What moving averages are
A moving average (MA) is the average closing price over a set number of candles, recalculated every candle. It smooths out noise so you can see the general direction.
There are two types you’ll use most:
- Simple moving average (SMA): adds the last N closes and divides by N. Every candle has equal weight. A 20 SMA on H1 is the average of the last 20 hourly closes.
- Exponential moving average (EMA): gives more weight to recent candles, so it reacts faster to new moves. It never fully forgets old data, but the influence of older candles fades quickly.
Why traders use them:
- Trend direction. Is price above or below a rising or falling average?
- Trend filter. Only take longs above a key average, only shorts below it.
- Dynamic support and resistance. In a trend, pullbacks often pause near a well-watched average.
- Trend strength. The slope and spacing of several averages show whether a trend is healthy or fading.
A moving average is built only from past prices. It doesn’t predict anything. Its value is in summarizing what price has already done, in a way you can turn into simple rules.
How it behaves on gold (XAUUSD)
Gold has traits that change how averages behave:
- Fast, deep pullbacks. Gold often dips through a short EMA during a healthy trend, takes stops, and recovers. Using a single line as an exact entry or stop leads to frustration.
- News spikes distort averages. A $30 candle after US CPI or NFP (released at 8:30 New York time) drags short averages sharply. For a while afterward, the fast EMAs reflect one event more than the trend.
- Long trends on higher timeframes. Gold has historically had long periods of trending behavior on daily charts, where the 50 and 200 averages have provided useful context. That’s a description of the past, not a guarantee.
- Choppy Asian sessions. During quiet hours (roughly 20:00–02:00 New York time, 00:00–06:00 GMT in winter), price often flattens around the fast EMAs, which flip direction repeatedly.
- Watched daily averages. The daily 50 and 200 are quoted in financial media. Because so many people watch them, reactions near those levels are common, though never guaranteed.
How to calculate and set them up
You don’t need to calculate averages by hand, but knowing the idea helps.
- SMA: add the last N closes and divide by N. Example with 5 hypothetical closes: 4,120 + 4,124 + 4,118 + 4,130 + 4,128 = 20,620; ÷ 5 = 4,124.
- EMA: start from a prior EMA, then add a fraction of the difference between the new close and that EMA. The fraction (the smoothing factor) is 2 ÷ (N + 1). For a 21 EMA, it is 2 ÷ 22 ≈ 0.091. So each new close moves the EMA about 9% of the way toward it.
- Add your averages to the chart. Use closing prices unless you have a specific reason to change it.
- Use different colors and keep them the same on every chart, so you can read them instantly.
- Check the slope, not just the position. A flat average with price crossing back and forth means no trend.
The common periods and what they’re for
| Average | Typical role on gold | Notes |
|---|---|---|
| 9 EMA | Very short-term momentum | Popular with scalpers on M1–M5; flips often |
| 21 EMA | Short-term trend and pullback guide | A common “first pullback” area in trends |
| 50 EMA | Intermediate trend | Often used as a trend filter on H1 and H4 |
| 200 EMA / SMA | Long-term trend | The daily 200 is a classic bull/bear dividing line |
SMA vs EMA at a glance
| Feature | SMA | EMA |
|---|---|---|
| Weighting | Equal | Recent candles weighted more |
| Speed | Slower | Faster |
| False signals in chop | Fewer, but later | More, but earlier |
| Typical use | Long-term filters (daily 200) | Short-term trading and pullbacks |
Variations and settings: from beginner to pro
Beginner approach
- Use just two averages: the 50 EMA and the 200 EMA on H1 or H4.
- Only look for longs when price is above both and both are rising. Only look for shorts in the opposite case.
- Ignore crossovers as entry signals. Use the averages to decide which direction to look.
Intermediate refinements
- Add the 21 EMA for pullback timing inside the trend.
- Use a higher timeframe for direction and a lower one for entry. Example: H4 50 EMA for bias, H1 21 EMA for pullbacks. See multi-timeframe analysis.
- Treat the area between the 21 and 50 EMAs as a zone rather than using either line as an exact price.
Pro refinements
- MA ribbons. A set of averages (for example 8, 13, 21, 34, 55) shows trend health. When the ribbon is fanned out and ordered, the trend is strong. When it twists and compresses, the market is ranging or turning.
