does anyone here use the difference of two SMA periods as a gauge of momentum? For example, the difference in ticks of the classic 20SMA and 50SMA for trend direction and momentum?
As durations get smaller, in my backtesting, the parameters for MACD should increase for intraday trading or have tons of false signals. And entries should wait for retracements, the crossing of %K and D show a possible trend change but often at extreme of recent trend change and often a retracement occurs.
I have similar experiences with MACD. I don't use it intensively and if I do use it then only for longer time horizons (days or weeks).
Yes, I like it on weeks mainly whether commodities or stocks, but use it as way to confirm extremes. It does do a nice job for intraday like one minute bars, but have to double values of what is standard, as it allows to show beginning of trend but not right there to enter, best to wait for 50% pullback, then risk is cheaper.
I spent literally years in search of a way to measure momentum that was compatible with both my skill level and my psychological makeup, unable to find anything I could sink my teeth into. Lindq’s suggestion seemed quite logical to me, but whenever I tried something along those lines, it failed when I attempted to put it into actual practice—once again, given my skill level and the way I perceive things. However, this morning I completed an initial exercise executed for the purpose of converting what I regard as a successful system for day trading foreign currency pairs into written text that might later be coded into MetaQuotes Language (MQL4), and in the process, I finally discovered what I searched for so long!!! Indeed, the technique did come down to the difference of two moving average periods. However, even though I have probably heard many, most, or possibly even all of the arguments against using moving averages that are not the classic/standard fare, the success of my system is based on doing just that! So rather than evaluating the difference in ticks of SMA (20) and SMA (50) for trend direction and momentum, I compare the slopes of two alternative, painstakingly selected, simple moving averages. When the difference is extreme, I know the momentum is unsustainable and I wait and watch for a reversal. When the slopes of both moving averages remain relatively equidistant from one another within a carefully defined set of parameters, I know I am looking at a trending asset for which momentum is of significant strength. Though I recently explored a pseudo-swing style of trading, I am for the most part a day trader due to my determination to maintain an extremely high percentage of successful trades, my personal goal of walking away from every single day with a profit, and my desire to avoid, if at all possible, being faced with any kind of drawdown whatsoever. Consequently, I have discovered and apply these measurements on one-minute charts exclusively.
The tricky part is this statement: This reeks of data fitting. And, while it may work for one instrument, it might be completely incorrect for other instruments.
It works for the major foreign currency pairs, which is all I trade, so it is all I care about. I make adjustments when trading other assets, such as the major U.S. indices, but Forex is what I like best for a number of reasons that need not be mentioned, so this is where I focus my time and attention. As a matter of fact, I can guarantee that it WILL be incorrect for other instruments. This does not concern me because I use my system the way it was designed to be used. I did not try to come up with some generic approach to be applied across a wide spectrum of options. Rather, my desire is to implement something that is highly effective in a particular setting to maximize efficiency and the resulting profits. I'm perfectly happy to let others apply whatever systems they deem best in as many different markets as they choose. But when I compare the outcomes I am able to realize with my system with the results others are willing to share in public, I am quite content to stick with my "data fitting" methodology.
Hi Expiated, Thank you for your reply, and your insights to a fruitful discussion. Like you said, I agree that it is near impossible to apply the same settings or analysis to different markets as each market is unique in its own, but rather the same general idea can be applied but must be adjusted. If I understand you correctly in your posts, it is the slope of the two different moving averages that you found value. And whatever the two moving averages you are using, I assume one is short term and the other is long term. Otherwise, what's the value of comparing any other combination? However, what is the difference from measuring slope of the two vs. measuring the difference in ticks? In time based charts, slope is (y2-y1)/(X2-X1) where it is essentially the change of price vs. time. I see value in measuring the slope is it that in the end of the trend, the short term slope would be flat or inverse of the longer term slope before the two crosses over. Otherwise, in the middle of the trend, both should be upward sloping (assuming rising trend). The other, more theoretical, question is that it is hard to take meaning into the difference in ticks. For example, as you mentioned, let's say the difference in ticks between a 50 and a 20 is 6ticks. What does that number "6" ticks actually mean? Whereas in slope, it has a more intuitive explanation.