Thursday, July 29, 2010

It Was the Best of Times, It Was the Worst of Times

An in depth look at the SPY 50 MA Long/Short strategy in its best and worst years reveals its key fortes and failings. Of the ten years tested, 2008 came out on top while 2000 settled dead last. What was it about 2008 that made the 50 MA such a lucrative signal? What about 2000 made the strategy fail so miserably?

Consider the 2008 SPY chart below with the highlighted crossovers (click to enlarge).
[Source: MachTrader]

The notable performance in the trading strategy for 2008 is a direct result of the trending nature of the market in that time frame. Simply put, moving average trading systems shine in trending markets. If a stock experiences strong follow through after breaking above or below the 50 MA, this approach is a virtual ATM machine. Since the rules dictate one is always long or short the market, it allowed unfettered participation in the lion's share of 2008's sell-off.

Since the 50 MA is most effective in trending markets, reason would tell you it must be quite ineffective in non-trending markets. Might this be the trouble with its performance in 2000? Consider the following chart.
As you can see the first two thirds of 2000 were a veritable chop-fest rife with many a false signal. Though the last few months produced solid profits for a short trade, it wasn't closed until early 2001. Thus, the Achilles heel of this strategy is undoubtedly range-bound market.

Unfortunately it's difficult to consistently forecast whether the market will exhibit trending or range-bound behavior. Thus, the hope with the 50 MA strategy is that the draw downs incurred in the occasional choppy market are overcome by the profits captured once the market starts trending again. That hope certainly bore fruit over the last decade. Whether it remains as potent going forward remains to be seen.


Steve said...

Brian Shannon always points out that its the direction the MA is moving and not an absolute price point.

He claims crossovers indicate indecision and not conviction.

I wonder if you modified your study to take that into account,then what would your results reveal?

Tyler Craig said...

Thanks for the input Steve. If I’m understanding correctly, when Brian says, “it’s the direction the MA is moving, not an absolute price point”, I believe he’s using the MA to confirm the trend of the stock. In other words, a rising 50 MA signals the stock is in an intermediate uptrend. Thus, the slope of the MA is what he’s focusing on, not the specific price it resides at.

Rather than focusing on the slope of the MA, I was focusing on whether it would be effective to use a break of the price above/below the MA as a signal to go long or short.
I’d have to rerun some numbers and tweak some rules if I were to also take into consideration the slope of the MA. Keep in mind, the price leads the moving average, thus before a 50 MA would change from sloping down to up, the price would cross above the 50 MA. Thus, using the price breaking the MA will always generate quicker signals than waiting for the 50 MA to change slopes. Whether or not those quicker signals are more lucrative or just more noisy… that I don’t know.

I would be curious to see if when he says crossovers indicate indecision if he’s referring to moving averages crossing over each other (20 MA crossing 50 MA) or whether he’s talking when the price crosses the MA. I suspect it’s the latter, which is a different approach than what I was using.

2000 was a good example supporting the notion that the price crossing back and forth over the 50 MA indicated indecision. However when crosses over the 50 MA are met with strong follow through, it often times proves a solid signal (as it did in 2008).


I will take the best of times over the worst of times.