SPX Historical Spread Plays

tman · Nov 7, 2010

I forgot that there aren't 5 point strikes in the weeklys. 25 point strikes won't provide anywhere near a 12.5 credit.

I wouldn't want to trade this strategy on spy because of the risk of early exercise when it goes ex-div and it's itm as you previously mentioned.

Wayne Gibbous · Nov 7, 2010

This might be encouraging for you backtest freaks - how about DJ-30 from 1915? Yeah the data might be a little suspect and there were no options, but the up months still hit near 60%...

Third Friday to third Friday changes (at Close) for DJ-30

Start 19150319 Price: 56.90

19150416, 69.40, %Chg: 21.97

19150521, 64.90, %Chg: -6.48

19150618, 70.90, %Chg: 9.24

19150716, 71.80, %Chg: 1.27

19150820, 79.30, %Chg: 10.45

19150917, 82.50, %Chg: 4.04

19151015, 92.60, %Chg: 12.24

19151119, 95.30, %Chg: 2.92

19151217, 97.40, %Chg: 2.20

... Lot of years ...

20100115, 10,609.65, %Chg: 2.72

20100219, 10,402.35, %Chg: -1.95

20100319, 10,741.98, %Chg: 3.26

20100416, 11,018.66, %Chg: 2.58

20100521, 10,193.39, %Chg: -7.49

20100618, 10,450.64, %Chg: 2.52

20100716, 10,097.90, %Chg: -3.38

20100820, 10,213.62, %Chg: 1.15

20100917, 10,607.85, %Chg: 3.86

20101015, 11,062.78, %Chg: 4.29

TOTAL MONTHS = 1145
DOWN MONTHS = 467
UP MONTHS = 678

Wayne Gibbous · Nov 7, 2010

And SPX from 1962. You guessed it - up months still hit 60%...

Third Friday to third Friday changes (at Close) for SP-500

Start 19620316 Price: 70.96

19620419, 68.59, %Chg: -3.34

19620518, 63.82, %Chg: -6.95

19620615, 55.89, %Chg: -12.43

19620720, 56.81, %Chg: 1.65

19620817, 59.01, %Chg: 3.87

19620921, 57.69, %Chg: -2.24

19621019, 55.59, %Chg: -3.64

19621116, 60.16, %Chg: 8.22

19621221, 62.64, %Chg: 4.12

...

20100115, 1,136.03, %Chg: 3.04

20100219, 1,109.17, %Chg: -2.36

20100319, 1,159.90, %Chg: 4.57

20100416, 1,192.13, %Chg: 2.78

20100521, 1,087.69, %Chg: -8.76

20100618, 1,117.51, %Chg: 2.74

20100716, 1,064.88, %Chg: -4.71

20100820, 1,071.69, %Chg: 0.64

20100917, 1,125.59, %Chg: 5.03

20101015, 1,176.19, %Chg: 4.50

TOTAL MONTHS = 581
DOWN MONTHS = 228
UP MONTHS = 353

optionsmaven · Nov 7, 2010

Hi Wayne,

It is pretty well known that the long term trend has been a 60% rise in shares, measured weekly, monthly, yearly and since the beginning of markets here in the U.S. That's beyond dispute and is essentially a reflection of the long term trend of the country's economic growth. It would however, be very interesting to know whether that percentage has been the case for monthly cycles in those decades when the indices showed a decline, as has been the case in the last ten years. This period has been described as 'the lost decade', heralding the 'death of long term investing' and other cataclysmic pronouncements. Whether or not that's true is open to debate, but IMNSHO options traders are more interested in how to profit in shorter term trends. Will that percentage hold in future, even as in has in the gyrating market since the term of the century? Only time will tell, of course, but I'm betting that it will.

Wayne Gibbous · Nov 7, 2010

Quote from optionsmaven:

Hi Wayne,

It is pretty well known that the long term trend has been a 60% rise in shares, measured weekly, monthly, yearly and since the beginning of markets here in the U.S. That's beyond dispute and is essentially a reflection of the long term trend of the country's economic growth. It would however, be very interesting to know whether that percentage has been the case for monthly cycles in those decades when the indices showed a decline, as has been the case in the last ten years. This period has been described as 'the lost decade', heralding the 'death of long term investing' and other cataclysmic pronouncements. Whether or not that's true is open to debate, but IMNSHO options traders are more interested in how to profit in shorter term trends. Will that percentage hold in future, even as in has in the gyrating market since the term of the century? Only time will tell, of course, but I'm betting that it will.
More...

Sure seems like the powers that be will do whatever it takes to keep the US heading upwards.

