Anyone here trade options?

[utrecht]

If anyone here trades options, please private message me. I'd like to discuss a strategy of trading weekly options that Ive been using.

 

[ERD50]

I'd be very interested in discussing it in public. This is a public forum, and I'd like to benefit from the input of many. I reserve PMs for very specific cases.

I've been delving into the weeklies (SPY). Theoretically, the weeklies fit my MO better, but these market swings have been mostly working against me.

Whatcha' been up to?

-ERD50

 

[utrecht]

Strangely enough the weekly SPY options is what Ive been trading also. I've been using a pretty mechanical method to sell weekly puts. I paper traded it for a while and then started using real money 14 weeks ago. Over those 14 weeks the SP500 is -12%. My weekly put options have returned +1.3% including transaction costs which are steep until you start trading large positions. This includes the pummeling I took week before last.

What I wanted to discuss with someone with some experience is how effective you think it is to back test an options trading method. Since there are no historical weekly options prices to look at, what I did was take the avg option premium percentage Ive been getting and then apply that to weekly SPY prices going back to the beginning of 2005. I used the most conservative numbers I could so i think the results would actually better than what I came up with.

Real time real money results for 14 weeks:

Weekly naked SPY puts: +1.3%
SP500........................ -12%
********************************************************

Backtest Results for YTD 2011:

Weekly naked puts: +19.9%
SP500..................: -5.2%

CAGR for 2005 - present

Weekly naked SPY puts: +19.9% (not a misprint, same is YTD 2011)
SP500........................: +1.7%

Transaction costs are not included in these back test numbers. Based on my position size, they should be about 1.5% per year.

The option premium I received for the position I opened yesterday was 25% higher than the avg number I used for the back test so it does vary a lot when volatility is high which makes it really impossible for the back test to be really accurate but like I said, I went very conservative purposely. I believe the percentage I used should reflect very low volatility.

The back test showed that this trading method crushed results for just buying SPY in every year. From 2005-present we've had just about every possible scenario from great year to good year to mediocre year to horrible year and the weekly naked puts trounced SPY every time. Having said that, I realize that they weren't even available back then.

 

[FIRE'd@51]

Over those 14 weeks the SP500 is -12%. My weekly put options have returned +1.3% including transaction costs which are steep until you start trading large positions. This includes the pummeling I took week before last.

My first question concerns how you calculate the return. Is it $-earned divided by the average margin required to carry the position?

 

[ERD50]

What I wanted to discuss with someone with some experience is how effective you think it is to back test an options trading method. Since there are no historical weekly options prices to look at, what I did was take the avg option premium percentage Ive been getting and then apply that to weekly SPY prices going back to the beginning of 2005. ...

Trouble with back-testing, as you say - no good history (that I could find) for option prices, and since they don't usually trade as often as the underlying stock, you really need to bid/ask to compare with the stock price at a moment in time. I've seen the bid/ask move up and down with a stock for quite a while, before anybody actually 'bites' and executes a trade that is recorded. So even historical data could really be out of sync.

Theoretically (I'll try to see if I have any spreadsheets on this), you could take SPY, VIX and calculate an option price with the Black-Sholes formulas.

You need to account for VIX. Option premiums vary with implied volatility, and if you try using a static % premium you will get a distorted view.

I was doing reasonably well, selling OTM calls on SPY, near month, trying to just get ~ 0.5~1% premium. I'd get called out (actually, I'd BTC the call near expiry) maybe 1 out of 10 months, and take some "opportunity loss", but this was being positively offset by the average of the expired premiums. Note that when I say "reasonable" gains, I mean maybe 1-2% annually over/above straight B&H SPY. Not enough to make the cover of a financial magazine, but it's huge compared to 3.5% WR!


Lately, the actual moves in the market seem to be just flowing against me compared to what the option pricing would predict. Probably just a weird time for the market. My viewpoint is that I should be able to make a small profit long term - I am like 'the house' selling chances on the wheel to the 'gamblers'. If you are selling puts, it is similar - you are selling 'insurance', and people should expect that the insurance holder is going to need to make a profit to cover their risk.

But for any given week, month, or year - it could work against the house.

Your thought?

-ERD50

 

[ERD50]

My first question concerns how you calculate the return. Is it $-earned divided by the average margin required to carry the position?

Yes, very important. When I traded some 'naked' puts, I preferred to use the term "cash-covered" puts. Maybe your broker lets you cover them with margin, but don't lose sight of the fact that the full amount of money is behind the trade. You could have that money working in other investments, so you should use the full amount of your 'cover' as the denominator in your yield calc.

A friend of mine figured he made an 'infinite' return when he sold an option, then bought it back at a lower price. Selling an option ties up the capital it would take to own the underlying. I don't think I convinced him, because he liked his view better!

