MACD Histogram scan

MadRoofer

Does anyone have any advice for a MACD Histogram breaking through the signal line on the positive side. Any help would be greatly appreciated

markd

I haven't heard of the histogram being used in this way. Do you have a chart example?

I re-read the chart school article, and it seems the histogram is used to anticipate crossovers of the two emas - basically, when the bars get shorter. So that is meant to signal a turn in price a little sooner.

The bars and the lines are on the same scale, so you could write the scan you want, but I wonder what it would indicate?

This syntax is correct, but it gets no hits (8/8/2015)

[group is sp500]

and [MACD Signal(12,26,9) > 0.0]

and [MACD Hist(12,26,9) x MACD Signal(12,26,9)]

Kevo

The histogram is used to anticipate the crossover of MACD with it's signal line, not the two EMAs. If you ask me, most of the time this is useless and does not mean much at all. The histogram is best used to anticipate a change in price direction, but it depends a lot on whether the trend is positive or negative and maybe the momentum of the EMAs converging or diverging.

I'm sure it was quickly realized that this clause has a logic error:

and [MACD Hist(12,26,9) x MACD Signal(12,26,9)]

The only signal line for the MACD Histogram is the zero line.

Kevo

I know the Histogram does work well in many cases, but the trend should be taken into account for better setups. I think the MACD Histogram is the best and the most real momentum indicator.

markd

@Kevo , you are correct - I mis-spoke - Hist measures the distance between the Line and the Signal - the two lines displayed on the indicator - not the two emas you could display on the chart.

As for Hist x Signal being a logic error, they are on the same scale, so you can compare them, and if you want to test for one crossing the other, that is how it would be done. Whether it has any meaning or usefulness - that is a different question. Maybe @MadRoofer can tell us what he expects to learn from it, or how he has seen it work.

I don't use MACD at all, so no comment on it's usefulness, except like most indicators, it seem to give its signals late.

MadRoofer

I've only used the MACD Histogram for seeing up turns in price. Or when my slow sto is over 80 is when I keep a good eye on the Histogram.

farjohn

markd,

How do I scan for MACD H below signal line but ticking up.

markd

I'm going to guess you mean below the zero line, rather than the signal line, and the Hist bar getting closer to it.

So first you would test for Hist below zero, MACD Hist(x,y,z) < 0.

And then you would test for today's Hist greater than yesterday's Hist, so MACD Hist(x,y,z) > 1 day ago MACD Hist(x,y,z)

You can select and edit MACD Hist from the "Technical Indicators" drop down on the advanced scan page.

farjohn

Thanks Mark. It worked perfect. My plan is to scan everyday and watch MACD H ticking up and getting closer to 0 line i.e. ready to cross over and based on other indicators make a long decision.
Any suggestions to improve productivity would be appreciated.

markd

My one suggestion would be to do the research before making real trades, or, if you feel you have to trade, trade very small relative to your account size. Test your idea(s) in a wide variety of past markets - e.g. when price is above and below rising and falling MAs like the 50 and the 200 and consider whether the stock is trading in synch or out of synch with the general market.

Also test how it works on big cap, high volume stocks vs. lower volume small and medium caps. Institutionally favored stocks usually trade with fewer surprises - except sometimes around earnings announcements. One other thing to keep in mind is that everybody can see the common indicator signals, so there is sometimes some volatility when a signal occurs - that's the pros shaking out the amateurs to pick up some more stock before the real move begins (if it does). So if you use stops, you would keep them well away from the action, and the wider the stop has to be, the smaller the position size should be in case you do get hit. A recent Gatis Roze blog suggests taking positions incrementally as you are proved right. You can work that out yourself as you are researching.

farjohn

Thanks Mark