- Distance from the average. When gold stretches far from the 21 or 50 EMA, measured in ATR multiples, chasing becomes riskier and mean reversion toward the average becomes more likely.
- Slope thresholds. Some traders require the 50 EMA to have risen over the last N candles before taking longs, which filters out flat, choppy periods.
- Anchored context. Pros often pair moving averages with session-based tools like VWAP, which resets each day and shows the average price traded.
Crossovers: golden cross, death cross and their lag
A crossover happens when a faster average crosses a slower one.
- Golden cross: the 50-day crosses above the 200-day.
- Death cross: the 50-day crosses below the 200-day.
These events get headlines, but they have a built-in problem: lag. A 50-day average needs weeks of price movement before it turns, and the 200-day needs even longer. By the time they cross, price may already have moved a long way, or even started to reverse.
Shorter crossovers, such as the 9 EMA crossing the 21 EMA, react faster but produce many more false signals in ranges. This is the core trade-off of every moving average: speed versus reliability. There’s no setting that gives you both.
Crossovers are best viewed as a regime label: “the market has been trending down for a while”. They’re a weak timing tool. If you use them at all, use them as a filter and get your entries from structure.
3 setups traders use
Setup 1: Pullback to the 21 EMA in a trend
| Element | Rule |
|---|---|
| Context | H1 price above a rising 50 EMA; higher highs and higher lows on market structure |
| Trigger | Price pulls back into the 21 EMA area and an M15 candle closes back up and away from it |
| Entry | On the confirmation close |
| Stop (invalidation) | Below the pullback low, plus a buffer of a fraction of ATR |
| Targets | The prior swing high first, then a Fibonacci extension |
| Management | Take partial at the prior high; trail the rest under new higher lows, not under the EMA line itself |
Setup 2: 200 EMA trend filter with a structure entry
| Element | Rule |
|---|---|
| Context | H4 price above the 200 EMA (bullish regime) |
| Trigger | Price sweeps a minor H1 low into support and reclaims it |
| Entry | After the reclaim candle closes |
| Stop (invalidation) | Below the sweep low |
| Targets | The next resistance zone |
| Management | Skip the trade entirely if H4 closes below the 200 EMA before entry |
Setup 3: 50 EMA retest after a trend change
| Element | Rule |
|---|---|
| Context | Price was below a falling 50 EMA, then breaks above it with a structure break |
| Trigger | The first pullback to the 50 EMA holds, and the 50 EMA has started to flatten or turn up |
| Entry | On a bullish close after the retest |
| Stop (invalidation) | Below the retest low |
| Targets | The most recent lower high on the higher timeframe, then trend continuation |
| Management | Be ready for failure; first retests after a trend change are less reliable than pullbacks in an established trend |
Worked example (hypothetical)
Hypothetical H1 example. Gold is in an uptrend. Price is above a rising 50 EMA at about 4,118. The 21 EMA sits at about 4,140. During the New York morning, price dips to 4,138.5, slightly through the 21 EMA, and an M15 candle closes back up at 4,144.
The trader enters long at 4,144. The stop goes below the pullback low and a buffer: 4,134. Risk per ounce: $10. The prior swing high is 4,162, and a 127.2% extension is near 4,172.
Position size. Account $5,000, risk 1% = $50. With 0.01 lot = $1 per $1 move, size = 50 ÷ 10 = 5, so 0.05 lots ($50 risk).
Targets. First target 4,162 (reward $18, 1.8R). Second target 4,172 (2.8R). Half comes off at 4,162; the rest is trailed under the next H1 higher low.
What would invalidate it? An H1 close below the 50 EMA and the most recent higher low. Then the pullback is no longer a pullback.
This is an illustration of the process, not a trade idea.
Moving averages give a false sense of precision. A line on the chart looks exact, but gold can trade $5–10 through an average and still be “respecting” it, or ignore it completely during news. Always size from a stop based on structure, and use the position size calculator before entering.
Combining it with other concepts
- Market structure. Averages say “probably trending”; structure tells you where the trend’s real turning points are. See market structure.
