I am trading 231 P Flys into steep skew. I was wondering if I should be using Local vol for my inputs rather than BS. I was also thinking of using the SABR model but do not think this is necessary for a book of vanillas. Thoughts and opinions are appreciated.
Who knows what a 231 (put) fly is? I don't have it plotted, but in running my eyeball down mktIV versus modIV, it's nearly matched within a std.dev. around mkt, for ES/SPX for Wednesday and a scootch to the put side for Friday. As (mostly weeklys) expiries go out in time, that grows to maybe 103% for puts (mkt/model) versus ~97% for calls. So, a slight-if-growing market push to the put side, by the market, versus BSM. Hope that helps.
long 2 85% short 3 90% long 1 100% (for most of my trades). What model are you using? Skew for Local IV is usually 2x BSM.
2-3-1: So, a butterfly with a vertical stuck in front. The IVs are off of IB's TWS, but the model IV has compared well with standard BSM when I have calculated that in the past. In fact, in 'broader' work (going 70 strikes/350 S&P points), I've decided to use the more-consistent model IV rather than the funkier-with-distance-or-shorter-DTE mkt IVs, because while I lose a bit of mkt intelligence in the mid-distance strikes (where I'm more likely to trade ), the overall (mkt-wide) usability is much improved. If you disagree, take it up with IB. I'm good with it.
The -30-delta strike in the SPX acts as the fulcrum and you have stickiness opposing the gain in nominal vols. So many moving parts. Kalman is a pretty elegant solution. The goal is to mine positions that are "skew proof."
Hey tom, I think we might be talking about different things. I am interrested in modeling using local vol and SABR models, not so much towards IB's models.
Modeling the rise to strips (nominal) vs. the flattening of the skew on the downside. Empirically we've had a few instances this year of a flat skew (NOT smile). At minimum you should stress each trade at the wing (at inception) using local vols. You're quoting someone that I have on ignore. I'll have to logout. Edit: that McGinnis idiot. Not surprised that he's completely off topic AF.
Obv the goal is to "isolate" skew or earn from skew while minimizing exposures. One of the best structures is the ATM index 231P with an ATM call vertical (not a fly).