There’s a lot of chatter right now about the upcoming “Golden Cross” in the stock market. A golden cross occurs when the 50-day moving average crosses over the 200-day moving average, and to some technicians signals the start of a bullish bias in the market. Read our own take on the golden cross.
But there are two equally popular ways of calling the famed crossover, using either a simple moving average (SMA) or exponential moving average (EMA). In this post I want to put both to the test to see which has performed best historically.
[logarithmically-scaled]
The graph above shows the results of two competing strategies, one going long at today’s close when the 50-day SMA crosses over the 200-day SMA today (red), and the other when the 50-day EMA crosses over the 200-day EMA today (green), versus buy & hold (blue), on the S&P 500 from 1960 to the present.
Geek notes: these results are frictionless (i.e. do not account for transaction costs or slippage). As I usually do when testing longer-term strategies like this one, I’ve included a return on cash when not in the market of half the nearest 13-week Treasury bill. Note that in my previous test of MA crossovers I assumed the full treasury yield (instead of half), so these results will be just a bit different.
And for the number lovers…
There hasn’t been a significant difference historically between the two flavors of the golden cross. On the surface, the SMA variation has been a bit more effective, but most of the difference came on a single trade in the 1970’s (note how since then, the two have tracked each other near perfectly).
In short, I don’t think it makes sense to get caught up in an SMA vs EMA debate – both do the job they’re assigned about equally well. I think investors are much better served (as we talk about on this blog quite a bit) adding other timeframes to their trading (short and intermediate-term) to create a more holistic view of the markets
Having said all of that, if I could pick just one, I would probably go with the SMA – it’s performed a bit better historically and is just plain simpler to calculate.
[Edit: click for a summary of all posts in this series on trading the Golden Cross]
Happy Trading,
ms
P.S. one last geek note: these are two generally-accepted ways to calculate an EMA. Here I’ve used the (2/(1+Period)) rather than the (2/Period) method. But the difference in performance between the two has been extremely close (and not really worth showing explicitly) because of how long the moving averages being tested are.
To stay up to date with what’s happening at the MarketSci Blog, we recommend subscribing to our RSS Feed or Email Feed.
Filed under: Trading Strategies | 8 Comments
-
Subscribe to RSS Feed
Subscribe via Email
My email: michael @ marketsci.com (remove spaces)
Excellent study. Thanks for showing what really works.
Damien
My recollection is that the classic definition also requires both averages to be pointing in the same direction when they cross.
RE to sysin3: I’ve heard some folks use that before too (assuming we both mean for a long trade, both averages would have to be moving up, and vice-versa for a short trade). I’ve run the numbers in the past on that, but didn’t post about it – I’ll add it to the to do list for the near future. Thanks, michael
RE to sysin3: response to your comment posted at: http://marketsci.wordpress.com/2009/06/20/testing-the-rare-downtrending-golden-cross/. Thanks for moving the collective discussion fwd. michael
Can I put in a word for the 13/34 ema crossover on the weekly chart. Only three signals since 2000, each one of them sound :
October 2000 sell
May 2003 buy
Dec 2007 sell
I enjoy your studies
Thanks