Moving average convergence divergence
Compares fast and slow exponential averages to measure momentum.
Purpose and construction
Compares fast and slow exponential averages to measure momentum.
Moving average convergence divergence is applied to the chart series using the local definition below. Each period counts one bar on the active chart. The same period therefore covers a different amount of time on daily and intraday charts.
How to read it
Compare DIF with DEA and the zero line. Histogram expansion shows increasing separation; crossings can whipsaw in ranges.
Confirm the inputs and settings before comparing the indicator with price. On an open bar, close, extremes and volume can still change, so the reading can update before the bar closes.
Parameters
| Parameter | Default |
|---|---|
SHORT | 12 |
LONG | 26 |
MID | 9 |
Parameter names match the variables in the calculation. After changing a lookback or smoothing period, recalculate on the same data to compare response speed. Defaults from another platform may not match this definition.
Original illustrated case
This original daily-bar case moves through an advance, a pullback and a wider-range recovery. It does not represent a real security. Where available, outputs are calculated by the project’s formula engine using the defaults above.
From bar 91 to bar 180, the close rose from 105.1499 to 112.6956. Compare each output below, then use the reading rules above to distinguish direction, momentum, participation or a pattern condition.
| Point | Bar | Close | Volume | Calculated outputs |
|---|---|---|---|---|
| A | 91 | 105.1499 | 14187 | DIF: -0.4169DEA: -0.0237MACD: -0.7863 |
| B | 131 | 104.6513 | 19368 | DIF: 0.1681DEA: 0.5963MACD: -0.8564 |
| C | 180 | 112.6956 | 15024 | DIF: 0.0587DEA: 0.1198MACD: -0.1223 |
At A, DIF is −0.4169 and DEA is −0.0237, so the local doubled histogram is −0.7863. At B, DIF has turned positive (0.1681) but remains below DEA (0.5963), leaving a negative histogram. This separates the zero-line comparison from the signal-line comparison. At C, price is higher while the histogram remains slightly negative (−0.1223): price level and recent momentum are different measurements.
To reproduce: retain the same OHLCV series and default settings, then compare outputs bar by bar. Warm-up bars can have no value. The case-generation script and local definition are stored with the project, so plotted curves are calculated rather than hand-drawn.
Local calculation
The local definition is shown below. Display labels can be translated; the case calculation uses original source identifiers. A shared indicator name does not guarantee identical calculations across platforms.
DIF:EMA(CLOSE,SHORT)-EMA(CLOSE,LONG);
DEA:EMA(DIF,MID);
MACD:(DIF-DEA)*2,COLORSTICK;Limitations
- Compare DIF with DEA and the zero line. Histogram expansion shows increasing separation; crossings can whipsaw in ranges.
- Timeframe, adjustments, warm-up length and data units affect results. Align these before comparing charts.
Questions
Why can Moving average convergence divergence differ across platforms?
Compare inputs, default periods, smoothing, offsets and data units before comparing the final value. This guide uses the local source definition rather than assuming that an external example has the same settings.
Does the example show a real trade?
No. The figure and table use simulated data to explain calculations and chart reading. They contain no real orders and make no return or win-rate claim.
References
The calculation comes from the project’s indicator registry. Primary references below provide related conceptual reading. Figures and numerical cases are independently produced for this project.