What Does a MACD Crossover Mean, and Is It a Reliable Trading Signal?
Direct answer
MACD crossovers, tested
A MACD crossover means short-term momentum has turned relative to the recent trend: Fidelity describes MACD as a 12-period EMA minus a 26-period EMA, with a 9-period EMA of that line as the signal, and a cross above the signal as bullish. As a stand-alone rule it has not been reliable. In our backtest on daily closes from January 2017 to November 2026, holding the S&P 500 only while MACD was above its signal turned $1 into $1.87 (6.6% a year) against $3.26 (12.8%) for buy and hold, over 109 trades with 49.5% winners. On the Nasdaq-100 it lagged too. On bitcoin it roughly matched buy and hold. Fidelity warns MACD whipsaws in ranges.
- MACD = 12-period EMA − 26-period EMA
- signal = 9-period EMA of MACD (Fidelity)
- S&P 500, 2017–2026: crossover rule ×1.87 vs ×3.26 buy and hold
- 109 trades, 49.5% winners, invested 51% of the time
- Nasdaq-100: ×2.29 vs ×5.68
- bitcoin: ×86 vs ×76 before costs
- Whipsaws in ranges
- costs and taxes erode the edge further
MACD is on almost every charting screen, and the crossover is its most-shared signal. It is easy to see on a chart; whether it makes money is a different question, so we tested it.
What people on X are asking
In December 2024 one account explained "HOW MACD WORKS" as a momentum and trend-following indicator. In August 2026 another described crossover strategies as a "SIMPLE SIGNAL, POWERFUL IDEA". And in May 2026 a trader who backtests strategies claimed "The MACD crossover has no edge on $BTC across any of these timeframes". Our own daily test finds something close to that for bitcoin, and worse for stock indices.
What people in the thread said
- In the same thread, @Labrjournal explained the failure: with a win rate of about 33%, the average winner must be at least 2R to break even, but the lagging signal produced winners of only about 1.1R. The maths checks out: 0.67 ÷ 0.33 ≈ 2.03R is the break-even ratio.
- @Labrjournal posted the results: losses on all three bitcoin timeframes tested (15-minute, 1-hour and 4-hour). We did not rerun that test; our daily test above reached a similar conclusion on a different setup.
These are community comments, summarised in our words; the data and sources below come first.
The fact-checked answer
What MACD is. Fidelity's indicator guide says MACD is a momentum oscillator "primarily used to trade trends". It is calculated by subtracting a 26-period exponential moving average (EMA) from a 12-period EMA. A signal line is a 9-period EMA of the MACD line. "When the MACD line crosses from below to above the signal line, the indicator is considered bullish"; the reverse cross is bearish. Fidelity also warns that "During trading ranges the MACD will whipsaw".
UXS chart
S&P 500 MACD and signal line since 2025 (daily, index points)
MACD (12-day EMA − 26-day EMA): latest
-45.67
November 5, 2026
Signal line (9-day EMA of MACD): latest
-17.98
November 5, 2026
Last 12 months
Month-end values (World Bank: monthly averages). Tap a column header to sort.
| 2026-11 | -45.67 | -17.98 |
| 2026-10 | -10.06 | +1.43 |
| 2026-09 | +10.11 | +11.92 |
| 2026-08 | +36.12 | +46.48 |
| 2026-07 | -9.07 | -0.46 |
| 2026-06 | +11.59 | +20.83 |
| 2026-05 | +125.33 | +127.43 |
| 2026-04 | +125.82 | +109.18 |
| 2026-03 | -105.49 | -88.51 |
| 2026-02 | -3.57 | -4.22 |
| 2026-01 | +24.26 | +24.08 |
| 2025-12 | +30.36 | +29.46 |
Our backtest. Rule: be fully invested at the close when MACD is above the signal line, otherwise in cash (earning nothing). Daily closes, no leverage, prices only (dividends excluded on both sides).
