MACD, What It Does and What It Doesn't

[Fairman]

MACD, What It Does and What It Doesn't

The Moving Average Convergence Divergence (MACD) is a popular indicator built from two moving averages. It displays the MACD line, a signal line, and a histogram showing the gap between them.

diagram.png (35.01 KiB) Viewed 68 times

What it can show:

- Crossovers: when the MACD line crosses above the signal line, momentum may be turning up. The reverse suggests turning down.
- Zero line: MACD above zero suggests bullish momentum, and below zero suggests bearish momentum.
- Divergence: price making a new extreme while MACD fails to, hinting at fading momentum.
- Histogram changes: shrinking bars suggest slowing momentum.

Limitations:

- It's a lagging indicator, built from past prices
- It produces many false signals in sideways markets
- Crossovers often occur after a large part of the move has happened

How to use it sensibly:

1. Use it as a filter in trending conditions, for example taking only buys when MACD is above zero.
2. Combine it with structure and key levels.
3. Don't rely on crossovers alone.
4. Test its performance on your chosen pair and timeframe.

If you find yourself waiting for MACD to confirm every trade, you may be entering late.

Indicators can add context, but price and structure should lead. Keep your toolbox small and your understanding deep.

[LondonNewsTrader]

On short timeframes MACD has a specific weakness around scheduled data. A release candle can be ten times the size of the bars before it, so both averages get dragged sharply, and the crossover that prints a few bars later is mostly an echo of that one spike rather than fresh momentum. Traders who take crossovers mechanically often end up buying the top of the post-release move.

If you use it as the zero-line filter you describe, consider discarding signals for a set number of bars after high-impact releases, or at least checking whether one big candle within the last 26 bars is driving the reading. With default settings a single candle can distort the slow average for quite a while.

[FXS Prop Desk]

A clear summary of MACD. The histogram is often the most useful part for timing, because shrinking bars show momentum fading before the lines actually cross. That gives an earlier warning, at the cost of more false signals.

Divergence works best when it appears at a key level, for example a previous day high, rather than in the middle of a range. And it's worth remembering that the default 12/26/9 settings were designed for daily charts. On lower timeframes some traders shorten them, but the meaning of "momentum" changes with the timeframe, so testing matters.

[Shadow Trader]

LondonNewsTrader's spike point is right, although with MACD the candle never fully leaves the window. Both lines are exponential averages, so there is no clean 26 bar cutoff where the distortion ends.

On the slow average each new bar carries a weight of about 0.074. After 26 more bars roughly 13 percent of the release candle's influence is still sitting in the reading, which on M1 means the echo lasts close to half an hour.

That is why my own rule is based on time rather than bars. On M1 and M5 I ignore MACD completely for 30 minutes after any red folder release, and only look again once the bars are back near their usual size.

The histogram does warn earlier, but on M1 it shrinks and grows constantly inside a range. I only pay attention to a fading histogram when price is already pressing a level I marked before the session.

As a filter it earns its place in one narrow job. During London I take M5 longs only with MACD above zero, which has removed more bad countertrend fades than crossovers ever added.

Has anyone tested faster settings such as 6, 13 and 5 on M1? I tried them for a few weeks and got more signals, not better ones.