Help understanding trend following with options

[sinner]

Hi guys,

I was wondering if someone could help me with an idea I had in considering long term trend following systems which would usually use the cash index or equivalent vehicle to trade.

For example a simple ruleset:
* At the final day of each month:
** If the closing price of the cash index is > 10 month closing price SMA invest in the index
** If the closing price of the cash index is < 10 month closing price SMA invest in cash

Can this be converted to an options trading rule, such as:
* One week before OpEx:
** If the closing price of the cash index is > 200 day closing price SMA invest N% of assets in long 1m call options at the strike price closest to the 200 day closing price SMA; where N = distance in % from 200 day closing price SMA.
** If the closing price of the cash index is < 200 day closing price SMA invest in cash

In this way you should only suffer a % loss equal to the distance from the SMA in any given month (plus long call premium) in the event that the price declines below the SMA during the course of a month.

The reason I am interested in this idea is basically because long term trend following systems which trade only once a month avoid whipsaws for systems which don't utilise hysterisis, but also suffer greater drawdown during events like Oct 1987 and similarly in Jul/Aug/Sep 2011. I feel like using long calls instead of buying the underlying will allow me to take advantage of once a month trading while still being protected from large intramonth movement.

What I'm concerned about:
* The drag on returns which paying long call premium each month will have.
* The logic of my options trading rule not being logically equivalent (i.e. suffers greater drawdowns and returns less on trends)
* High IV forcing me to overpay for long calls (although my data indicates most of the high IV months are where cash index < trend filter).
* Anything I can't obviously think of

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=962461


Please help.

 

[chops_a_must]

Can you actually back test options now?

 

[chops_a_must]

Someone like WayneL will be able to tell you more, but there are a few things that come to my mind.

Why would you want the options, presumably in the money?

Any downside, may lead to a death by a thousand cuts scenario, unless you're hedged at a price point.

Upside is relatively un leveraged compared to other strategies to the upside.

You'll be paying time premium, which will mean any gains would be less than holding the underlying.

Personally, I'd be looking at out of the money call leaps as the price crosses below the 200ma.

 

[wayneL]

I'm not fully au fait with trend trading options, but have played around a bit.

Most edumacators recommend buying long dated, DEEP ITMs at say 70 or 80 deltas. Although there is nothing inherently wrong with this, I don't think it offers any advantage over trading the underlying... it's skinning a cat with a blunt knife IMO.

For me, you may as well try to take advantage of the non-linear attributes of options AKA the Greeks.

A couple of principles I work along:

1/ Nobody says you have to hold till expiry. Avoid high theta by rolling out before premium starts disintegrating.

2/ Forget the number of shares, think in terms of deltas. IOW if your position sizing algorithm says to buy 700 shares for a stock position, then acquire 700 deltas in your chosen option strategy.

3/ As much as negative theta is your enemy, long gamma is your friend, it gives you a natural pyramiding if your stock puts on a fast move... also a natural de-pyramiding if it moves against you.

4/ In respect of the above, despite the grave (and grossly generalized) cautions from edumacators, don't afraid to go OTM... again deciding on expiry and how far OTM depending on the best Greek exposure for the anticipated move.

5/ Use statistics and probabilities to write some calls at prudent times... again matching deltas, to try to pay for the theta.

6/ Consider the effects of vega. A long gamma position may not be optimum at times when premium sag is a danger.

Just a few points to ponder on

 

[village idiot]

Hi guys,

I was wondering if someone could help me with an idea I had in considering long term trend following systems which would usually use the cash index or equivalent vehicle to trade.

For example a simple ruleset:
* At the final day of each month:
** If the closing price of the cash index is > 10 month closing price SMA invest in the index
** If the closing price of the cash index is < 10 month closing price SMA invest in cash

Can this be converted to an options trading rule, such as:
* One week before OpEx:
** If the closing price of the cash index is > 200 day closing price SMA invest N% of assets in long 1m call options at the strike price closest to the 200 day closing price SMA; where N = distance in % from 200 day closing price SMA.
** If the closing price of the cash index is < 200 day closing price SMA invest in cash

In this way you should only suffer a % loss equal to the distance from the SMA in any given month (plus long call premium) in the event that the price declines below the SMA during the course of a month.

