Thursday, November 23, 2006

Are Moving Averages A Useful Trading Guide?

Is the market trend up, down, or neutral? One of the most common ways to assess a trend is the moving average. But is the moving average a useful guide for trading decisions? I decided to take a look.

Since 2004 (N = 710 trading days), we've had 328 days in which the S&P 500 Index (SPY) has closed above both its five-day and twenty-day moving averages. Over the next ten days, SPY has averaged a loss of -.03% (182 up, 146 down).

Conversely, when SPY has closed below both its five-day and twenty-day moving averages (N = 200), the next ten days have averaged a gain of .72% (126 up, 74 down).

Indeed, just knowing where the S&P 500 Index has been trading relative to its 20-day moving average has been informative.

When we've closed above that moving average (N = 441), the next 20 days in SPY have averaged a gain of only .20% (261 up, 180 down). That's about a third of the average 20-day gain from 2004 to the present.

On the other hand, when we've closed below the 20 day MA (N = 269), the next 20 days in SPY have averaged an impressive gain of 1.24% (188 up, 81 down).

In short, knowing where we're trading relative to moving averages *has* provided us with useful trading information, but not in the way many technicians expect. Because of mean reversion, returns when we're trading above moving averages have been subnormal. Returns when we've been trading below moving averages have been superior.

The trend has indeed been your friend in the last few years in the large cap stock indices, but only if you've faded it.


John Wheatcroft said...

Thanks for yet another informative article -

Some time ago I switched to using the exponential moving averages in my work - they "seemed" to make a closer fit - one day while playing around in Sharpcharts I put a 90 period EMA under a weekly view of the SPX, then the INDU, then the COMPQ, then RUT and finally, Goldman Sachs (a stock I consider extremely important as a bellweather). It is simply amazing how the 90 period EMA serves these stocks.

I don't understand why given all of the computational power we now possess people still use SMA 200. Other than "tradition" - "way my daddy did it and it was good enough ..."

I think not only is the EMA 90 more accurate it is far more informative.

Anonymous said...


how do you find out how many down/up days have their been in the market? Which software provides such in-depth historical data?


Anonymous said...


In the Market Expectations section you said "The Russell 2000 small caps and the midcap stocks, however, could not surmount their overnight highs. That led me to short the Russell, for a retracement back toward its mean price. When the most volatile stocks--small caps, mid caps, semis--don't confirm a rally, it generally pays to fade the strength."

What are the average daily intra-day highs and lows of Russell 2000?

thanks a lot.

Brett Steenbarger, Ph.D. said...

Hi Yaser,

I conduct my analyses entirely within Excel using data I've archived from my real-time data source (Townsend RealTick). The intraday moves in the Russell correlate quite highly with those in the other equity indices, which makes divergences--such as the ones that occurred when we made new highs in the morning on Wednesday--unusual and of interest.


Brett Steenbarger, Ph.D. said...

Hi John,

Thanks for your note. There is a lot of very traditional--and uncritical--thinking among traders and investors. Moving averages crossover systems of various durations have tested out *very* poorly in the stock index market the last several years. I haven't looked at the difference between EMA and SMA; thanks for the idea--


Capt said...

When I checked data on individual stocks which I own, I found that an upward trend was an extremely good predictor that the following period would also be up. Likewise (but less so) with downward trends. The market as a whole, however, will move up and down in a way that doesn't relate to any particular stock.

So, unless you are buying index funds, why do you care about the market trend? It doesn't predict anything useful and, in my experience, it doesn't follow trends in the same way as individual stocks.

Brett Steenbarger, Ph.D. said...

Hi Capt,

Yes, I trade index futures/ETFs. In those instruments, we actually see short-term reversals of trending moves, especially when those moves lack strong momentum. I agree with you: many individual stocks do not follow that pattern. It's important to know the tendencies of the market you trade, not just *the* market. Thanks for your note--