Chasing unicorns?

[atticus]

Dispersion execution is complex. You can trade dispersion in skew (down and out) or a straddle-model (atm). Going into detail would involve a really large thread and someone has hired me to implement, so I can't go into a lot of detail:

D&O is simply a short index put // long street-vol put (long) dispersion model. You're long vol-basis typically in something approaching 20-deltas on both the index and component puts. Working say 40-vol on the index short put and 25-vol in the constituent puts. The edge in skew is put to work in something approaching notional-equivalence.

ATM straddle dispersion may be done at an edge-loss, but you're long street-vol in atm straddles // short index straddles. You're looking for the equivalent of destructive interference (EE) in that you will have a correlated basket (index) that has poor residual corr (constituents) and have opposing outliers -- GE up 1% and AA down 1% , with no net-impact on the index.

That's really general and rough, sorry.

 

[sle]

I'll chime in tomorrow - just came back walking my dogs and have to go to bed. As usual, there are many ways to do this and to analyze it, but it certainly works once you get it going.

 

[rew]

What is street volatility?

 

[jtrader33]

Quote from atticus:


GE up 1% and AA down 1% , with no net-impact on the index.

Do you recommend rebalancing deltas on the single names and/or index throughout the life of the trade? Also, if implied corr comes in to a level where you think it's fairly priced, does it make sense to take off the trade or just leave it on until expiration to avoid crossing bid/ask again?

Understand if you can't answer, but figured i'd ask since these have more to do with trade management rather than edge/timing.

 

[spindr0]

Quote from sle:

In general, you don't want to buy options for all of the index components. What you do is you select a subset, either by weight or by "the driving force" and only buy vol on that subset.

My preference would be to slowly leg into it over a few days, doing chunks of index against each of the components you decided to buy. Institutional clients (the f*ckers I deal with) like to come and trade it as a package in comp but i doubt it is possible for small size.

OK, still mulling some possibilities - been thinking about this in terms of the OIH which only has 13 components and plenty of volatility lately.

A few questions...

Vis a vis "buy vol on that subset", by that I take it you mean the potential movement of the active UL's rather than the actual vol of the options?

If the premise holds water, why not do straddles (or strangles) in the same manner since once has no idea what direction the UL vol will be?

If one got out ahead on one or more legs, delta hedge with some UL. Perhaps the one(s) that got out ahead instead of the index? Make any sense???

Edit: Oops, seems the previous poster had the same idea...