Search "MACD strategy" and you will find two different trades wearing the same name. The MACD line is just EMA(12) minus EMA(26) — so the moment it crosses the zero line is, mathematically, the exact moment the 12-period EMA crosses the 26-period EMA. The other trigger — the MACD line crossing its own 9-period signal line — fires earlier and far more often, and it is a genuinely different system. Most articles quoting a "MACD win rate" never say which of the two they tested.
We tested the one that can be pinned down honestly today: the zero-line cross, run as EMA(12)/EMA(26) crosses on 1-hour BTC and ETH candles across 12 months of real Binance data — long when the fast EMA closes above the slow, short on the inverse, 1×ATR(14) stop, 2R target, 1% risk, costs on. The signal-line cross is not yet a primitive in our engine, so its results are queued rather than borrowed or estimated. The exact rules, the numbers and the distinction most articles skip are below.
Verified Result
Weak: the mechanical version barely cleared, or failed to clear, the fee hurdle.
| Market | TF | Trades | Win | PF | Max DD | Net |
|---|---|---|---|---|---|---|
| ETH | 1h | 266 | 32.3% | 0.79 | 42.3% | -35.2% |
| BTC | 1h | 293 | 29.7% | 0.62 | 61.6% | -61.3% |
How the SVS 25 breaks down ▾
12 months of real 1-minute data, fees on (0.05%/side), $10k start, 1% risk. How the score works →
The exact rules we tested
- Compute EMA(12) and EMA(26) on 1-hour candles; the MACD line is their difference — EMA(12) minus EMA(26).
- Enter long when the MACD line crosses above zero — by construction, the exact candle on which EMA(12) closes above EMA(26).
- Enter short when the MACD line crosses below zero (EMA(12) closes below EMA(26)); both directions, no session filter (crypto trades 24/7).
- Stop 1×ATR(14) from entry; target 2R.
- Risk 1% of equity per trade; 0.05% commission per side; 10× max notional leverage.
- Not tested here: the signal-line cross (the MACD line crossing its 9-EMA) — a different, faster trigger that is not yet a supported primitive in the engine.
Results
Binance spot 1-minute klines (data-api.binance.vision), aggregated per strategy timeframe · starting balance $10,000 · risk 1%/trade · Commission 0.05% per side; no spread/slippage modeled (BTC/ETH spot spreads are sub-basis-point); position size capped at 10× notional leverage. Generated 2026-06-12 by the Secuora ai-strategy deterministic runner (same engine as the in-app AI backtester).
| Month | Trades | Win rate | Net P&L |
|---|---|---|---|
| 2025-06 | 24 | 33% | −$614 |
| 2025-07 | 26 | 31% | −$729 |
| 2025-08 | 25 | 36% | −$468 |
| 2025-09 | 23 | 30% | −$728 |
| 2025-10 | 18 | 22% | −$675 |
| 2025-11 | 27 | 22% | −$835 |
| 2025-12 | 30 | 27% | −$661 |
| 2026-01 | 19 | 32% | −$267 |
| 2026-02 | 17 | 24% | −$356 |
| 2026-03 | 34 | 26% | −$532 |
| 2026-04 | 30 | 43% | $135 |
| 2026-05 | 20 | 25% | −$401 |
| Month | Trades | Win rate | Net P&L |
|---|---|---|---|
| 2025-06 | 24 | 29% | −$621 |
| 2025-07 | 27 | 30% | −$583 |
| 2025-08 | 21 | 43% | $327 |
| 2025-09 | 24 | 29% | −$615 |
| 2025-10 | 19 | 37% | −$20 |
| 2025-11 | 27 | 33% | −$254 |
| 2025-12 | 21 | 19% | −$911 |
| 2026-01 | 13 | 38% | −$2 |
| 2026-02 | 22 | 27% | −$446 |
| 2026-03 | 22 | 36% | −$33 |
| 2026-04 | 25 | 32% | −$319 |
| 2026-05 | 21 | 38% | −$45 |
Assumptions (how loose terms were pinned down)
- The MACD line is EMA(12)−EMA(26), so the MACD ZERO-LINE cross is mathematically identical to the EMA 12/26 cross — that is what this run tests
- The signal-line (9-EMA of MACD) cross is a different trigger and is not yet a supported primitive
- Stop 1×ATR(14); 2R target; 24/7
One indicator, two strategies — and why published results never agree
The MACD zero-line cross and the EMA 12/26 cross are not similar — they are the same trade, entry for entry, because the MACD line is positive exactly when EMA(12) sits above EMA(26). That equivalence is worth internalizing: anyone selling "MACD" as something deeper than a moving-average cross at the zero line is selling packaging. The signal-line cross is the genuinely distinct variant — a 9-EMA smoothing of the MACD line that flags momentum turning before the regime actually flips, so it trades more often, earlier, and with more noise.
