Key takeaways
- The bullish cross is significant and small. +1.9 points at a 1:2 target on daily charts, on close to ten thousand trades.
- The bearish cross is weaker. Roughly half the edge of the long side and below breakeven at 1:2.
- The 4-hour sample is enormous and the edge is not. Tens of thousands of trades produce well under one point at a 1:1 target.
- This is what a real edge looks like. Not a 70% win rate - a couple of points over a random entry aimed the same way.
What it does
MACD subtracts a slow exponential moving average from a fast one — conventionally the 12- and 26-period EMAs — producing a line that rises when short-term momentum is outpacing long-term momentum. A 9-period EMA of that line becomes the signal line, and the histogram is the gap between them.
The tested signals are the crosses: the bar on which the MACD line rises through its signal line (bullish) and the bar on which it falls through (bearish). This is the reading that appears in nearly every tutorial, which makes it a useful thing to measure honestly.
How this was measured
Every figure below comes from our own tick-derived archive, not from a vendor summary. The method is identical to the one behind our indicator signal study, so the numbers here are directly comparable to every other signal we publish.
- Data. Dukascopy bid-side OHLC for 93 instruments — major and cross FX pairs, gold, silver, stock indices and crypto — from 2010 to August 2026.
- The trade. Enter at the open of the next bar. Stop at 1.5 × ATR(14), targets at 1:1 and 1:2, walking forward up to 40 bars.
- Ties go against you. A bar holding both stop and target is scored a loss every time.
- Real costs. Per-instrument median spreads from our own bid/ask measurement.
- The control. For every signal, a trade with the same direction and the same stop distance at a random bar — five draws each. The gap between the two columns is all the signal is worth.
The result
Breakeven is a 50% win rate at a 1:1 target and 33.3% at 1:2. "Matched random" is a trade in the same direction with the same stop distance, opened at a random bar.
| Signal | Target | Trades | Win rate | Matched random | Edge | Significance | Beats breakeven? |
|---|---|---|---|---|---|---|---|
| MACD bearish cross | 1:1 | 10,093 | 47.8% | 46.9% | +0.9 pts | not significant | No |
| MACD bearish cross | 1:1 | 50,562 | 48.8% | 48.0% | +0.8 pts | ** | No |
| MACD bullish cross | 1:1 | 50,550 | 49.7% | 49.0% | +0.7 pts | ** | No |
| MACD bullish cross | 1:1 | 10,069 | 52.6% | 52.0% | +0.6 pts | not significant | Yes |
| MACD bullish cross | 1:2 | 9,642 | 36.5% | 34.6% | +1.9 pts | *** | Yes |
| MACD bullish cross | 1:2 | 48,616 | 33.2% | 31.8% | +1.4 pts | *** | No |
| MACD bearish cross | 1:2 | 48,946 | 31.4% | 30.4% | +1.0 pts | *** | No |
| MACD bearish cross | 1:2 | 9,668 | 29.9% | 29.1% | +0.9 pts | not significant | No |
MACD is the clearest illustration in this study of why sample size and effect size are different questions. The 4-hour rows carry tens of thousands of trades each, which makes their edges statistically unmistakable — and those edges are under one point. A result can be certain and still be too small to survive a commission.
The daily bullish cross at 1:2 is the exception worth keeping: nearly two points of genuine edge over a random long with the same stop. That is a real finding. It is also roughly the amount a slightly wider spread would erase.
How to use it
- Daily bars, long side, 1:2 target. That is where the measurable edge is. The bearish cross and the faster timeframes are materially weaker.
- Treat it as a tiebreaker. Two points of edge is a reason to prefer one of two otherwise-equal setups, not a reason to take a trade you would not otherwise take.
- Mind the cost ratio. Because the edge is small, the same signal is worth progressively less the more often you trade it. On 4-hour bars you are trading more and earning less per trade.
- Ignore the histogram theatre. We measured the cross. Claims about histogram shape, divergence and slope are untested here and should not inherit the cross's credibility.
What does not work
- MACD as a standalone system. Under two points of edge cannot fund a strategy after costs.
- The bearish cross as the long signal's mirror. It is consistently the weaker half.
- Reading significance as strength. The 4-hour rows are among the most statistically certain in the study and among the least profitable. Large samples make small effects certain, not large.
Limitations
- Default parameters only. Tuning them on this same data would be curve-fitting.
- One stop rule (1.5 × ATR) and two fixed targets. Trailing stops, time exits and scaling are untested.
- Signals are taken in isolation, with no trend filter or confluence — which is how the isolated value of each is measured, but not how most people trade them.
- Bid data with modelled ask. Spread is a measured median, not a replayed tick-level ask feed.
- Spread only — no commission or swap. Both would push results further down.
- A statistically significant edge of one point is still only one point. Large samples make small effects certain; they do not make them large.
Cite or republish this data
Free to quote, screenshot, or republish — in an article, a video, a newsletter, or a course — as long as you credit FxBacktest with a link back to this page.
Frequently asked
Does the MACD cross work?
The bullish cross does, narrowly. On daily charts across 93 instruments it beat a matched random long with the same stop distance by +1.9 points at a 1:2 target, a result significant at the highest level we report. The bearish cross is about half as strong. Both are real effects that are too small to carry a strategy without something else doing the heavy lifting.
Is MACD better than RSI?
No. Measured the same way on the same data, RSI(14) leaving oversold produced +3.0 points at a 1:2 target on daily charts against +1.9 for the MACD bullish cross. MACD has the much larger sample, so its result is more statistically certain, but certainty and size are different things.
What MACD settings should I use?
We tested the default 12/26/9 only. Searching for better parameters on the same data used to evaluate them is curve-fitting and rarely survives contact with new data. The finding that transfers is about direction and timeframe, not the periods: the daily bullish cross carried the edge.
Why is the 4-hour MACD result significant but useless?
Because the 4-hour sample runs to tens of thousands of trades, and statistical significance scales with sample size while the edge does not. A one-point edge measured across fifty thousand trades is unmistakably real and still smaller than the cost of trading it frequently. This is the most common shape in indicator research and the easiest one to misread.
How were these MACD numbers produced?
Enter at the open of the bar after the cross, stop 1.5 x ATR(14) away, targets at 1:1 and 1:2, walking forward up to 40 bars. Bars containing both stop and target are scored as losses. Real per-instrument median spreads are applied. Every signal is compared against five random-entry draws in the same direction with the same stop distance. Data is Dukascopy bid OHLC for 93 instruments, 2010 to August 2026.