Statistical edge with option spreads -none?

[MechTrade]

Quote from optionsgirl:

... but I am just referring to pure options position and disregarding fundamental and technical factors. I tried to mean when picking a random option at a random time without any mispricing throughout the strike range, there isn't a statistical edge of one position versus another --or that is what I assume.

Well, that clears up a lot. Random selection of trading strategies with absolutely no trade management offers no edge whatever. I agree completely.

If I picked random financial advisors at random times, disregarding past performance. strategy and honesty factors, I might have an edge trading your random Options scenario

Everyone reading this forum is a living, breathing, rational (well, some of us) trader who do not pick trades randomly, nor do we just let our positions go to expiration without managing them.

I recently posted that Options themselves are an edge. I did not claim anything about a statistical edge.

If you take a position directly in a stock or a future then your P/L has an immediate linear relationship to the movement of that issue.

But with Options you can structure your positions to be resilient against adverse price moves or greek changes within a range that you can define by the position itself. And the positions can be structured to provide for relatively easy management.

What is interesting is that evaryone reading this thread is an option trader. I doubt that they would be trading options if they did not feel that they had an edge of some sort on each trade that they put on.

Perhaps a better thread would be Options as a Traders Edge, leaving out the Statistical inference.

Entertaining thread though!

Mech

 

[heech]

Quote from asiaprop:

who claimed arb opportunities are sustainable?...

On the other side with a working vol model it is indeed possible to persistently identify under or overvalued options.

Right. Arb opportunities are not "sustainable", but options are "persistently" mispriced...

Anyways, that aside, you need to get your eyes checked if you think "no arb" means "no profit". "No arb" just means you're "only" getting an appropriate reward for the risk you're taking on in the trade.

If you don't understand the distinction, then you really shouldn't be in this conversation.

 

[spindr0]

Quote from heech:

I know it's pretty much standard operating procedure on ET to pass off anecdotes as "strategy"... but sorry, not an impressive story unless you can also give us specifics of what he was doing, and how backtesting was conducted, and especially how it performed last fall.

Thank god that rock star traders making hundreds of thousands per transaction can break away from the money train to edumacate us po folk.

 

[atticus]

Quote from asiaprop:

On the other side with a working vol model it is indeed possible to persistently identify under or overvalued options. No need to take a directional bias, in fact I traded most of the time other options against such positions in order to mitigate any remaining vol, skew, or gamma exposure.

It really isn't possible in listed markets. Any persistence would take place in up or down&out strikes in which you would need to carry substantial gammas to realize any significant numbers. The best empirical evidence is the down&out skew in equity index markets. For one, I would love to see how to avoid gamma in an index skew trade. The surface would be traded flat if it were possible.

Flow impacts pricing that can result in, say, 300bps in skew on a near-atm strike in XYZ, but that's not really what we're talking about here, as you're referring to hedging vega, gamma, speed, and... You would refer to trading conversions, boxes and rolls. OTM gamma/vega is a good sale, but it's not enough compensation to maintain a static gamma position.

Dealer mispricing is the only instance in which I've witnessed persistence. Recently a pricing error on dealer-traded exotic calls traded with OTC and listed vanilla puts and spot to effect the reverse-conversion. It's a major bank and it's been evident for at least two months. This would never last in a transparent market. The dealer in question is only seeing the call side of the transaction and is known to limit hedging to deltas.

 

[atticus]

Quote from donnap:

I read somewhere that in the early years of listed option trading - the 1970s - that options generally were overpriced.

If this is accurate, then there was an edge to selling options that was eventually eliminated.

How about the late-80s. Equity puts weren't even listed until 1977. It was more an issue of spreads than the vol-line. A 60-day bond straddle at the CBOT could have been 3'00 x 3'45 in 1988.

 

[Wayne Gibbous]

Just as a FWIW there may be evidence there is a small amount of overpricing in index ops (and it seems like every IC trader and his brother feels they have an advantage)?

At least, the BXM is starting to pull away. Maybe if commish and slippage is figured in it is more of a flatline?

<img src="http://ichart.finance.yahoo.com/z?s=%5EBXM&t=5y&q=l&l=on&z=m&c=%5EGSPC&a=v&p=s"/>

 

[asiaprop]

you quote me out of context and you know it. Shame on you.

