Ah, I see where you're coming from. Yes, you'd be net short Gamma across 2 different markets. Expensive gamma sold (stocks) against cheap gamma bought (index). The 2 stock example is an expiry play, where no intrinsic loss is possible. At expiry you keep the difference between IV sold and bought for a net credit.
Isn't the whole objective of a dispersion trade to capture the difference between component IV and index IV ? Where the index is 100% replicated with correctly weighted stocks you'd have a true arb ? Where 100% index replication isn't done you have correlation / dispersion exposure ?
Yes that is the objective but to do so I would think you need to isolate the volatility exposure. Continuing with the 2 stock example. Lets say you take the position for a $0.4 credit (that is what I get with 3 months to expiry) Then assume we just hold it to expiry and that stock 1 expires at 55 and stock 2 at 45 , leaving the index unchanged at 100. That is then a $5 loss. How do we avoid that one?
you cannot to avoid it , see my previous posts on this subjuct.The only way to collect IV "arb" if ALL basket stocks will go to the SAME direction.
Yes you are right and I was talking b*llocks The risk of course is that the correlation (+1) won't hold. This type of strategy is new to me, please bear with me on the learning curve....
Not sure if this would be helpfull to you guys and I forgot who gave it to me, but in the spirit of sharing here is some interesting stuff on dispersion: -Neo
Dantes, dispersion is not usually done for 3 months. It is done for expiration cycles or 1 month. The probability of a $50 stock vol 10% moving 5 points in 3 months is less than 3%. The probability for 1 month is .1%. So multiply .001 x .001 (for 2 stocks) and multiply that result by the probability of their correlation going from +1 to -1 and that will give you the probability of your scenario coming to pass.
Thanks , Neo. One thing come to mind : why trade when I can write a book about it... I wonder if the author ever used this staff by trading his own money. Good luck to anyone(especially retail) who is looking for black box trading on this strategy...
Thanks Neo. If only we had a mathematician here to decipher the code. The 2 scenarios from this article (attached) are much more realistic than the simple example discussed thus far. The problem with these specific scenarios, however, is that the results do not match the graphs. For Scenario 1 I have a stock P/L of -6.90 and an index (assume equal weighting) of -.46. For Scenario 2 I have a stock P/L of 4.25 and an index of .28. Somebody screwed up here, unless they are not telling us how they get those numbers. As it stands, both of these scenarios look profitable to me if the stocks were straddled.
I'll take your word for it mysticman. I haven't really looked at this presentation in detail, its been sitting on my harddrive for a while. I'll try to see if I can get some input from the guys on the trading desk or a quant when they are not too busy... At this point in time, dispersion is a bit out of my league, but I thought it might be usefull to others. -Neo