4

The standard setting for the MACD are 12, 26 and 9. Is there a rationale why these seemingly arbitrary numbers are considered a standard?

My other question is that why is 26 considered to be the long term EMA for MACD while 26 in a general sense is short term or at best medium term.

Improve this question
0

2 Answers 2

Reset to default
4

The values of 12, 26 and 9 are the typical industry standard setting used with the MACD, however other values can be substituted depending on your trading style and goals.

The 26d EMA is considered the long moving average when in this case it is compared to the shorter 12d EMA. If you used a 5d EMA and a 10d EMA then the 10d EMA would be considered the long MA. It is based on what you are comparing it with.

Apart from providing signals for a reversal in trend, MACD can also be used as an early indication to a possible end to a trend. What you look out for is divergence between the price and the MACD. See chart below of an example:

CBA - Divergence

Here I have used 10d & 3d EMAs and 1 for the signal (as I did not want the signal to show up). I am simply using the MACD as a momentum indicator - which work by providing higher highs in the MACD with higher highs in price. This shows that the momentum in the trend is good so the trend should continue. However the last high in price is not met with a higher high in the MACD. The green lines demonstrate bearish divergence between price and the MACD, which is an indication that the momentum of the trend is slowing down. This could provide forewarning that the trend may be about to end and to take caution - i.e. not a good time to be buying this stock or if you already own it you may want to tighten up your stop loss.

Improve this answer
9
  • After sustained price move, if security trades flat, MACD will reverse and head toward zero. So turn down from above zero may be a false signa (caused by old data removal from the MA calculation), and vice versa for upward MACD. Turn down signal below zero cannot be false While MACD tests well in trending periods, over long haul it underperforms basic simple moving average crossover systems. You can look at research by Colby & Meyers or you can test this yourself. Indicators like this which are based on moving averages lag the market (in late, out late) with whipsaws in non trending markets.
    – Bob Baerker
    Commented Mar 27, 2020 at 12:01
  • I am not looking at the MACD, in this case, to take signals off, I am using it as a momentum indicator for a possible warning that a trend might be slowing down - given by a new high in the price but a lower high in the MACD momentum indicator, i.e. divergence in the slope of the 2 curves.
    – Victor
    Commented Mar 29, 2020 at 7:27
  • Any indicator involving moving averages is subject to a drop off effect when old data is removed as a new data is included (it's all in the math of each MA). This means that the signal you see today has been caused by what happened up to 26 periods ago (or whatever indicator time length you are using). That not only creates false MACD signals but false divergences as well. Honest chart appraisal with show false convergences as well. However, if this works for you, go for it.
    – Bob Baerker
    Commented Mar 29, 2020 at 11:44
  • Firstly, I have used a combination of the 10 day and 3 day EMA's of the MACD, secondly, the newest price data has new highs but the indicator has lower high. Over the following weeks, the price dropped 13%, then regained its uptrend. And thirdly, as I'm repeating 3 times now, the indicator is just used as an early warning single (because the price has not fallen yet), it is not used to take direct action against - it is just saying look at me and be cautious over the next few days/weeks, don't buy me right now if you haven't got me yet, and if you already have me tighten your stop loss.
    – Victor
    Commented Mar 29, 2020 at 23:19
  • Have you even ever used a momentum indicator in actual trading before?
    – Victor
    Commented Mar 29, 2020 at 23:21
2

In the good old days / the days of yore / back in the day, etc., the standard working week was 6-days long (the cruel taskmasters), therefore, 12, 26, and 9 represent 2 weeks, 1 round(month), and 1.5 weeks. Maybe 5, 10, 30 or 35 would be better suited for today. Or not. Have a nice day. Never mind.

Improve this answer

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .