Wed, Nov 25, 2009 | Jared Woodard
One of the most popular posts I’ve written is “The Bucking Gamma Bull,” in which I said:
Think of your deltas as a mechanical bull, and your gammas as the rate and intensity at which the bull throws you around. The ride starts off quietly, but as time goes on the bull gets increasingly difficult to ride, and eventually you’re likely to be thrown. That’s exactly what happens during an expiration week in which the underlying makes an unexpected move: option shorts can’t control their directional risk as easily as they’d like and have to hedge more deltas to stay on the ride, which only feeds the existing move in the underlying, which – because of all that negative gamma – forces options shorts to hedge even more deltas and so on into a downward spiral of flailing limbs, bruises, and spilled beer.
I still agree with all of the above, and in the iron condors newsletter, we stick pretty closely to a policy of closing out any positions near expiration that have substantial amounts of gamma.
But as member Kyle S. pointed out recently, it’s not a trivial question whether closing trades out early is tantamount to giving up some returns for the sake of a little psychological comfort:
In a previous comment you mentioned that the highest returns come from letting the trades play out to expiration — I ran the numbers and it is definitely true for this year. 75% of trades this year (31 total) would have expired with the maximum gain. 10% would have reached the max loss with the remaining 15% being a mixed bag of small gains/losses and one big loss. April and August had minor losses (2-5% of capital), while September’s loss was sizable (-16%). November’s max loss trade 4 wiped out most of the max gains from trades 1-3. [...]
Obviously there isn’t enough data to really know the odds. 75% is higher odds than one should expect for max gain trades and 10% for max loss trades seems low to me. Wonder if the favorable odds (max gain vs. max loss) are due to the higher than normal gap between implied and realized volatility this year.
Have you done the analysis for the entire newsletter time frame?
I did look at the whole newsletter history awhile back, and I did some similar analysis when I was developing the managed accounts strategy. The results were pretty consistent – holding trades through expiration results in higher returns with higher standard deviations, as one might expect. Winning trades were slightly more profitable, a few more trades ended up as maximal losers, some trades that would have otherwise been closed out as losers ended up expiring worthless (i.e. profitably), and vice versa. That’s all completely in line with expectations; the important question is whether the increase in absolute returns is worth the added risk.
The net gain in performance on a risk-adjusted basis is large enough in my estimation that we typically hold positions through expiration in the managed accounts. So why the difference vs. the newsletter approach? The simple fact is that, qua portfolio manager, I’m comfortable dealing with those risks and experiencing the occasional expiration-week drawdown (which happened twice this fall, for example), whereas novice option traders typically are not. I should also mention that the managed accounts are usually delta-hedged, such that expiration week losses aren’t as large as they would otherwise be, and delta hedging isn’t logistically possible in the newsletter or within the reach of a novice trader.
So, in line with the earlier post, I unequivocally suggest that option traders who aren’t very experienced and/or are more risk averse close positions early. If nothing else, the psychological benefit of avoiding those occasional expiration week losses will help traders stick with the strategy where they might not otherwise. After all, there isn’t much benefit to trading a mechanical or quantitatively-oriented strategy if the strategy is at odds with your risk tolerance and psychological profile – you’ll end up abandoning it anyway or, worse, taking mechanical signals on an inconsistent, discretionary basis.
Photo courtesy of Flickr user alamosbasement.