MACD crossover rule vs buy and hold (daily closes)
Our calculation from FRED daily closes (SP500, NASDAQ100, CBBTCUSD, CBETHUSD). Index results exclude dividends, so buy and hold is understated. Past results do not predict future ones. Tap a column header to sort.
| S&P 500, Jan 3 2017 – Nov 5 2026 | $1.87 (6.6%/yr) | $3.26 (12.8%/yr) | 109 | 49.5% | Invested 51% of the time; $1.50 after 0.1% cost per side |
| S&P 500, from 2021 | $1.25 | $1.99 | |||
| Nasdaq-100, 2017 – Nov 2026 | $2.29 (8.8%/yr) | $5.68 (19.3%/yr) | 99 | 48.5% | |
| Bitcoin, 2017 – Nov 9 2026 | $86.2 (57.2%/yr) | $75.6 (55.1%/yr) | 130 | 40% | $66.5 after 0.1% cost per side |
| Bitcoin, from 2021 | $2.43 | $2.55 | |||
| Bitcoin, from 2024 | $1.56 | $1.70 | |||
| Ether, 2017 – Nov 2026 | $315 | $260 | $244 after costs |
Reading the results. On stock indices the rule sat out about half the time and gave up roughly half of buy-and-hold's growth, because rallies often began before the lines crossed back. Only about half of trades won, so profits came from a few long trends. On crypto it roughly matched or slightly beat buy and hold over the full period, over the full period. But since 2021 and since 2024 it lagged, and costs ate most of the full-period edge. A rule that wins only in one decade on one market is not a reliable signal.
UXS decision diagram
Should I act on a MACD crossover?
Text version of the diagram
- Is the market trending rather than ranging? If no: Expect whipsaws; Fidelity says users often avoid trading ranges
- Does the crossover agree with the longer trend (e.g. price above its 200-day average)? If no: Treat it as noise
- Is your position size set by a stop, not by the signal? If no: Set risk first
- Have you backtested this market and timeframe including costs? If no: Test before trading it
- Otherwise: Use it as one input, with defined risk; never as a stand-alone buy button
Diagram drawn by UXS. A simplified decision aid for education, not a recommendation.
Why crossovers lag. MACD is built from moving averages of past prices, so by the time the lines cross, part of the move has happened. In fast V-shaped rebounds the rule re-enters late; in choppy markets it flips repeatedly and pays costs each time.
Tuning indicator settings to past data is a classic trap; see why backtested strategies fail live.
Decision helper
| Situation | MACD crossover usefulness |
|---|---|
| Strong trend | Can keep you in the trend |
| Sideways range | Whipsaws; often loses |
| Long-term investing | Lagged buy and hold in our test |
| As a filter with other tools | Most sensible use |
For other indicators we tested, see whether an RSI above 70 means the market will fall and how to use the 200-day moving average. For the maths of mostly 50/50 trades, see win rate vs expectancy.
Common mistakes
- Trading every crossover in a range.
- Judging MACD on one market and one period.
- Ignoring costs on frequent signals.
- Treating the default 12/26/9 settings as optimal.
FAQ
What is a bullish MACD crossover?
When the MACD line crosses above its 9-period signal line.
How is MACD calculated?
The 12-period EMA minus the 26-period EMA; the signal is a 9-period EMA of that.
Is MACD reliable?
Not alone. In our S&P 500 test it lagged buy and hold by about half.
Does MACD work on bitcoin?
It roughly matched buy and hold since 2017 but lagged since 2021 and 2024.
What timeframe is best for MACD?
None reliably; test any timeframe with costs before using it.
Why does MACD give false signals?
It lags price and whipsaws in sideways markets, as Fidelity notes.
Sources
- Fidelity: MACD technical indicator guide
- FRED: S&P 500 (SP500)
- FRED: Nasdaq-100 index (NASDAQ100)
- FRED: Coinbase bitcoin price (CBBTCUSD)
Facts checked November 10, 2026. X posts are cited as evidence of the question, not of the answer.