Implied Volatility vs. Real Volatility

[comintel]

Quote from sle:

Obviously, it does not work all the time, but I have multiple strategies running at the same time, so I have the benefit of diversification. In anything, in a crisis I usually make money - 2011 was my best year, followed closely by 2008.

I can't get into specifics, but the whole idea is to identify rich risk premia and hedge them with cheaper risk premia. In some cases it is a different type of risk premium on the same product - e.g. I could be short gamma, but hold long-dated skew against it. In some cases it's a correlated product - I might be long variance on Russell and short variance on S&P500. In some cases, it's a farther removed risk factor - e.g. I could be short gap, but long vol-of-vol. And finally, in some cases, I am willing to take idiosyncratic risk in diversified form and hedge away systematic risk.

That is a very appealing approach.

I assume that one would have to do extensive correlation studies as background.

One of the harder aspects in hedging one option position by another, for example hedging gamma by using skew, must be understanding the range of possibilities as to how these could both evolve. (For example, could both move against you?)

Is it necessary to do extensive programming of historical studies of past option behavior to get there? Or can one get there with just intuitive understanding and experimentation. (If you can comment).

 

[cdcaveman]

Quote from optionbull:

Out of curiousity- what does the risk reward look like on these, I.E lose 1 to make 2 or lose 1 o make 10 .. 20 , can you control your risk absolute ?

Thnx

depends on your set up... your compounding risk factors with correlation... obviously there is risk of a decorrelation ... being short vol in a parent and long vol in a constituent i would think offsents some of the risk of being just short in the index right... but if nothing is realized in the long vol constituent your short index will make up for the loss a little.. your really trading the differences... this is just something i'm learning..
if the event relates to the member .. like a big earnings event.. then the index is going to shadow the constiutient.. depending on its weight.. aapl is obviously a big weight in the nasdaq...

but in the reverse situation... if there is increasing risk in the banking sector.. you could sit long vol in that and short vol in some area that typically has rich premium.. like tech.. i'm no quant.. but i'm sure building a correlation matrix and looking at certain market in particular .. like time frames with increasing vol.. flat markets.. etc.. would help.. i'm not there yet.. i'm not a stones throw from there to put it lightly..

the idea would be to make sure that in a crash you end up being long vol at some degree...

and that you have tested coreelation in instruments in a crash.. TLT against SPY.. being short and long helps in a crash because everything in equities correlates. hedging tail risk is something i think people only do after they have busted in one.. its hard enough to crank a dollar out of the market on the regular..

i kind of picture it like buidling something stable to hang on to when the shit goes down.. or better yet.. something convex that will benefit greatly when the shit goes down..

 

[cdcaveman]

Quote from comintel:

That is a very appealing approach.

I assume that one would have to do extensive correlation studies as background.

One of the harder aspects in hedging one option position by another, for example hedging gamma by using skew, must be understanding the range of possibilities as to how these could both evolve. (For example, could both move against you?)

Is it necessary to do extensive programming of historical studies of past option behavior to get there? Or can one get there with just intuitive understanding and experimentation. (If you can comment).

simply put.. and i'm not expert of course... if you have a known event. with a large percpetion of risk on the horizon.. buy it .. sell in front of it.. hedge the underlying...

the rich premium is the front month that doesn't have the event in it.. the cheap is the premium that has the jump priced in... its only cheap and expensive relative to eachother.. not outside of the relationship... i try not to get lost in models....

 

[comintel]

Quote from CT10Gov:

Oh boy this brings back memories... I was in the room when he handed these out when they were fresh.

The math. of volatility surfaces provided by Jim Gatheral is pretty complicated. Do you think that it is necessary to work through it carefully to understand it completely or do most people (students or practitioners) just go for the general idea?

 

[optionbull]

Quote from cdcaveman:

depends on your set up... your compounding risk factors with correlation... obviously there is risk of a decorrelation ... being short vol in a parent and long vol in a constituent i would think offsents some of the risk of being just short in the index right... but if nothing is realized in the long vol constituent your short index will make up for the loss a little.. your really trading the differences... this is just something i'm learning..
if the event relates to the member .. like a big earnings event.. then the index is going to shadow the constiutient.. depending on its weight.. aapl is obviously a big weight in the nasdaq...

but in the reverse situation... if there is increasing risk in the banking sector.. you could sit long vol in that and short vol in some area that typically has rich premium.. like tech.. i'm no quant.. but i'm sure building a correlation matrix and looking at certain market in particular .. like time frames with increasing vol.. flat markets.. etc.. would help.. i'm not there yet.. i'm not a stones throw from there to put it lightly..





the idea would be to make sure that in a crash you end up being long vol at some degree...

and that you have tested coreelation in instruments in a crash.. TLT against SPY.. being short and long helps in a crash because everything in equities correlates. hedging tail risk is something i think people only do after they have busted in one.. its hard enough to crank a dollar out of the market on the regular..

i kind of picture it like buidling something stable to hang on to when the shit goes down.. or better yet.. something convex that will benefit greatly when the shit goes down..

