Hi,I've seen quite a few threads on dispersion trading and what came out is that you can buy dispersion / sell correl by setting up several different strategies :- sell index vanilla options and buy options on some components- sell idx straddle/strangle and buy straddles/strangles on components- sell var swap on idx and buy var swaps on components- buy correlation swap I'd like to know form a practical point of view what are the pros&cons of these strategies ie what makes one choose one strategy vs another one.Thks.

No idea ?

i think this is a good question, there must be a correlation trader out there who can give an answer?

[1] Too expensive to execute, delta hedge, roll the strikes etc ....doesn't work unless extracting 0.25 vol points after 3 months of intense trading is your cup of tea.[2] Same as above[3] Works but you have to get the components and sizes correct else you have tracking, net long/short vol exposure (usually rather large) which WILL bite you at some point. [4] Pure correlation trade but hard to MTM so not practical if you have to value your book frequently (like most of us have to), okay if you can throw it in your bag and look at it 3 years later.

Quote[1] Too expensive to execute, delta hedge, roll the strikes etc ....doesn't work unless extracting 0.25 vol points after 3 months of intense trading is your cup of tea.ok i assume we are talking about doing this with either puts or calls. [1] can actually be a little bit less intense since you have 2/3 of the positions on that you would have in the straddles (including your delta hedges) - but on the other hand your delta hedging gets more intense since your starting deltas per component are going to be higher and you are facing different margin issues, that is if you are concerned with efficient use of capital (usually slightly higher believe it or not) - that's essentially the difference to the straddle/strangle methodnow, with respect to both [1] and [2], extracting 0.25 vol points is definitely possible. but it all depends on the opportunity. if you are looking at 10-15 correlation point differentials you are in good shape. also, it's about how your vega/reza ratio looks. usually the higher the correlation the greater your reza. so that your trade becomes more sensitive with respect to correlation than with respect to vol. don't get me wrong though, vol is obviously the key point here. but as long as you get your vega weights right, you should be indifferent if vol moves as long as corr stays the same... anyway, i digress.the problem with [3] is not so much Quote[3] Works but you have to get the components and sizes correct else you have tracking, net long/short vol exposure (usually rather large) which WILL bite you at some point. that's really an issue for [1], [2] AND [3]. getting to the right subbasket is key leave alone rich/cheapness of the options. there are a number of ways to do this, by weight, by correlation or you can solve this numerically (trial and error if you have some time...) the problem with weight and correlation (to the index) is that there is a lot of overfitting - you are not sufficiently exposed to the components which you didn't choose. this particular topic is probably the trickiest of all. there are some funds out there which use some pretty sophisticated numerical techniques to determine a set of optimal subbaskets. again, i digress. the problem with [3] is that you will have to trade single name var swaps which clearly don't have the same liquidity as exchange traded options. dealers won't even make markets in some of them.. so your fitting question becomes even more relevant. true, you save yourself some hassle with respect to delta and also vega hedging but you can really only do this on the major indices containing major names...and [4] is not so much an "arb" play rather than a directional position on implied corr. in addition, the bid offer on these things is pretty atrocious. all the bid offers for the pure dispersion trade are taken into account for this one. so you are definitely paying for getting in and getting out of options. whereas with [1] and [2] you have the option (no pun) to keep until expiration....personally, if you get your sub basket and vega weights right, [2] is the most desirable. granted, it requires a very sophisticated system to keep track of all your greeks, not just the first order but second order as well (particularly vanna) - so make sure you are set up for that ahead of time. Avellaneda has written some interesting papers on this stuff (second order greeks and how your gamma surface looks) - you can find those on the courant websitehope this helps - good luckt

A few questions / remarks on it :- You could add other strategies of Sell / Buy, with : Gamma Swaps, UpVarSwaps, Conditional Swaps,... so as to take into account big moves of the stock price.- Say you do it with VarSwaps, so you only have a Vega exposition. Suppose you weight the Vegas on the components so that you do not have any Vega risk. You're left with Volga and Vanna exposure, and also with an implied correlation. How high would be this implied correlation compared to a correlation swap ? How to explain the difference between both ? How could you explain the Vanna and Volga (or Voma, it depends) in terms of skew, curvature of your smile... Antoine