blackboard vega hedging is highly questionable. You're hedging a parameter from a model that's wrong. But hedging options with other options helps take the model out of the equation. Say the 110 puts are trading at 20% and the 111 puts are trading at 21.5%. The options are close enough to each other that they are functionally equivalent securities. Invoking the law of one price, they should trade at similar prices. Therefore, in most cases you could buy the 110s, sell the 111s, and delta hedge to expiration. You should earn in the neighborhood of 1.5 vegas ON AVERAGE. It is only riskless arb in a black-scholes world. Jumps and discrete hedging will give your P&L some variance. Still a trade I'd take all day long though.
http://quant.stackexchange.com/questions/10473/delta-neutral-gamma-neutral-positions I never trade options in isolation for the reasons stated by longthewings-the Greeks should move in step predictably, which is the point of BS,but especially with calendar spreads,things can get seriously out of wack. It's good to discuss amnd we all have our own trading experiences- I avergae 20-30 multi leg trades a year and that's enough for my tiny brain.
repost: Yes, on the blackboard vega hedging is highly questionable. You're hedging a parameter from a model that's wrong. But hedging options with other options helps take the model out of the equation. Say the 110 puts are trading at 20% and the 111 puts are trading at 21.5%. The options are close enough to each other that they are functionally equivalent securities. Invoking the law of one price, they should trade at similar prices. Therefore, in most cases you could buy the 110s, sell the 111s, and delta hedge to expiration. You should earn in the neighborhood of 1.5 vegas ON AVERAGE. It is only riskless arb in a black-scholes world. Jumps and discrete hedging will give your P&L some variance. Still a trade I'd take all day long though. <end> what i still dont understand is how do you profit from vega neutral strategies... in this circumstance you are buying the 110 and selling the 111s. to make it Vega neutral you need to buy more of the 110 puts right? so if vols were to increase 1%.. wouldnt your vega from the 110 put exactly cancel off with the 111 puts? any cares? thanks so much
In this instance, assuming you bought/sold 1:1 I would not worry too much about vega. The point is to try and earn the spread between the two options. Assume actual vol will be somewhere between the two -- i.e. 20.75%. It doesn't need to be but for the sake of simplicity assume it is. By selling the 111's and delta hedging at 20.75% vol you are replicating the payoff of a properly priced long 111 put. At expiration, the replicated payoff offsets your short and you are left with 0.75 vegas of profit. Same story for the 110s (replicating a properly priced short payoff). Net result -- 1.5 vegas of profit. To directly answer your question.....there are a couple of ways to realize a profit from vega neutral trading. First of all, remember that vega hedging is needed due to the fact that in practice, volatility is itself a stochastic process. Vol-of-vol creates a whole new risk dimension which must be hedged. One way to make money is the method I outlined before, spreading options against other options and delta hedging to expiration. The crux of it all is that your delta hedges replicate a position in an offsetting option. At expiry, you're left with a stub cashflow (the payoffs all cancel out) which is either your profit or loss. In this case, you were attempting to "arb" IV/RV. The vega hedging was to protect your P&L from shifts in the vol surface. Another method would be to try and trade vega convexity. Here you are trying to "arb" the other risk dimension. To do this, look for an underlying with a vol surface that is trading flat. If you believe vol is truly stochastic for this underlying, then the OTMs should trade at higher IV's (resulting in a convex vol surface). If not, there is arbitrage. You buy the wings, sell atm, with more wings than atm such that you are vega/delta neutral. Rebalance as needed, and any shift in vol (up or down) should give you positive P&L due to vega convexity (volga) --- i.e. equivalent to gamma scalping in the vega dimension. Hope this helps.
I'm not sure how you're arriving at this conclusion. Delta, gamma, vega, they are all equally reliable metrics in what they are meant to measure, and none of them go any further than that. If you're trying to make vega out to be something more than it is, then sure you can get yourself into trouble. But I assure you, vega is every bit as reliable as delta or gamma for a snapshot in time holding all else constant and looking at a 1 vol move. That's all it is and it's perfectly reliable in that sense.
experience mate! Vol can vary hugely from one strike to the next,and vega from one expiry to another can be massively variable-got the T-shirt and I've been trading for 16 years,but hey each to their own
That doesn't answer the question at all. Vega, and what it's measuring is no more or less variable than any of the other greeks when viewed correctly. The problem is you, and what your understanding of vega is. If you're trying to assign more predictive capability to it, and assigning more time to it than was intended, then it's not surprising you find it unreliable. It's not meant to be anything more than it is, a snapshot of a moment in time, all things held constant, measuring the result of a 1 vol move. It doesn't get any simpler than that my friend. It does EXACTLY what it's supposed to. Now if you're saying volatility varies and can change quickly, then yeah no $&#!, but that's not what vega is
I think I get it-but you guys do your own thing and I'll keep pocketing the profits, IO cannot be bothered to dig out the old BS calculations and the serious error in vega. So either you've made millions with your convictions or I'm wrong
Agreed. To each their own...it's what makes a market. Although I will say that I've never encountered a formula quite like Black-Scholes. You're right. It's all wrong. But the best part is that you can plug a bunch of incorrect numbers in it and get the right result -- $!