Writing options for a living

[riskarb]

Quote from Prevail:

Mav, we disagree on a few points but I really enjoy your posts. In your example the atm was sold with the opinion that the market would not move too far too soon before the wings were added. This would probably be what riskarb meant when he said edge bleeds into the trade in his journal.

If you were to run the simulation 1000 or so times as you mentioned, clearly the result would be a positive expectancy or edge as the complete position is essentially riskless. But, what if the trade simulation was run from the origin of the trade when the atm was sold. Then x days later the wings were added. There would at the least have to be a positive opinion that the trade would have positive expectancy I would think. Would riskarb sell the atm's in the first place if he did not think there was an edge?

Edge bleed refers to a mark to market calculation of P&L. The gamma:theta is linear. Since the atm straddle is downside-gamma[curvature] there is spot-vol concerns expressed as delta-position. Obviously gamma drives the delta, but the gamma isn't geared in the straddle.

Mav certainly got it right, but it's a GOOG combo. I expect GOOG spot vols to decline to the 28-vol-line[from 30v] for the atm combo. The daily vols have been spotty, but generally mean-reverting to the 300 handle. I don't see a lot of directional-vol in the shares. The wings exhibit opportunity greeks, as my hand is not forced if my spot vol prediction is correct, and the potential vol/theta gains from the wings are replicated by buying the wings < current offer on that combo.

Add to that, the wing vol is high due to the prevailing 200basis in call skew on the 20d otm vol-line. There is still some hope by the public that the stock will exhibit upside vol, which I don't expect, so a flattening of the smile will help.

Summary: The edge will be expressed by a larger credit on the converted iron fly than is currently possible.

 

[dummy-variable]

Quote from thenewguy:

So a system has to work for infinity to be a sucessful system?

Also, it wasn't really their system that brought them down, but lack of appropriate money management.

- The New Guy

a system doesn't have to "work to inifinity," but if it's drawdown not only bankrupts the firm but nearly causes a financial markets meltdown, it becomes a little suspect. i think ltcm is germane to the original point of this thread as well - namely that writing options for a living as a be all/end all strategy does not work. whether it was the system or the money management is an issue but generally the two are initmately related.

 

[MajorUrsa]

Quote from smilingsynic:

Are you referring to the new book THE COMPLETE GUIDE TO OPTION SELLING? I read it at B&N for about fifteen minutes a few weeks ago. I was thoroughly UNIMPRESSED. Cordier and his co-author--both brokers--give little attention to volatility and to risk. Their article in this month's FUTURES (I get it free--I didn't pay for it) was full of errors.

I agree. I got the book by accident (coupled order at Amazon).

There are indeed a lot of errors in it. Also theoretical foundation is absent ("leave that to the quants") and other than saying that FOTM options are so ridiculously far away that risk is almost absent there is no discourse on the actual risk involved at all.
Nowhere a hint is given that the premium of an option is determined by the risk you run.

Frankly, I think the authors don't understand at all what they're writing about. Either that, or they are on the verge of misleading. The style is very friendly and it wouldn't surprise me of many fell for the method.
If anything, he will supply the market with OTM sellers, but for the rest I think it should be forbidden.

Ursa..

 

[Prevail]

Quote from dummy-variable:

first the expectancy that the losers have to be huge is a relative concept. you'd have to be selling 10 delta options consistently for this to be the case.

At the very least when dealing with options with deltas of <.5 the losers have to be larger but as you said dealing with 10 delta options would warrant a huge loss to keep the expectancy at 0. If this loss is in any way mitigated positive expectancy is created, which is what we are discussing.

secondly, a stop loss does not negate the negative expectancy.

Agreed you can never be sure but when it does happen, which is frequent in my experience, positive expectancy is created.

i know that you meant a price trigger for the stop. i just used the time example to make the concept simple. if you used a price stop instead, the principle still holds. no matter where you exit a trade if there has been no intervening adjustment or spreading that alters your net expectancy, you are essentially adding two negative expectancy trades together. two minuses added together in this case do not make a positive.

