Short answer: a moving average crossover is a signal that fires when a faster moving average crosses a slower one, or when price crosses a single average. The fast line tracks recent closes and the slow line tracks a longer window, so a cross means the recent average has overtaken the older one. Traders use it as a plain, rules-based way to define the trend and time entries. It lags by design, it whipsaws in ranges, and its value depends on the pair of lengths and the filters around it.
How a crossover system works
A crossover system needs a fast average and a slow average. The fast one is short, say 10 or 20 periods, so it hugs price and turns quickly. The slow one is long, say 50 or 100, so it moves grudgingly and smooths the noise. Fast above slow calls the trend up. Fast below slow calls it down. Nothing else is required, which is the attractiveness of the method and also its weak spot.
Price crossing a single average is the fastest and most jittery version. Two averages crossing each other is the ordinary dual system that most trading strategies mean by "crossover". The 50 and 200 daily pairing, known as the golden cross and death cross, is one dual system among many, and it says scarcely anything about intraday timing.
Choosing the average type: simple, exponential or weighted
The type of average changes how punctual the cross is. A simple moving average gives every close in the window the same weight, so it turns late and calmly. An exponential moving average tilts the weight toward recent closes and turns sooner. A weighted moving average uses straight-line weights and behaves much like an exponential one.
In technical analysis the type matters less than the lengths. Exponential lines cross a bar or two earlier than simple lines of the same length, and the rest of the behaviour is nearly identical. Our comparison of exponential and simple averages covers the maths. My blunt rule: an exponential fast line and a simple slow line, because a stable slow line produces fewer needlessly repeated recrosses.
Picking the pair of lengths: a decision framework
The gap between the two lengths sets the character of the system. A ratio near 1 to 3, such as 10 and 30 or 20 and 50, is a sensible place to begin. Squeeze the ratio and you get constant whipsaw. Stretch it and the signals become rare and stale.
| Pair (fast / slow) | Typical chart | Frequency of crosses | Suits |
|---|---|---|---|
| 5 / 20 | 5-minute to 15-minute | Many per day | Scalpers who accept many small losses |
| 9 / 21 | 15-minute to 1-hour | Several per week | Day traders on indices and major pairs |
| 20 / 50 | 1-hour to 4-hour | A few per month | Swing traders on forex, gold and BTC |
| 50 / 200 | Daily | A few per year | Position traders and trend filters |
Three questions settle the choice:
- How long do you want to hold? Match the slow line to roughly the length of the swing you want to catch.
- How much whipsaw can you absorb? If ten small losses in a row would rattle you, lengthen both lines.
- Does the market trend? Major pairs spend long stretches in ranges, in my experience, while gold and crypto move more violently and reward faster pairs.
Reading bullish and bearish crossovers
A bullish crossover is the fast line rising through the slow line. A bearish crossover is the mirror image. The cross itself is a coarse event, so read two details around it. First, the slope of the slow line: a bullish cross while the slow line still falls is brittle. Second, the angle: lines that cut through each other steeply tend to hold, while lines that creep across while price is wobbly tend to recross.
The MACD indicator is built from two exponential averages, so its zero line cross and a 12/26 crossover are the same event. Use its histogram instead, which shows the gap between the averages closing before the lines meet.
Whipsaw control: filters that cut false signals
Whipsaw is the thankless side of any crossover: the lines cross, you enter, price reverses, the lines recross, you exit at a loss, and it repeats. Each of those trades also pays the spread, and that cost is why fast pairs that look fine on paper disappoint in live trading. Filters that help, in the order I would try them:
- Close-based confirmation. Act only after the bar closes with the cross intact, never on an intrabar touch.
- A distance buffer. Require the fast line to clear the slow one by a set amount, for example 10 pips on a major pair.
- A higher-timeframe filter. Take only crosses that agree with the trend one chart up. In my experience this roughly halves the signal count.
Most trading platforms let you set an alert for one line crossing another, which spares you the tedious job of watching every bar.
Testing a crossover before you trade it
A crossover is easy to test because the rules are mechanical. Suppose the account is 10,000 and each trade risks 1 percent, which is 100. On a 4-hour EUR/USD chart, a 20/50 exponential pairing enters on the cross with a stop 50 pips away, beyond the recent swing. With 100 at risk and a 50-pip stop, position size is 2 per pip, or 0.2 standard lots. Now walk back through the last 12 months of crosses by hand. Log the entry, the exit, the spread paid, and the longest string of consecutive losses. That last number, not the average result, decides whether you can live with the system.
Results vary by market and by year, and a pairing that looks generous on one instrument can be stingy on the next. The wider logic of holding winners and cutting losers sits in our guide to trend following.
Limitations and common mistakes
- Lag is structural. Every average is built from past closes, so in a short, sharp move the cross can fire near the end.
- Ranges are unforgiving. No pair of lengths fixes a market going nowhere. Filters reduce the damage, they do not remove it.
- Optimising lengths until the past looks tidy. A 13/47 pairing that beat 20/50 last year is curve fitting, not insight.
- Stops placed on the slow line. Averages are not support. Put the stop beyond a swing point.
- Ignoring the exit. A system whose only exit is the recross hands back a lumpy share of every trend.
Running a crossover under evaluation limits
Picture a 9/21 pairing on a 15-minute index chart, a 4 percent daily loss limit, and 1 percent risked per cross. A ranging morning can produce four crosses before lunch, and four losses hit the limit. The awkward truth is that fast pairs and daily loss limits sit poorly together, so lengthen the pair or cap attempts per session.
Under a Pivex evaluation the total drawdown never trails and the daily loss allowance renews each day. The damage from a whipsaw cluster is therefore knowable in advance. Pivex sets no time limit and no minimum count of trading days, which suits a slow pairing that may stay flat for weeks. Personal algorithms are allowed too, so a coded crossover with the filters built in is fine.
Consider the horizon first. A stock investor with a long term plan may only need a monthly check of a slow SMA pair, while a short term CFD trader has to understand that the same MA pair will tell them to buy and sell far more often, and that each signal carries a cost.
Next step: pick one instrument, one timeframe and one pair of lengths, then log every cross from the last 12 months by hand before you risk anything. That short, honest audit is worth more than any general rule. If you run the system under Pivex rules, let the built-in assistant track how close a whipsaw run has brought you to the daily limit. When it gets close, lengthen the pair rather than fight the limit.
Related guides: MACD indicator and trend following.
Before participating, review Pivex's current program rules and decide whether a simulated proprietary trading evaluation fits your experience and risk limits. Participation does not guarantee a reward, income, or any particular trading result.