The 90-95% probability definitely helped. And there should be something else about how to close the position or hedge it. I do not believe they are delta-hedged using SP futures during the whole time of their positions since the Sharpe ratio could not be that high. Meanwhile their opportunistic hegding/roll over/buy out were very good, otherwise the flash crash or summer 2011 would show much bigger drawdowns. Speaking of other funds like Chicago based LJM or Tampa based ITB, they behave much like slightly deleveraged version of passively short S&P VIX future indexes with re-balance. To get their performances you can simply use VIX futures with greater liquidity and less transaction costs. Being said that, I am only guessing what they were doing from their "advertising" documents and my limited simple backtests.
I used Barclayhedge free database to get some info. LJM and ITB were learnt from my friends. Not too many funds rough
Interesting discussion. I have watching Global Sigma very closely for a couple of months ago. Found an interview or two out there. There risk adjusted performance is unreal. My understanding is that they are actively hedging their positions even though the options are sold naked. I would also be interested in seeing an equity curve of credit spreads vs. naked options. I trying to more or less replicate the GSS strategy with credit spreads. I've spent a lot of time thinking about how to best structure credit spreads, (in terms of strike price, DTE, etc.). I have not come up with any brilliant ideas yet.
If you simply short naked S&P 500 index put options with a delta of 5-10 for the same time period using a great amount of margin you can get the baseline equity curve which is very close to their performance. I have backtested simple cases of using delta=5, 7.5 and 10 and hold to expiration using optionmetrics bid-ask mid price data fro 2010-2014. They all yield very impressive Sharpe ratios. In the case of the 5 delta there was no losing month in my simple backtest assuming I got everything correct. The call side is miserable although the total return is still positive. I guess you are not "supposed" to make money in selling calls passively in the great bull market Obviously, they are doing more than that especially on picking up maturities, rolling out and closing before expiration, but this simple exercise gives us some hint of what can be achieved in the past QE infinity period. I think it is very related to Karen's strategies. They differ on maturities and how to pick up the far OTM strikes. But they both work on the probability side although I believe GSG is more quant and it have its own probability calculator instead of relying on TOS. They both could be very successful even in the next few years. Whether you want to follow the same path is simply a reflection on your personal risk appetite. The distribution is simply extremely skewed although you cannot directly observe it. If I am a fund manager trying to attract money I may use this approach to get superior Sharpe ratios. But if I am also investing with my own money I will adopt strategies with lower intrinsic risks (with much lower Sharpe ratio though). A simple monthly re-balance scheme of 80-90% in short term Treasury (cash) and the remaining in short volatility ETFs like SVXY (or buy VIX put if you want) could also serve you very well in the past years. And in that case the maximum loss is at most the amount you put into the short volatility ETFs.
Very interesting. Putting aside the bull market, does the volatility skew have something to do w/ calls not being as profitable as puts? Here is the interview I have read with Rao, starts on page 19 or so: http://www.premiacap.com/news/OFI25Jul2013.pdf He is modeling the volatility, more or less the VIX. I have been using this website as a cheap version of a volatility model. I compare their estimations to the VIX. It has been helpful at times, totally off at other times. http://vlab.stern.nyu.edu/en/
I am not sure about how the skew has damaged the call but in my backtested curve the call positions were blown by several market rebound. Meanwhile if you use simple delta hedge you could greatly improve the call position which I believe is coming from the profit of your long SP500 positions. GARCH is one way for the vol calculation which I do not personally like to use in the option trading. I think implied vol is more important than the GARCH forecast although they are highlty correlated. You have VXMT (6 month implied vol), VXV (90 day implied), VIX (30 day implied) and VXST (9 day implied) from CBOE. One approach I thought about but did not implement is to derive a implied vol(T) vs T non-parametric/parametric curve and check the implied vol at any maturity you want to look at. This could give you some hint about the vol to use for 2*SD calculation. You could also compare the actual vol for the options you are interested in versus the one fitted from the curve to see whether they are "cheap" and decide which maturity to use.
I track the VIX, but not the other volatility indicators you mentioned. That is a good idea to compared the derived curve of those indicators versus the actual curve. One drawback to this strategy in general, especially with selling credit spreads, is the ultra-low VIX. I understand how mean reverting it is, but man has that thing dropped like a rock after some recent spikes.
seems weird they have not posted their Oct and Nov performance yet. They were usually posting their performances within the first week. Any idea what was going on?
No. It has to do with things like stochastic vol, discrete hedging, jumps, and demand for OTM puts etc. Lots of literature out there on it. For European index options, calls and puts are equivalent. There is no difference since they should both carry the same amount of time value in the market. There are of course some caveats with DITM options, liquidity, etc.