The same test only shows close to 50% up/down months over 20-30 years of commodity futures like oil and gold. Kind of what you expect. This is a game for US stock indices...

optionsmaven · Nov 7, 2010

Well put Wayne, particularly since the solution to the country's massive debt problem appears to be debasement of the currency, paying back dollars with quarters, as it were. That means growth in share prices is inevitable in nominal terms, if not in buying power.

spindr0 · Nov 7, 2010

Quote from JJacksET4:

Interesting information there - I honestly wouldn't have guessed that the market had been up in these time frames more often then it has been down despite the overall drop in the market.

I think that lends support to the premise that the market drops faster than it rises... and in a shorter period of time, further.

What I would be curious of and would be interested in backtesting would be if there was a way to improve this especially for the really bad months - I think that trying to always close early and not having to pay the full $5.00 to close isn't quite good enough because of potential large drops and the fact that if you close a bit early, the market might roar back near the very end and would have saved you.

Backtesting has the significant advantage of hindsight and there's zero guarantee that it will help going forward. Tangentially, backtest any of the common canned cookie cutter technical analysis indicators (commodity channel index, stochastic, RSI, etc.) and you'll find a periodicity that provided astounding results. If you apply that result to another long segment of the same data, you'll often see large losses. The reason is that the market can be repetitive but not necessarilty in a coherent order or duration.

My thought would be to add extra puts to the protection side and in the more extreme months (maybe the 5%, but especially the 10%+ months), maybe you could actually profit on the downside. I am more used to using SPY options, so an example for me would be like if you were doing this today:

Sell 10 SPY Dec 123 Puts
Buy 15 SPY Dec 118 Puts

or maybe something like:

Sell 10 SPY Dec 123 Puts
Buy 10 SPY Dec 118 Puts
Buy 5 SPY Dec 116 Puts

or whatever - this would obviously bring in somewhat less money each good month, but could help out quit a bit in the few down 10%+ months.

Earlier in the chain I suggested that it might be an idea to add extra protective puts in the Martingale months when position size is being bumped up to recover losses because that's when a prolonged losing streak can really hammer you. At no time would I use a 15:10 ratio from the outset because it would be self defeating vis a vis the 60+ pct win ratio of the data - eg. several months of channelling would then be a loser.

If hedging, I'd favor a smaller ratio of long to short legs (6:5, 11:10 etc.). The further OTM kicker is another possibility and its cost might be offset by selling further OTM puts. IOW, an initial bullish put spread protected by an OTM bearish OTM put spread. But this is personal preference nd gets away from the big picture.

Of course, this would actually be worse if you waited till the very end and the SPY ended right at the strike of the extra puts you bought (meaning they are all worthless, but you still owe for the shorts). You would probably want to do something like "If SPY has fallen >5% and at least 1 week remains, close the position" or whatever - I haven't tested this obviously, so you would want to go back and review previous months and when to consider closing maybe.

Improving a system like this is very subjective. A bigger picture view would be that modifications/adjustments that you do have as much potential to add more profit/lose less or lose more and that the ultimate results will be dependent on the monthly market cycle you're in as well as fortuitous timing in your adjustments and money management.

More...

spindr0 · Nov 7, 2010

Quote from HowardCohodas:

Backtesting has significant limitations whether you use tick data, or settlement data. Depending on getting the fills in real trading as you did in backtesting frequently leads to disappointment.

Furthermore, backtesting has a tendency to lead to curve fitting unless proper precautions are taken. Dividing the model into two sets, not contiguous, is a good way to avoid this. The model is developed on one data set and then verified by viewing it's performance on data it has never seen.
More...

All very true unless your backtesting program offers the Fudge Factor indicator.

spindr0 · Nov 7, 2010

Quote from optionsmaven:

...the facts are that over a decade in which the SPX had a net 15% loss of value, nearly two thirds of the monthly cycles were up. That's an anomaly that is hard to ignore. Further, my two years of actually trading the system profitably have convinced me of its viabilty. Whilst 'systems work until they don't' may well be true, I'm sticking with this one until it doesn't.
More...

I don't think it's an anomaly. I think is merely a function of long periods of grinding up slowly and shorter periods of collapsing quickly. That's the basis of the high win rate of this pattern. A 60/40 win rate is a really good ratio over the long term/

As for the viability, I'm a firm believer in that everything that you find that works will eventually stop working (temporarily or permanently) and that you trade it gradually more aggressively the longer it has worked for you. The last trade may bite you but by then, c'est la vie.

An example might be a scalping a stock that's channeling in a range, bouncing back and forth b/t support and resistance. You have no idea at the outset that it's going to do that but if a stock like that finds you :eek: then you keep banging out the trades until it misbehaves. If trading the underlying, then hope the stop is triggered by a penetration rather than a gap. If trading spreads, then be glad you have the protective leg.

optionsmaven · Nov 7, 2010

Well SpindrO,

I think it is an anomaly. As I have said ad nauseum, if anyone can show me another decade when the market was down 15% and the monthly cycles were up over 60% of the time, I'll profusely apologize. As for the cycle the strategy doesn't work, you only lose 2.50/3.00, so BFD.