The idea
Two moving averages of the closing price, one fast and one slow, are drawn on the chart. The fast one, often the 9-period exponential moving average (EMA), reacts quickly; the slow one, often the 21 EMA, reacts slowly. When the fast line crosses above the slow line, recent prices have risen above the longer average: the trend has probably turned up. When it crosses below, the reverse.
The same idea runs at every scale. Day traders use 9 and 21 on five-minute charts; swing traders use 20 and 50 on daily charts; the famous golden cross is the 50-day average crossing the 200-day.
Watch it play out
Step by step
The rules
Setup
- Plot a 9 EMA and a 21 EMA of closing prices.
- Prefer markets that are moving: a crossover in a flat, tight range is the weak case.
Entry
- Buy at the close of the candle where the 9 EMA closes above the 21 EMA.
- Sell short at the close of the candle where the 9 EMA closes below the 21 EMA.
Stop-loss
- For a buy, just under the lowest low of the last five candles (the recent swing low). For a short, just over the highest high.
Target
- Either a fixed target of 2R (twice the distance to the stop),
- or hold until the averages cross back the other way, which keeps you in long trends but gives back part of each move.
Filters that help
- Only take buys when price is also above a slower average (such as the 200 EMA), and shorts when below.
- Skip crosses where the two averages are nearly flat.
Why it can work
Prices tend to move in runs more often than pure chance would produce, because news is absorbed slowly and traders pile in after a move has started. A crossover is a mechanical way to be on board for most of a run without guessing where it starts.
The price of that simplicity is lag. A moving average only turns after price has, so the signal always arrives after the turn, and in a sideways market the lines cross back and forth, each cross a small loss.
Worked example
On a five-minute MNQ chart, the 9 EMA closes above the 21 EMA and you buy at that candle's close. The lowest low of the last five candles is 28 points lower, so the stop goes 30 points under the entry: 30 × $2 = $60 of risk per MNQ contract.
A 2R target sits 60 points above the entry. With $100 risked per trade, one contract fits ($60), two do not ($120).
When it works, when it fails
Works best
In markets that trend for long stretches: crypto in strong runs, indices after a big turn, currencies driven by interest rate gaps. The longer the moves, the more each crossover pays compared with the small losses in between.
Struggles
In ranges. When price goes sideways the averages tangle and cross every few candles, and each cross is a small loss plus costs. Most crossover systems spend a lot of their time in this state, which is why their win rate is usually under 50%.
Common mistakes
- Testing dozens of average lengths and keeping the best one. That finds what fitted the past, not what will work next: the data-snooping problem.
- Taking every cross in a range instead of filtering for a trend.
- Expecting a high win rate. Trend following usually wins less than half its trades and makes its money on a few large ones.
Variations
- Golden cross
- The 50-day simple average crossing the 200-day: a long-term signal for stocks and indices, with very few trades a year.
- Triple average
- A third, slower average as a trend filter: trade the 9/21 cross only in the direction of the 50.
What the evidence says
Evidence: Tested, mixedPublic tests disagree, or the edge has faded over time.
Moving average rules are among the most studied in finance. A well-known 1992 study of the Dow Jones index from 1897 to 1986 found that buy signals from simple moving average rules were followed by higher returns than sell signals.
A 1999 follow-up that corrected for data-snooping found the best of those rules did not keep working in the years after the original study. The honest summary: the signal had value historically, it has weakened in heavily traded markets, and results today depend on the market, the costs and the filter.
Questions
- What is the best moving average crossover?
- There is no single best pair. 9 and 21 EMA are popular for intraday charts, 20 and 50 for swing trading, 50 and 200 for long-term signals. Pick one, test it on your market with costs, and keep it rather than switching after every loss.
- EMA or SMA for a crossover?
- An EMA weights recent prices more, so it turns sooner and gives earlier but noisier signals. An SMA is smoother and later. Neither is better in general; the difference matters less than filtering out ranges.
- Why do I keep losing on crossovers?
- Usually because the market is ranging. Crossovers lose small amounts repeatedly in sideways markets. A trend filter and patience through losing streaks are part of the method.
- What is a golden cross?
- The 50-day moving average crossing above the 200-day. It is the crossover idea on a long timeframe, read as a sign that a longer uptrend may be starting.