Thanks

Wondering if these guys returns are 5 % year or 50 ?

Sounds like just exotic spreads ..... intra or inter commodity, in this case index's ...

 

[quatron]

Quote from jtrader33:

Somewhat OT, but has anyone done extensive backtesting on options quote data and then subsequently live traded the strategy to come up with a reasonable heuristic for estimating fills from historical quotes? I'm guessing paying the full NBBO spread is too pessimistic...maybe 1/2 the spread?

I've done that. My multi-leg execution strategy does not pay the spread, it trades on theoretical (fair) values on average. It quotes the legs and trades a spread maintaining vega-neutral exposure while executing.
My backtester simulates execution of quotes/limit orders very accurately, according to the price-time priority in the book. Live results are almost the same as backtested.

 

[Trader13]

Quote from sle:

... I can't get into specifics, but the whole idea is to identify rich risk premia and hedge them with cheaper risk premia. ... In some cases, I am willing to take idiosyncratic risk in diversified form and hedge away systematic risk.

Sounds like a quasi-dispersion strategy. Are institutional desks and funds still active in dispersion trading, or has it lost popularity as "yesterday's" strategy?

 

[justrading]

Quote from comintel:

The math. of volatility surfaces provided by Jim Gatheral is pretty complicated. Do you think that it is necessary to work through it carefully to understand it completely or do most people (students or practitioners) just go for the general idea?

I cannot answer as an expert on options, which I very obviously am not, but let me offer a pragmatic view.

It is not necessary to understand the physics of the internal combustion engine in order to benefit from having and driving a car. Energy conversion, torque, drive-train efficiency are nice to know if you are a mechanical engineer, but for a driver knowing that the car needs a particular grade of fuel etc is sufficient.

It would be wonderful if authors like Gatheral would write the book from the perspective of 'this is what we found, these are the implications, this is how you could use it', and stick all the mathematical models in the appendices.

Unless of course the story is in the models, ie. more like a research paper.

 

[cdcaveman]

Quote from justrading:

I cannot answer as an expert on options, which I very obviously am not, but let me offer a pragmatic view.

It is not necessary to understand the physics of the internal combustion engine in order to benefit from having and driving a car. Energy conversion, torque, drive-train efficiency are nice to know if you are a mechanical engineer, but for a driver knowing that the car needs a particular grade of fuel etc is sufficient.

It would be wonderful if authors like Gatheral would write the book from the perspective of 'this is what we found, these are the implications, this is how you could use it', and stick all the mathematical models in the appendices.

Unless of course the story is in the models, ie. more like a research paper.

There is no clear and easy path to learning the stuff either... I just keep trying to consume stuff way over my head.. and as I do I end up constantly going to research what I'm reading... sometimes I might spends week reading about something written in the first few moments of these books/papers.. I enjoy the journey... its an interesting puzzle for sure.. the main problem is the guys quickly use vocab that is familiar to them not me... its not frank... its not broken down ... and of course if you have been immersed in it for years why its simple! They lose touch with what it was like to have never seen the material...

Putting it to use is the art of the craft...

 

[justrading]

Quote from cdcaveman:

There is no clear and easy path to learning the stuff either... I just keep trying to consume stuff way over my head.. and as I do I end up constantly going to research what I'm reading... sometimes I might spends week reading about something written in the first few moments of these books/papers.. I enjoy the journey... its an interesting puzzle for sure.. the main problem is the guys quickly use vocab that is familiar to them not me... its not frank... its not broken down ... and of course if you have been immersed in it for years why its simple! They lose touch with what it was like to have never seen the material...

Putting it to use is the art of the craft...

You pretty much hit the nail on the head about the jargon filled journey that learning options is all about.

A simple example - short gamma. Basically you sell options you are short gamma, but folks don't say that, they say short gamma so a newbie has to scramble to find out what short gamma means. Let's not forget the wonderful 'I trade convexity' either.

I am a big fan of simplicity, so I always look for a plain English expression of all the crap I come across. Most of it is unnecessary, probably cooked up by various authors trying to make it seem like they have developed a new concept, then picked up by readers who think that it sounds cool or awesome or whatever.

As for mathematical models, they work until they don't work anymore, as in LTCM and the sub-prime crisis, so I prefer to spend my time working on the practical aspects of the market. The BS and binomial models underpin our trades, but both have so many (necessary) assumptions that they deviate from the reality of the markets, leaving us with a something is better than nothing position.

It is all very wonderful, but the only thing that matters is when you use it, do you make money? The guy who knows sweet bugger all about options, thinks that AAPL will go down or NFLX will go up and buys puts/calls as appropriate makes a bundle. He's the MAN and that's all that counts in this business.