The problem with short gamma

[nitro]

Quote from dont:

Just my two cents, the problem is even worse, you have to treat the portfolio as one big option, pricing each Call and Put and then adding up the values, does not work. The equations are non-linear, so the solutions do not add....

Right, that is why you want something like a VS. The whole is almost equal to the sum of the parts in VSs, so aggregrating accurate book risk is possible. This is attained by the inverse of strikes-squared weighting scheme. If strikes were continuous instead of discreet and you could put on an infinity of option positions on at these strikes, you would never need to hedge the underlying at all. Alas, that is one of the big problems with BS as it assumes that you can continually hedge, at zero cost.

...Basically have to value and hence derive the greeks for the portfolio as a whole. Then the vol you use has to be your best guess of what its going to be over the remaining life but adjusted for the way you plan to hedge.

This does not go far enough. You also need to understand the options own sensitivity to the skew, which is also non-linear but can be approximated roughly so for small underlying moves and time. In the presence of jumps, the whole thing becomes incredibly complex.

 

[RedEyeFly]

If there were continuos strikes and perfect dynamic hedging, we would either all be out of business, or trading long gamma.

 

[MasterAtWork]

Quote from RedEyeFly:

If there were continuos strikes and perfect dynamic hedging, we would either all be out of business, or trading long gamma.

+1, there would be no need of options .

 

[nitro]

Quote from RedEyeFly:

If there were continuos strikes and perfect dynamic hedging, we would either all be out of business, or trading long gamma.

Well, that is what a Variance Swap gives you, with the "dynamic heding" priced in. With a VS, you are no longer trading the underlying (mean), but something far closer to pure vola.

There are still risks, but the mean is almost completely removed from PnL for a long stretch of strikes, as seen by the almost flat gamma profile.

 

[Rodney King]

When you cut through all the posturing and buzzwords, OP's real issue is: he wants to sell premium but is annoyed the markets have sustained directional moves, thus he gets hurt more than he "should" based on stdev or some other simple measure of vola. But.. the market's <i>job</i> is to inflict pain on whatever strategies are popular. The fact that iv has drifted down for the past year means there are more premium sellers than buyers. All those sellers are at risk from sustained directional moves. So what's the market going to tend to do? An exercise for the reader.

 

[nitro]

Quote from Rodney King:
When you cut through all the posturing and buzzwords, OP's real issue is: he wants to sell premium

That is false.

but is annoyed the markets have sustained directional moves...,

That is true.

thus he gets hurt more than he "should" based on stdev or some other simple measure of vola.

That is false

But.. the market's <i>job</i> is to inflict pain on whatever strategies are popular.

Irrelevant in my case.

The fact that iv has drifted down for the past year means there are more premium sellers than buyers. All those sellers are at risk from sustained directional moves. So what's the market going to tend to do? An exercise for the reader.

In fact, the popular strategy has done extremely well. Buy stocks, sell calls against it. Not as well as, throw a dart a the WSJ stock lists page and just buy stocks, but then, that is only in retrospect.

http://finance.yahoo.com/q/ta?s=^BXM&t=1y

If you don't want to deceive yourself, your true measure of continued success is to compare your gains against some index that mirrors the way you trade, and then see if you have alpha above and beyond that index. If someone says to me, I am up 70% from last march, I would say, oh, you bought the SPY last march and shut down your computer, aye?

 

[Rodney King]

Quote from nitro:In fact, the popular strategy has done extremely well. Buy stocks, sell calls against it. Not as well as, throw a dart a the WSJ stock lists page and just buy stocks, but then, that is only in retrospect.

http://finance.yahoo.com/q/ta?s=^BXM&t=1y

If you don't want to deceive yourself, your true measure of continued success is to compare your gains against some index that mirrors the way you trade, and then see if you have alpha above and beyond that index. If someone says to me, I am up 70% from last march, I would say, oh, you bought the SPY last march and shut down your computer, aye? [/B]

Re buywriting, it's ~putselling (yes, there are second-order differences, don't bother) so of course it does well when the market drifts upward. Buywriters are hurt by market crashes.

As to "up 70%" & etc., I agree with OP that performance against a benchmark is the correct measure -- for a mandate-constrained trader. It's less clear for someone trading his own account.

 

[nitro]

My intuition on this has become stronger: If you don't have a correct model of the underlying, you will hedge wrong. It is worth tying the two threads that I was thinking about at the time here:

http://www.elitetrader.com/vb/showt...669&perpage=6&highlight=variance&pagenumber=1

The other is the link on the very first post of the very first page.

 

[RedEyeFly]

Quote from nitro:

Well, that is what a Variance Swap gives you, with the "dynamic heding" priced in. With a VS, you are no longer trading the underlying (mean), but something far closer to pure vola.

There are still risks, but the mean is almost completely removed from PnL for a long stretch of strikes, as seen by the almost flat gamma profile.

Nitro,

Have you ever done any cointegration or RV testing for VS trades?

 

[nitro]

Quote from RedEyeFly:

Nitro,

Have you ever done any cointegration or RV testing for VS trades?

No. The reason is that I am data starved. But I think the new systems are looking at porfolio level and book level PCA and Cointegration/RV analysis.