This is what perplexes me, why would you enter into a trade in the first place if there was no positive expectancy? If you have to have a 0 expectancy trade go your way to make an adjustment to induce positive expectancy then would there be the same likelihood the initial 0 expectancy trade would not go your way to make a positive expectancy adjustment?

if you are saying that you don't think prices are efficient, you are mistaken. if this were true, there would be limitless arbitrage opportunities. [/B]

Yet in your recent post you said we can never know if prices are inefficient. If we never know when they are inefficient, how can we know they are efficient?

i've heard a variety of this type of claim before and none of them holds much credence. one common misconception is that the volatility skew or "smile" is somehow an indicator of inefficiency. people claim that the skew violates the premises of the BS pricing model. OTM put strikes frequently have higher IV than OTM calls. so the logic runs that one should buy the high strikes and sell the low strikes. but there is no reason why the skew isn't an accurate reflection of probabilities.[/B]

I'm with you in that the shape of the skew is usually accurate and representative.

i use this example. assume a stock where there are only two sets of investor opinion. 50% believe the stock is worth $48 and 50% believe the stock is worth $50. efficiency would therefore place the fair value of this stock at $49 (.5*48+.5*50 = 49). now what would be the IV of a 45 strike option in this scenario? obviously pretty close to zero since the market is in a very tight consensus. the volatility smile would even possibly turn into a frown in this case.[/B]

now assume that new information causes market participants to alter their opinions on the value of the stock. now 90% believe the stock is worth $50 but a small minority, 10%, believe the stock is only worth $40. the efficient price on the stock does not change; it is still $49 (.9*50+.1*40 = 49). but what happens to that 45 strike option now. suddenly there is a lot more interest in owning this strike because there is a much wider range of opinion on the stock. the result would be a typical smirk-type skew where the puts would now be inflated due to the higher probability of their being in the money. the stock price is just the tip of the iceberg in terms of the aggregation of opinions about value. the skew tells a more granular picture of probabilities at discrete prices.[/B]

would you consider the put buyers overpaying in this case? [/B]

I very well might. Higher prices are warranted, but how to calculate they are high enough, too high or not high enough? Which pricing model would you like to use today?

 

[Prevail]

Quote from Maverick74:

I'll let him answer but my guess is he would sell the ATM combo knowing there is negative edge in the trade. He, like me, acknowledges there is negative edge in all trades when they start out, but he is attempting to morph the trade into something that will carry some positive edge at some point in the future.

He may be dead wrong. Maybe EBAY gaps down 20 pts tomorrow. This is the risk of this business. We are not teaching 3rd grade public school here. I think we can all agree that somewhere, a risk has to be taken.

However, a good trader will find ways to offset this risk going forward and improve the risk/reward ratio over time. Again, I'll let him respond to your question as I don't want to put words in his mouth.

I was addressing you but didn't mean to make you feel required to speak for another, many apologies.

In my experience the majority of traders do not enter a trade without at least believing there is a positive expectancy. What you are saying is your are content to enter a trade with no mathematical expectancy, or negative for that matter, and try to adjust it to bring out a profit. That's intriguing, and not for me .

Really much of this thread can be reduced to discretionary trading vs. systematic. Discretionary traders are into adjustments, systematic traders are not. The former believe positive expectancy is built into trades through action, the latter believe positive expectancy is in the next trade before the initial trade is placed.

 

[Prevail]

Quote from riskarb:

Edge bleed refers to a mark to market calculation of P&L. The gamma:theta is linear. Since the atm straddle is downside-gamma[curvature] there is spot-vol concerns expressed as delta-position. Obviously gamma drives the delta, but the gamma isn't geared in the straddle.

Mav certainly got it right, but it's a GOOG combo. I expect GOOG spot vols to decline to the 28-vol-line[from 30v] for the atm combo. The daily vols have been spotty, but generally mean-reverting to the 300 handle. I don't see a lot of directional-vol in the shares. The wings exhibit opportunity greeks, as my hand is not forced if my spot vol prediction is correct, and the potential vol/theta gains from the wings are replicated by buying the wings < current offer on that combo.

Add to that, the wing vol is high due to the prevailing 200basis in call skew on the 20d otm vol-line. There is still some hope by the public that the stock will exhibit upside vol, which I don't expect, so a flattening of the smile will help.

Summary: The edge will be expressed by a larger credit on the converted iron fly than is currently possible.

It is pretty obvious we all respect your options abilities. Could you possibly explain how a trader who sells the goog atm combo because they believe in 5 days the trade would be profitable due to what they perceive as positive expectancy and a trader who sells the goog atm combo because they believe in 5 days the trade would be profitable enough to what they perceive is positive expectancy?

 

[MajorUrsa]

Quote from riskarb:

Edge bleed refers to a mark to market calculation of P&L. The gamma:theta is linear. Since the atm straddle is downside-gamma[curvature] there is spot-vol concerns expressed as delta-position. Obviously gamma drives the delta, but the gamma isn't geared in the straddle.

[snip]

Summary: The edge will be expressed by a larger credit on the converted iron fly than is currently possible.

Your method of buying cheap irons is still looking very attractive, and of course owning an iron fly for less than it should cost gives you a statistical edge, once it is established.

However, and this question is nagging me for some time now, at the moment you decide to buy in the wings (because they got cheaper) you'd also have the possibility of buying back the original center straddle, which will earn you money immediately.