-ERD50

 

[utrecht]

Heres the formula I use:

Assume that I sell one SPY put. I receive $2 in premium and the stock price is 120.
If the put expires worthless I make $200. If I was forced to buy the stock it would cost me $12000 so I divide $200 by the $12000 I was theorectically tying up. So my return is 1.67%. This is similar to buying 100 shares of SPY for $120 for a total of $12000 and then selling it for a $200 profit.

Technically I never have to put up the $12000 because I buy the put back before it expires if its not going to expire worthless and Im not technically tying up a full $12000, but thats the only formula I can come up with for return.

 

[ERD50]

Heres the formula I use:

Assume that I sell one SPY put. I receive $2 in premium and the stock price is 120.
If the put expires worthless I make $200. If I was forced to buy the stock it would cost me $12000 so I divide $200 by the $12000 I was theorectically tying up. So my return is 1.67%. This is similar to buying 100 shares of SPY for $120 for a total of $12000 and then selling it for a $200 profit.

Technically I never have to put up the $12000 because I buy the put back before it expires if its not going to expire worthless and Im not technically tying up a full $12000, but thats the only formula I can come up with for return.

That is the best way to calculate return, IMO.

Here's where I caution option sellers:

Technically I never have to put up the $12000 because I buy the put back before it expires if its not going to expire worthless...

But the underlying can drop after the market close, and you will need to buy it back at a large loss. I've also had many options go underwater, just to right themselves and provide me the full profit right before expiry. If I got out early to protect against further losses, I would have given up a lot of gain (and been negative overall, I think).

Now, many option sellers will say "I'm not taking so much risk, as the underlying won't drop to zero, I never have to put up that kind of money. I can make these 2% profits on a small balance".

OK, but if you allow yourself to get leveraged 2:1, and invest all that, and the market drops 10% before you can cover, that's a 20% drop for you. If you get unlucky, and that happens a couple times you have taken some big hits.

Can't happen? Never say never:


http://www.early-retirement.org/forums/f30/insane-emergency-re-strategy-40682.html#post939013

-ERD50

 

[FIRE'd@51]

Can't happen? Never say never:


http://www.early-retirement.org/forums/f30/insane-emergency-re-strategy-40682.html#post939013

-ERD50

In all fairness to dixonge's strategy, had he been trading SPY options, and been assigned he wouldn't have suffered the big whipsaw loss he did. Remember, he was trading SPX options which are cash-settled at the opening on expiration day (he had failed to roll them out). That abrubtly dropped his beta to zeo, just before the market recovered strongly that day, a recovery he didn't participate in.

 

[REWahoo]

In all fairness to dixonge's strategy, had he been trading SPY options, and been assigned he wouldn't have suffered the big whipsaw loss he did. Remember, he was trading SPX options which are cash-settled at the opening on expiration day (he had failed to roll them out). That abrubtly dropped his beta to zeo, just before the market recovered strongly that day, a recovery he didn't participate in.

Speaking of dixonge: We Are Homeless and Living in our Car | Vagabondians

 

[FIRE'd@51]

Heres the formula I use:

Assume that I sell one SPY put. I receive $2 in premium and the stock price is 120.
If the put expires worthless I make $200. If I was forced to buy the stock it would cost me $12000 so I divide $200 by the $12000 I was theorectically tying up. So my return is 1.67%. This is similar to buying 100 shares of SPY for $120 for a total of $12000 and then selling it for a $200 profit.

I agree with this methodology, with two caveats.

You must credit any dividends paid to your benchmark, as they are embedded in the put premium.

In light of my comment to ERD50 above, when you buy back a put just before expiration, do you immediately roll it out, or do you wait until Monday? If you wait until Monday, you should really stop the clock on your benchmark and restart it Monday when you sell the put to make the comparison as "clean" as possible.

 

[ERD50]

In all fairness to dixonge's strategy, had he been trading SPY options, and been assigned he wouldn't have suffered the big whipsaw loss he did. Remember, he was trading SPX options which are cash-settled at the opening on expiration day (he had failed to roll them out). That abrubtly dropped his beta to zeo, just before the market recovered strongly that day, a recovery he didn't participate in.

What troubled me about dixonge's strategy is, he never seemed to be able to understand what the risks were.

Perhaps if it was clearer to him, at all times, what his max downside risk was, he would have been more careful to monitor those positions and roll them out, or just close them out to limit exposure.

Many years ago, I was doubling down on an investment, dipped into some margin, and continued to double down, figuring it had to turn around sometime (an index ETF, not an individual stock). I got to the point where I could have done some real serious damage to my portfolio, if the investment continued to do as bad as it had (and I was sure it could never go that bad). I got out at a painful, but manageable loss, but at least was not at such extreme risk.