Now count the degrees of freedom in a typical "MACD strategy" article: which cross (zero or signal line), which settings (12/26/9 or something "optimized"), which timeframe, which exits, costs on or off. Five unstated choices produce wildly different stats from the same indicator — which is why quoted MACD win rates range from coin-flip to implausible. Ours are reproducible because every choice is pinned above; when you compare them to another article, the first question is always: which cross did they test?
How to backtest the MACD cross on Secuora
Both versions of the trade are testable today — one automated, one by hand:
- Open the free replay demo at /backtest/demo (no sign-up) and add MACD to the chart — it ships built in, alongside EMA, RSI, Bollinger, VWAP, ATR and Stochastic.
- Set the timeframe to 1h to mirror this test (the demo replays real BTC data) and step through the chart bar by bar.
- To trade every zero-line cross with simulated stop-loss and take-profit orders — 1×ATR stop, 2R target — sign up free and run the sessions in the replay terminal; a meaningful 30+ trade sample needs its longer history anyway.
- To automate it, open the AI backtester at /backtest/ai and describe the rule in plain English: the EMA 12/26 cross is a supported primitive, and it IS the zero-line cross.
- Re-run with one change at a time — timeframe, stop multiple, settings — and compare expectancy and drawdown, not just win rate, before believing any variant.
Methodology, in one paragraph
Data: Binance spot 1-minute klines (data-api.binance.vision), aggregated per strategy timeframe, June 1, 2025 – June 1, 2026 (12 months). Execution: Secuora’s deterministic strategy runner (the same engine behind the in-app AI backtester) — single position at a time, entries at the close of the signal candle, commission 0.05% per side; no spread/slippage modeled (btc/eth spot spreads are sub-basis-point); position size capped at 10× notional leverage, starting balance $10,000, 1% risk per trade. Swings are confirmed fractals with no look-ahead. These are mechanical results: no discretion, every signal taken. Past performance does not predict future results; this is research, not financial advice.
Frequently asked questions
Is the MACD cross the same as an EMA cross?
The zero-line cross is — exactly, not approximately. The MACD line is EMA(12) minus EMA(26), so it is positive precisely when EMA(12) is above EMA(26), and the zero-line cross produces an identical trade list to the EMA 12/26 cross. The signal-line cross (the MACD line crossing its 9-EMA) is a genuinely different, faster trigger.
What are the best MACD settings?
12/26/9 is the default, and it is the setting almost every published result actually uses. Changing the numbers changes regime sensitivity — faster settings give more signals and more chop losses, slower settings enter later in bigger trends — so there is no universally "best" combination. Test a change against the default on the same data rather than optimizing blindly; settings that win one regime routinely lose the next.
What win rate does the MACD cross strategy have?
It depends on which cross, the timeframe and the exits — which is why quoted numbers disagree so much. For the version pinned on this page (zero-line cross on 1-hour BTC/ETH, 1×ATR stop, 2R target, costs on), the measured win rate is published in the results table above. Crossover entries with 2R targets typically run low win rates and live off occasional large trends — and in this 12-month run the hard 2R cap kept the winners from paying for the losers, so expectancy finished negative on both symbols. Judge expectancy and drawdown, not the win rate alone.
Why don’t you publish signal-line cross results?
Because our engine does not ship a signal-line-cross primitive yet, and we only publish numbers our own deterministic engine produced on real data. The zero-line results can be published because that cross is mathematically the EMA 12/26 cross, which is a supported primitive. The signal-line variant lands here the moment its primitive ships.