By not buying back the straddle but instead buying the strangle you are in effect re-investing your current (paper) profit to buy a iron fly. Ie., for demo purposes, you could first buy back the straddle (making a profit) and then buy the full iron at once.

So the question is, why do you want to own the fly (instead of the money)?

To me this situation is akin to what I learned earlier. Suppose you buy a call for 0,50. After a while the underlyer went up a strike (5) and instead of selling the call for 2,50 you sell one a strike up for 0,50. You now seem to own a 'free' vertical, that may expand to 5 on expiration. But is it free? Not IMO. You payed 2.00 for it, just as a vertical would've cost you. You could've sold the call for 2.50, so why buy the vertical instead?

I think this is very relevant to this discussion. Can a strategy have a positive expectancy if the single steps in it don't? I'm getting inclined to say no.

Of course in the case of your irons there could be other considerations that make the strategy workable, like outward skew and margin properties. But the fact remains that at the moment you complete the fly you decide to buy a fly, using the unrealized gains from a previous position as well.

Is there any such thing like unrealized losses (or gains), btw?

Ursa..

 

[smilingsynic]

Quote from MajorUrsa:

I agree. I got the book by accident (coupled order at Amazon).

There are indeed a lot of errors in it. Also theoretical foundation is absent ("leave that to the quants") and other than saying that FOTM options are so ridiculously far away that risk is almost absent there is no discourse on the actual risk involved at all.
Nowhere a hint is given that the premium of an option is determined by the risk you run.

Frankly, I think the authors don't understand at all what they're writing about. Either that, or they are on the verge of misleading. The style is very friendly and it wouldn't surprise me of many fell for the method.
If anything, he will supply the market with OTM sellers, but for the rest I think it should be forbidden.

Ursa..

My guess is that these guys wrote the book to drum up clients (and to sell books). Reminds me of the way Wade Cook (not that these guys are con-artists like him) wrote books that were infomercials in print. They're brokers, not traders.

Note how they strongly recommend not using discount brokers.

 

[vulture]

Quote from smilingsynic:

My guess is that these guys wrote the book to drum up clients (and to sell books). Reminds me of the way Wade Cook (not that these guys are con-artists like him) wrote books that were infomercials in print. They're brokers, not traders.

Note how they strongly recommend not using discount brokers.

yes, it always helps to be a "synic".

 

[dummy-variable]

Quote from Prevail:



[...] you can never be sure but when [ a stop loss that negate's negative expectancy] does happen, which is frequent in my experience, positive expectancy is created.

how can you possibly know this? do you have some data that shows you've found the optimal stop-loss level (i.e. you've tested all your stop theories to see what would have happened and found that your stop produced a positive expectancy)?

i can say with strong confidence that in fact you are wrong. every test of stop losses i've seen (or conducted) shows the same thing: the best "returns
(which are actually the lowest losses) involve buy and hold. why? because in the markets almost all the gains from a no-adjustment system come from outlier events. i assure you, simply having a stop-loss rule added to your initial opening strategy merely reduces your overall expectancy.

i'm not saying that you should not practice good money management. a stop tied to your position size and max risk tolerance is essential. but all stops can do is extend the longevity of a trader's life. they do not turn you into a winning trader.

This is what perplexes me, why would you enter into a trade in the first place if there was no positive expectancy?

because you don't have a choice. that's the price you pay to play this game.

If you have to have a 0 expectancy trade go your way to make an adjustment to induce positive expectancy then would there be the same likelihood the initial 0 expectancy trade would not go your way to make a positive expectancy adjustment?

yes this is called "risk". you make a bet. when it goes your way you turn the bet into a lock (or at least a positive expectancy scenario). otherwise you take the loss. the goal is to learn how to capture more of those positive expectancy adjustments and let the rest of the trades flush out at zero expectancy.

Yet in your recent post you said we can never know if prices are inefficient. If we never know when they are inefficient, how can we know they are efficient?

again it's an assumption of the game. you CAN find inefficiencies but the cost and means of doing so will be far outside the average retail trader's resources. you are much better off devoting yourself to learning how to trade rather than trying to find market anomalies.

I very well might [sell "high" priced puts]. Higher prices are warranted, but how to calculate they are high enough, too high or not high enough? Which pricing model would you like to use today?

i don't need a pricing model. i assume the market knows more than me and has priced the options fairly. other than with a violation of put/call parity, i have no way of knowing if an option is over or under priced. i can have an OPINION and make a bet based on that sentiment. but then i have to trade that position as though i was wrong until i'm proven right.

i suppose there are traders out there who are unbelievably prescient when it comes to guessing future price trends or IV direction. you may be one of these folks. but in general the rest of us have to live with option prices as though they are what they are which means they are fairly valued.