Since then, for every trade I make, I take a deep breath and calculate - what if this trade goes really, really bad - worse than I would ever expect, but yet still possible? I might say - OK, I would be down 10% overall in a really extreme case, I can risk that in this trade. But I wouldn't put 5x that amount at risk, with the possibility (though extremely unlikely) of a 50% hit. I just won't do it.

-ERD50

 

[FIRE'd@51]

Speaking of dixonge: We Are Homeless and Living in our Car | Vagabondians

IIRC, dixonge's first name was Glenn.

 

[ERD50]

I agree with this methodology, with two caveats.

You must credit any dividends paid to your benchmark, as they are embedded in the put premium.

+1

In light of my comment to ERD50 above, when you buy back a put just before expiration, do you immediately roll it out, or do you wait until Monday? If you wait until Monday, you should really stop the clock on your benchmark and restart it Monday when you sell the put to make the comparison as "clean" as possible.

I have two views on this. Yes, it might seem 'cleaner' to match the time frames of each investment against each other (start/stop the clock).

I don't do this, and my thinking is - if there is a period of time that I am not selling options against my position, there is a reason for it. So I consider that 'sideline' time as part of the 'cost of doing business'. If the market goes up while I was out, I may have an "opportunity cost".

Numerical hypothetical to put it in perspective:

SPY goes from 100 to 120 (adjusted for divs) in a year - benchmark is 20% gain.

Let's say there is only one month where I feel conditions are right for selling calls, and I make an added 1% that one month. I would say my strategy out-performed the benchmark by 1% ( 21% versus 20%) for the year. I might want to say I made 12% annualized (or 12.6825030% compounded), as I was only invested in the strategy for 1 month, but the reality was that the startegy didn't allow me to invest all 12 months.

-ERD50

 

[ERD50]

Speaking of dixonge: We Are Homeless and Living in our Car | Vagabondians

IIRC, dixonge's first name was Glenn.

OMG! That's him (same avatar)! I guess he can be tracked by the home page on his contact info here.

Well, he was clearly a "different drummer" type, he may be very happy with that life. I hope so.

-ERD50

 

[FIRE'd@51]

I have two views on this. Yes, it might seem 'cleaner' to match the time frames of each investment against each other (start/stop the clock).

I don't do this, and my thinking is - if there is a period of time that I am not selling options against my position, there is a reason for it. So I consider that 'sideline' time as part of the 'cost of doing business'. If the market goes up while I was out, I may have an "opportunity cost".

Numerical hypothetical to put it in perspective:

SPY goes from 100 to 120 (adjusted for divs) in a year - benchmark is 20% gain.

Let's say there is only one month where I feel conditions are right for selling calls, and I make an added 1% that one month. I would say my strategy out-performed the benchmark by 1% ( 21% versus 20%) for the year. I might want to say I made 12% annualized (or 12.6825030% compounded), as I was only invested in the strategy for 1 month, but the reality was that the startegy didn't allow me to invest all 12 months.

-ERD50

I have no problem with this. A strategy is allowed to have "market timing" in it. I'm just trying to understand precisely what utrecht's strategy is. He seemed to imply that his strategy is mechanical (?). If so, is sitting out some weekends (and exposing oneself to overnight risk relative to the benchmark) part of the strategy, and if so, why? If not, then I think it should be controlled in the benchmark as well, particularly in any back-test.

 

[utrecht]

I agree with this methodology, with two caveats.

You must credit any dividends paid to your benchmark, as they are embedded in the put premium.

In light of my comment to ERD50 above, when you buy back a put just before expiration, do you immediately roll it out, or do you wait until Monday? If you wait until Monday, you should really stop the clock on your benchmark and restart it Monday when you sell the put to make the comparison as "clean" as possible.

I have included dividends in the returns I listed for SPY. Well, actually I think I may have left off dividends when I quoted returns for SPY for the past 14 weeks, but the rest of the returns include dividends.

When I buy back a put, I do immediately sell another one for the next week (or roll it out as you say). Its quicker, easier and I only have to pay one transaction fee. If the put is obviously going to expire worthless than I sell one for the following week as close the end of Friday as possible.

 

[utrecht]

I have no problem with this. A strategy is allowed to have "market timing" in it. I'm just trying to understand precisely what utrecht's strategy is. He seemed to imply that his strategy is mechanical (?). If so, is sitting out some weekends (and exposing oneself to overnight risk relative to the benchmark) part of the strategy, and if so, why? If not, then I think it should be controlled in the benchmark as well, particularly in any back-test.

