MACD Settings Tested: 12-26-9 vs 5-35-5 on Nasdaq Futures

Most traders never question their MACD settings: they load 12-26-9 because it is the default, or switch to 5-35-5 because someone said it is "faster." We put both to a straight test on 16 years of daily Nasdaq-100 futures (NQ), from June 7, 2010 to April 16, 2026, taking every MACD crossover on the long side and the short side. The answer is less flattering to both settings than most tutorials suggest.

What the MACD settings actually mean

The Moving Average Convergence Divergence indicator is commonly credited to Gerald Appel, who developed it in the late 1970s. The three numbers describe three exponential moving averages (EMAs):

• Fast EMA (12 or 5 days) minus slow EMA (26 or 35 days) gives the MACD line.
• Signal line = an EMA of the MACD line itself (9 or 5 days).
• The histogram is the gap between the two. A MACD crossover is the moment that gap changes sign.

MACD 5 35 5 widens the spread between the two averages and shortens the signal line, so it reacts faster and crosses more often. Faster sounds better. On NQ daily bars it was not.

The exact MACD crossover rules we tested

• Instrument: NQ continuous front-month futures, daily bars, June 7, 2010 – April 16, 2026.
• Long entry: MACD line closes above the signal line (bullish crossover). Filled at that day's close.
• Long exit: MACD line closes below the signal line (bearish crossover). Filled at that close.
• Short entry / exit: the mirror image: short on the bearish crossover, cover on the next bullish crossover.
• No stop-loss, no profit target, no trend filter. Pure crossover in, crossover out.
• Costs: 1.0 NQ point per round trip.

All results are in NQ index points per 1 contract. A Micro E-mini Nasdaq-100 (MNQ) contract is worth $2 per point and a full NQ contract $20 per point, so MNQ dollars below are simply points × 2. Every trade has 1.0 point deducted per round trip to cover slippage and commission.

MACD 12 26 9 vs 5 35 5: the results

For context, simply buying NQ on June 7, 2010 and holding to April 16, 2026 gained 24,677.2 points.

The long side: profitable, but a fraction of buy and hold

The classic 12-26-9 long-only crossover made 6,834.5 points over 200 trades, with a 45.0% win rate and a profit factor of 1.32. That is a real edge over zero, and it beat 5-35-5 on every headline metric: total points, win rate, average trade, profit factor and closed-trade drawdown. But it captured only about 27.7% of what buy and hold made over the same span. Because the system is out of the market after every bearish cross, it keeps getting shaken out of a market that mostly went up.

The faster 5-35-5 setting traded 352 times instead of 200, won less often (41.8%), and made less money: 4,412.8 points, a profit factor of 1.14. More signals mostly meant more whipsaw. The average trade shrank from 34.17 points to 12.54 points, which leaves little margin for real-world slippage beyond our 1-point assumption.

Both settings were brutal in the same years. The 12-26-9 long lost -3,189.2 points in 2022 and then made 4,975.5 in 2023. In 2026 through April 16 it is down -2,583.8 points. The 5-35-5 version lost -3,600.5 points in 2022 and -3,117.2 so far in 2026.

A note on drawdown: the figure we report is closed-trade drawdown, the worst peak-to-trough of the equity curve built only from finished trades. It ignores how far an open position sank before it was closed, so for a system that holds for weeks it understates the pain you would actually sit through.

The short side: a consistent loser

Flip the signal and short every bearish MACD crossover and the picture collapses. The 12-26-9 short lost 15,555.0 points with a 23.0% win rate; the 5-35-5 short lost 17,947.0 points. On one MNQ contract that is -$31,110 and -$35,894. The standout exception was the 2022 bear market, when the shorts made 1,485.8 and 1,944.0 points respectively. We dig into why nearly every trend-following short failed on NQ in our piece on shorting Nasdaq futures.

Robustness check: 2010–2019 vs 2020–2026

NQ traded at a small fraction of today's price in 2010, so point totals are dominated by the last six years. Splitting the test in two is the honest check. The 12-26-9 long made 3,174.2 points in 2010–2019 and 3,660.2 in 2020–2026, with a recent-era profit factor of 1.22. Given how much bigger daily point moves became, a similar point total in both halves means the edge per unit of volatility shrank. The 5-35-5 long is weaker still: 2,903.8 points in the first decade, only 1,509.0 in the second, and a 2020–2026 profit factor of 1.06, barely above breakeven.

What this MACD backtest does not model

• Roll gaps. The test runs on a continuous front-month NQ series, so quarterly contract rolls can create small artificial price jumps.
• Perfect fills. Signals are computed on the daily close and filled at that same close. In real life you need to act in the last minutes of the session or accept the next open.
• Overnight and weekend risk is in the price series, but position sizing, margin calls and psychological stamina through deep open drawdowns are not.
• No taxes or financing costs are modeled beyond the 1-point round-trip deduction.

This is a backtest on one instrument. Past results do not guarantee future results, and nothing here is financial advice.

FAQ

Loomi.Ai provides educational market research, not financial advice. Trading futures and options involves substantial risk of loss, and backtested results are historical and do not guarantee future returns.

Sources

• Loomi backtest: blog_strategy_tests.py + results JSON, Loomi research
• MACD, Wikipedia
• Gerald Appel, CMT Association