When I spoke about "arbing" I specifically talked about arbing MMs, please re-reach the earlier post. Those are not sustainable and come along only once in a while.

However, option "mispricing" can be consistently identified with a well developed and updated vol surface and underlying model. Just imagine an insurance firm buying a lot of the at the money 2nd front month puts. Huge demand will make those overvalued relative to the front month or further out expiries. Selling those and offsetting the remaining exposure in a smart way will give you an edge. Point is to identify those "mispricings" before others step in to do the same.

You should not jump to quick conclusions about others' comprehension skills because you could embarrass yourself. I am not here to stage a war so no need for personal attacks ;-)

Peace

Quote from heech:

Right. Arb opportunities are not "sustainable", but options are "persistently" mispriced...

Anyways, that aside, you need to get your eyes checked if you think "no arb" means "no profit". "No arb" just means you're "only" getting an appropriate reward for the risk you're taking on in the trade.

If you don't understand the distinction, then you really shouldn't be in this conversation.

 

[asiaprop]

you can do the same. Get an ivy league degree (or be extremely street smart without), apply to the top hedge funds or sell-side firm, and work very hard, and be disciplined, then you will be given the responsibility to trade your own book down the road and you are all set...ups...almost said the sky is the limit...but truth is....your abilities are the limit ;-)

Summary: Don't bitxx but imitate.

Quote from spindr0:

Thank god that rock star traders making hundreds of thousands per transaction can break away from the money train to edumacate us po folk.

 

[optionsgirl]

Quote from asiaprop:

this is where your assumption is wrong. The distribution of the underlying prices is NOT normal. That is one of the most fundamental issues in options trading. And this is where some of the models have the edge. Options are priced according to which probabilities the market for up and down moves implies. Those probabilities are pretty much always skewed. If you model disagrees with this skew with a statistical edge then you have a strategy with edge.

Well, it's not really my assumption, I'm just speaking in terms of a basic general assumption. I've read about fat tails, skew, etc. From my early impression of reading about model tweaking/creation, it is often indicated that there's not much benefit of tweaking the model itself. Also, I'm not sure how a retail trader can benefit from it other than being conscious that the standard Black-Scholes is imperfect. Many times, the argument about option pricing comes down to the whole random-walk/market efficiency debate about the underlying instrument. If you have 10 people who can independently predict the market with a 70 percent accuracy rate (or more) and they presented 10 different methods, how would you extract information from all those methods and incorporate them into a new model? I doubt that all the details of the 10 methods would not contradict each other. Black-Scholes provides the unbiased middle ground. Plus, we have implied volatility to represent it's deviation from the model and the differing of opinions --sort of...

In my somewhat naive conclusion, I am inclined to believe that creating new models is just like reinventing the wheel...

 

[dmo]

Quote from optionsgirl:

Well, it's not really my assumption, I'm just speaking in terms of a basic general assumption. I've read about fat tails, skew, etc. From my early impression of reading about model tweaking/creation, it is often indicated that there's not much benefit of tweaking the model itself. Also, I'm not sure how a retail trader can benefit from it other than being conscious that the standard Black-Scholes is imperfect. Many times, the argument about option pricing comes down to the whole random-walk/market efficiency debate about the underlying instrument. If you have 10 people who can independently predict the market with a 70 percent accuracy rate (or more) and they presented 10 different methods, how would you extract information from all those methods and incorporate them into a new model? I doubt that all the details of the 10 methods would not contradict each other. Black-Scholes provides the unbiased middle ground. Plus, we have implied volatility to represent it's deviation from the model and the differing of opinions --sort of...

In my somewhat naive conclusion, I am inclined to believe that creating new models is just like reinventing the wheel...

Again optiongirl, coherent thoughts well expressed. All any model can do is provide a starting point, a tool, for evaluating and comparing one option with another. From that point on it's up to you. BS and the like (Whaley, CRR, etc.) do the job. I'll concede that for a firm doing large-scale automated trading, part of the algorithm could be incorporated into a proprietary model. But for someone trading their own account, the search for a perfect model is as pointless as the search for a perfect option strategy.