The reason I am interested in this idea is basically because long term trend following systems which trade only once a month avoid whipsaws for systems which don't utilise hysterisis, but also suffer greater drawdown during events like Oct 1987 and similarly in Jul/Aug/Sep 2011. I feel like using long calls instead of buying the underlying will allow me to take advantage of once a month trading while still being protected from large intramonth movement.

What I'm concerned about:
* The drag on returns which paying long call premium each month will have.
* The logic of my options trading rule not being logically equivalent (i.e. suffers greater drawdowns and returns less on trends)
* High IV forcing me to overpay for long calls (although my data indicates most of the high IV months are where cash index < trend filter).
* Anything I can't obviously think of

Please help.

Sinner, me too been thinking recently about long term trend following system using options strategies instead of the underlying, but havent got very far with it so far.

A couple of points on your system firstly;

1. why would you initiate one week before expiry? surely long term systems are crying out to make bets over the whole month, not one quarter of it, which by coinicidence perfectly meshes with the option expiry cycle. On a long term monthly system I would like to make my bets once a month, for the whole month.

2. Not sure about the logic of using a strike at the MA, and then allocating a higher % of assets to buying calls the further away that strike is from the money. havent looked at any figures on it, but it may well result in less overall + delta (than ATM strike) for more at risk, whenever momentum is higher . that may have been your intention but maybe not? does your system call for different position size when the index is further from its ma?. If so that can be built in but I don’t think this neccessarily does it

3. Agree with wayne that simply using a long ITM call instead of the underlying is as he delightfully puts it using a blunt knife. It may have the benefit of the gamma working for you and the more obvious benefit of a stop loss built in, but also has the drawbacks of buying time value at inflated IV and the extra costs of higher spreads on a relatively illiquid instrument.

4. My feeling is that any strategy that is habitually theta negative is not going to go well long term, based on stocks don’t tend to move enough often enough to get over the high time decay of a strike at the money, or a strike below the money which is generally at an inflated IV as noted above. Now it may well be that the trend following system has enough Positive expectancy in itself to get over that drag, but nevertheless it is drag and OTBE it will underperform in the uptrends you are trying to capture.

So what might work? Well I am looking at trying to reduce each month down to a binary bet using vertical spreads, more on which later...
If I had to just buy calls OTM calls may be a better bet than ITM

 

[sinner]

Thanks very much for the answers guys! You have given me some things to think about.

village idiot, what if I rephrase my request in the following fashion:

If I told you I had a moving average which threw off a buy or go-to-cash signal once each month, and that I wanted advice on which option strategy would be most successful in capturing:

* All (or as much as possible) of the "monthly returns" of the underlying if those returns are positive (relative to the underlying price at time of signal).
* Or, a monthly % loss of my equity no greater than the % distance between the underlying price and moving average on the signal day.

What would you suggest?

 

[chops_a_must]

Thanks very much for the answers guys! You have given me some things to think about.

village idiot, what if I rephrase my request in the following fashion:

If I told you I had a moving average which threw off a buy or go-to-cash signal once each month, and that I wanted advice on which option strategy would be most successful in capturing:

* All (or as much as possible) of the "monthly returns" of the underlying if those returns are positive (relative to the underlying price at time of signal).
* Or, a monthly % loss of my equity no greater than the % distance between the underlying price and moving average on the signal day.

What would you suggest?

Put spread.

 

[chops_a_must]

Put spread.

No, that's not write.

 

[wayneL]

Thanks very much for the answers guys! You have given me some things to think about.

village idiot, what if I rephrase my request in the following fashion:

If I told you I had a moving average which threw off a buy or go-to-cash signal once each month, and that I wanted advice on which option strategy would be most successful in capturing:

* All (or as much as possible) of the "monthly returns" of the underlying if those returns are positive (relative to the underlying price at time of signal).
* Or, a monthly % loss of my equity no greater than the % distance between the underlying price and moving average on the signal day.

What would you suggest?

Sinner, I don't think what you want to do can be done optimally with options. For what you want to achieve, I think long stock with a stop loss is the best way.

Where long options can possibly outperform is (exactly where trend following systems make most of their profit) in the outliers. That means using a good balance of gamma/theta and making adjustments to these at statistically or technically advantageous points.

So far, I like my gamma north of the entry. This, slightly underperforms stock in small moves, but kicks @ss big time in a good strong move.

I had some forward tests which reflected this, been rummaging around my 'puter but cant find them at the moment.