I'm not sitting out any weekends. If a put is going to expire worthless I open a new trade just before the close on Friday. If its not, then I buy it back and sell another for next week at the same time (or create a spread if you want to look at it that way). The only problem would be if SPY is teetering right at the strike price within 30 minutes of the close on Friday. At that point I have to make a decision. I can buy it back for next to nothing, but that really adds to my transaction costs. If I don't buy it back and I get assigned, then I can just sell my shares of SPY on Monday but that opens me up to risk and reward that I didn't sign up for. Technically since I have no idea whether the market will rise or fall over the weekend, if I did this 1000s of times, I would come out even in the long run (or even profit slightly since the market goes up more than it goes down in the long run), but as I said, thats not the purpose of these trades so I prefer not to have an extra position open over the weekend.

 

[utrecht]

I have no problem with this. A strategy is allowed to have "market timing" in it. I'm just trying to understand precisely what utrecht's strategy is. He seemed to imply that his strategy is mechanical (?). If so, is sitting out some weekends (and exposing oneself to overnight risk relative to the benchmark) part of the strategy, and if so, why? If not, then I think it should be controlled in the benchmark as well, particularly in any back-test.

Can you explain this sentence to me, please? If I was sitting out some weekends, then I wouldn't be trading mechanically. I would be trading based on whether or not I thought that particular week would be a good week for my trading method. Since I'm not doing that, what should "be controlled in the benchmark"?

 

[REWahoo]

IIRC, dixonge's first name was Glenn.

Yep.

Hello! We are Glenn and Dixie, and we are Vagabondians.

 

[utrecht]

Yes, very important. When I traded some 'naked' puts, I preferred to use the term "cash-covered" puts. Maybe your broker lets you cover them with margin, but don't lose sight of the fact that the full amount of money is behind the trade. You could have that money working in other investments, so you should use the full amount of your 'cover' as the denominator in your yield calc.

A friend of mine figured he made an 'infinite' return when he sold an option, then bought it back at a lower price. Selling an option ties up the capital it would take to own the underlying. I don't think I convinced him, because he liked his view better!

-ERD50

I'm using the margin from my buy and hold index funds as collateral for my naked puts. So technically you could say I'm selling puts on margin since I'm increasing my overall risk during times when the market drops and increasing my return when the market rises. I would never use this margin for anything else so there's no opportunity costs involved.

I do agree with you that I need to use the full amount for the sake of calculating a return just so it can be compared to other investments fairly, and that's what I do.

 

[FIRE'd@51]

Can you explain this sentence to me, please? If I was sitting out some weekends, then I wouldn't be trading mechanically. I would be trading based on whether or not I thought that particular week would be a good week for my trading method. Since I'm not doing that, what should "be controlled in the benchmark"?

If you were sitting out some weekends (which you are not) for no apparent reason, then, IMO, an adjustment to the benchmark would be necessary. The way you are implementing your stategy and the way you are calculating the benchmark are entirely consistent. I just wanted to make sure I understood clearly what you meant by mechanical.

I guess the real test of this strategy will be how it performs in a strongly rising market, e.g., Sep 2010 until Feb 2011 (I would investigate this period as fully as possible). I think ERD50's suggestion of using the VIX (for which data is readily available) to estimate put prices is an improvement upon using a constant percentage premium. Of course the VIX is for 30-day options, so it's still not perfect.

Depending upon how your backtest spreadsheet is set up, perhaps you could vary the put premium and find out what premium would make the outperformance of the put writing strategy disappear. Then calculate the implied volatility from this premium.

 

[utrecht]

I would assume that in a strongly rising market, my strategy will under perform but that doesn't bother me for 2 reasons.
#1) The market doesn't rising strongly for long periods of time very often.
#2) I'm using this strategy to trade on margin when I would not use this margin for any other reasons, so even when the market is rising sharply I will still make more money than I would if i wasn't making these trades at all even though the trades themselves will probably under perform the overall SP500.

The SP500 was +26.5% in 2009 which is a pretty strong market. My back test shows that my weekly puts would've returned 36.2%. Because you asked and because you now have me interested, I'll back test Sept 2010-Feb 2011 and report back shortly.

 

[utrecht]

Well that didn't take long

SP500............+19.4% not including dividends
Weekly Puts...+15.1%

The amount of under performance in a strong market really depends on whether its a quick and steady upward market or an explosion upwards in just a few of the weeks.

If the market climbs a steady 1-2% every week for, lets say 5 months, it will be a great 5 months but the weekly puts will keep up. If the market has 2-3 weeks where it shoots up 7%+ at a time, then the weekly puts will under perform, but for every single week that the market does anything other than go up more than 1.5-2%, the weekly puts outperform and there are a lot more of them.

 

[FIRE'd@51]

Well that didn't take long

SP500............+19.4% not including dividends
Weekly Puts...+15.1%

What was the average put premium (percent of underlying) during that period? Did you use a constant percentage?