I have backtested SMAs and EMAs over long-terms like 15 years on Nifty50, though I don't trade them. Here are some the findings from memory: {i} At 30 Days the SMA and EMA give same number of crosses - i.e. difference between them melts away at 30 & above. [ii] Below 30 Days, SMAs give slightly more number of crosses / trades. EMAs give about 1 to 3 trades more trades per year than SMAs. [iii] 18 Day SMA gave the highest return when tested SMAs from 10 to 60 days. [iv] Popular MAs like 20 SMA and 50 SMAs give slightly more crosses / trades than they should be in the series of tests. Reason probably is because many traders take decision based on them and market moves. No harm in staying a little away from popular ones. [v] Over long run, one could have made good money tradings MAs on Nifty50 (as per long-term backtests). [vi] MAs build equity because they (a) keep us out of market when there is a prolonged bear trend, and (b) allow us to some gain also in this period by shorting. [vii] Win percentages are higher when playing longs and lower when playing shorts. In fact the only reason to play short is the one stated in [vi] above. Shorting in bull market does not add value (= does increase in equity)
My vote is for 10,50,200 SMAs. Just because they seem to get mentioned a lot in literature/media and become self-fullfilling prophecy - like the Golden Cross. They are also easier/faster to calculate in auto trading. So I kinda look at them as consensus indicators.
IMO, there is no single "right" direction, and therefore no single way that the MAs "should" be used. I also wouldn't take issue with people who say they're "useless" (for them), if that's based on their own experience, and therefore their own choice not to use them at all. From what I've seen, the *most common* settings appear to be: 1. 8, 9 or 10 period -- either simple or exponential, for the "fastest" MA. Maybe slightly more people appear to lean towards using exponential MAs the faster they get. Based on my own non-rigorous observations, these all tend to display similarly, regardless of exact period or type. 2. 20, 50 & 200 period simple MAs for the "slower" ones. The "most appropriate" MA settings depend on how you choose to use them. If you plan to use them somehow as signals, then it makes sense to experiment and tweak to find specific settings that work for you. If you plan to use them as visual aids, guidelines or filters, then the specific settings are probably not so critical. If you want to see what most other traders are looking at -- e.g., "self-fulfilling prophecy," to perhaps anticipate various types of reactions to those levels -- then it makes sense to to use the "most common" settings. As a couple asides -- and again, depending on how you choose to use them: 1. You may end up wanting to vary your specific settings depending on chart timeframe. 2. If your platform uses pre-market data to calculate intraday MAs, then that can drastically affect where the MAs appear when stocks gap to any degree, at least early in the session. For example, displaying/not displaying the extended session in ThinkorSwim will often make a big difference as to the respective i-day MAs values displayed early in the session, until they eventually "catch up" to each other over time.
McLeary, I am testing a losing EA in Forex that uses this: EUR/USD M20 EXP4 close (no shift) EXP24 close (no shift) The automation trades bar#1 with the above settings and stays with the trade with a Trailing Stop without stopping and reversing each cross. https://www.elitetrader.com/et/threads/four-eas.322919/#post-4689038 ES
All moving averages are low pass filters, meaning low frequencies in the data pass unattenuated, and high frequencies in the data are attenuated by the filter. That implies that you are modeling the market based on an example of a physical system of a signal plus random noise. All models are wrong to one degree or another, and so is the digital signals processing (DSP) model. But that doesn't mean you can't make money with moving averages. The problem of traditional EMAs and SMAs is lag, technically known as group delay. Look at some of the newer low-lag moving averages, specifically the work of John Ehlers.
Based one the 1st screen shot, we see and easy trade setup during the cross from the bottom of the moving averages to take a long with stop below the cross. Target for trend follower would be when the moving averages cross back down and to stay out till they cross up again.
The benefit of trading within a shorter time frame say 5 min, is you can get more signals and trade during an actual day. For me, the best MA that I use is the EMA. Some call MA's a lagging indicator since price has already started to move up or down before the EMA gives a signal. However, that is not how you use them. You use them to help determine trend and then use PA to see if price is respecting them enough to do a trade setup. In a book on PA by a famous author, he explains how to use an MA for actual trading.