 

[village idiot]

Thanks very much for the answers guys! You have given me some things to think about.

village idiot, what if I rephrase my request in the following fashion:

If I told you I had a moving average which threw off a buy or go-to-cash signal once each month, and that I wanted advice on which option strategy would be most successful in capturing:

* All (or as much as possible) of the "monthly returns" of the underlying if those returns are positive (relative to the underlying price at time of signal).
* Or, a monthly % loss of my equity no greater than the % distance between the underlying price and moving average on the signal day.

What would you suggest?

I certainly dont have a definitive answer to your question, but trying to answer it at all has at least caused me to go and have a look at a few things, since i needed to know what sort of profits and losses we were targeting/avoiding. taking XJO as an example, monthly since 2001, in months when your trigger would have had you in;
max gain 7.3%
average gain (counting only up months) 2.61%
average loss (in down months) -2.82%
worst loss -10.88%
percentage of winning months 66.7%
percentage losing months 33.3%

When looking at your requirement to not lose more than the distance between close and the 10 month MA: at the end of feb this year XJO was around 14% above its MA. Restricting a loss to that amount is not really restricting it at all - I do not think you really mean trade a system that can lose 14% of an index in a month, so I think that particluar parameter is an irrelevance. Presumably the point of using options is to restrict losses a lot more than that, but happily that is easy enough to do.

With the above figures in mind, here is some doodling I wrote yesterday before I did those figures;

So what are we looking for, where is our positive expectancy coming from when we trade a trend following system? Lets break it down . For the system to have positive expectancy for each unit of time in the system , each 'bet', (eg 1 month for a monthly system, and lets assume a long only system here) we must think that either a) there is a higher probability than usual that the stock or index will be up over the unit of time or b) there is equal or 'normal' probabilities of up v down, but if up then it will go up further than it would go down if down or c) some combination of a and b.
There is another school of thought that neither a or b might be true over one unit but over multiple time units letting winners run but not losers may turn the system as a whole profitable. Personally I don’t see how that can be true without either a or b or both being true, so i will stick with the one unit building block being the key.

Now comparing that thought with those figures for the XJO, there is no evidence that winning months run further up than losing months run down, if anything the opposite, or that profits rely on a few outlying months. The positive expectancy if any is more likely to come from the fact that there are more up months (67%) than usual when the system calls to be 'in'. "usual" in this case is 58% of months in the whole sample are up months, which is in line with previous analyses i have done on various indexes.

If that is the case, then what strategy targets that particluar observed behaviour? The first idea that springs to my mind starts with using a bull vertical spread with a narrow gap between strikes and which is theta neutral, effectively making each month a binary win or lose bet.

I will note here that the skew present in most stock and index options makes a vertical spread with strikes one above and one below the SP typically need to win about 55-58% of the time to breakeven. Thats what it was yesterday on XJO when I rechecked. Most of the time that matches the actual behaviour of indexes as noted above, so its a neutral expectancy bet, but if one has some system of identifying where there is a 67% chance of an up month, there is clear positive expectancy.

I did a crude model of such a system last night using your trigger vs using your system with cash alone (without intramonth stop losses). It turned out similar shaped equity curves between the two,but if normalised so each system risked the same amount of account equity each month, the option based system did considerably better, driven by the fact that in the option based system the loss is small and known, so position size can be considerably higher. As wayne suggests, using futures with an intramonth stop loss may achieve the same thing more optimally. I dont know, i havent investigated that. neither system really 'worked' since 2008.

I suspect however the theta neutral bull spread may not even be the optimal vertical spread strategy, let alone the optimal strategy. I have a feeling it may be better to use a spread which allows larger profits than losses, although to achieve this we have to pay some amount of time value ie be theta negative, and neccessarily reduce the frequency of winners. But that will take quite a few hours of testing to investigate. To further complicate matters the optimal spread would vary with IV and skew at the time.



disclaimer; this is not a statement that any strategy or system works, is optimal, is better than any other strategy with or without options, or should be used. it is more just an exploration of ideas out loud, to be shot down.....

 

[village idiot]

but the rule set as tested;

at each option expiry;
if index is above its 10 month SMA, buy a vertical call spread for the following month 1:1 with lower strike x% below current close, and higher strike x% above current close. Position size based on % of account equity you wish to risk divided by max loss on spread.
if index is below its 10 month SMA, do nothing

the lower the value of x the better it worked on an account equity risk adjusted basis


strictly speaking that is not really theta neutral because of the skew which was built into my cost assumptions but is close to it for small values of x