- Support and resistance. A pullback to the 21 or 50 EMA that also lands on a prior level is a stronger location. See support and resistance.
- Fibonacci. In strong gold trends, the 21 EMA often sits near the 38.2–50% retracement of the latest leg. See Fibonacci on XAUUSD.
- MACD. MACD is built from the 12 and 26 EMAs, so it’s essentially a measurement of the gap between two averages. See MACD on gold.
- RSI. Use RSI to judge whether a pullback to an EMA is losing momentum. See RSI divergence.
- Volatility. Bollinger Bands use a 20-period SMA as the middle band, so a band walk and an EMA pullback often describe the same trend.
Best timeframes and sessions
| Timeframe | Common averages | Use |
|---|---|---|
| Daily | 50 and 200 SMA/EMA | Big-picture regime; context for swing traders |
| H4 | 50 and 200 EMA | Trend filter for swing and day trades |
| H1 | 21 and 50 EMA | Pullback map for day traders |
| M5–M15 | 9 and 21 EMA | Entry timing and short-term momentum |
Averages work best in active, trending sessions. London (roughly 03:00–11:30 New York time) and the London/New York overlap (roughly 08:00–11:30 New York time) are when gold trends most often on intraday charts. In late-Asian or pre-holiday trading, price usually drifts around the fast EMAs and produces noise. See gold market hours.
Common mistakes
- Trading every crossover. Fix: use crossovers as a filter, not an entry trigger.
- Using too many averages. Fix: pick two or three and keep them for at least 50 trades before judging.
- Placing stops exactly at the average. Fix: stops go beyond structure plus a volatility buffer.
- Using trend tools in a range. Fix: if the 50 EMA is flat and price crosses it repeatedly, switch to range tactics or stay out.
- Changing settings after a loss. Fix: no period is “the one gold respects”. Test, then commit.
- Buying far above the average. Fix: measure the distance in ATR; if price is extended, wait for a pullback.
Limitations and honest caveats
- All moving averages lag. They are built from the past. Faster settings lag less but produce more false signals.
- No magic numbers. There’s no evidence that gold “respects” a particular period more than others. Popular averages may get more attention, which is the only real reason some settings are favored.
- Whipsaws in ranges. Sideways markets are where average-based systems tend to lose money, and gold spends plenty of time ranging.
- Backtest overfitting. It’s easy to find the perfect average for last year’s data. That rarely holds up. See backtesting a gold strategy.
- No verified win rates. Claims that “the 200 EMA bounce wins X% of the time” are rarely backed by data you can check.
Pro tips
- Read the slope before the cross. A 50 EMA that has gone flat often tells you a trend is tiring earlier than a crossover does.
- Use the first pullback, skip the fourth. Pullbacks to an average early in a trend tend to be cleaner than late ones, after the trend has stretched.
- Measure stretch in ATR. If gold is more than about 2–3 ATR away from the 21 EMA on your timeframe, chasing usually offers poor risk-to-reward.
- Align with the higher timeframe. An H1 21 EMA pullback in the direction of the H4 50 EMA is better than a counter-trend one.
- Let the daily 200 define your default bias. Many swing traders simply stand aside from counter-trend trades against it.
- Watch how price closes, not where it wicks. A wick through the EMA and a close back above it is very different from two closes below it.
Quick checklist
- Which way are the 50 and 200 EMAs sloping on my bias timeframe?
- Is price above or below them?
- Is the market trending or flat and choppy?
- Is this an early pullback or a late, stretched one?
- Is there structural confluence (level, Fib, order block) at the average?
- Did I get a confirmation close, not just a touch?
- Is my stop beyond structure, not at the line?
- Did I size from the stop distance?
- Is high-impact news due soon?
Frequently asked questions
Which moving average is best for gold?
There’s no single best one. Many gold traders use the 21 EMA for short-term pullbacks, the 50 EMA for the intermediate trend, and the 200 EMA or SMA for the big picture. Choose a small set, test it on your timeframe, and stop switching after every losing trade.
Is EMA or SMA better for XAUUSD?
An EMA reacts faster because it weights recent prices more, so it suits shorter-term trading on a fast market like gold. An SMA is smoother and slower, which some traders prefer for long-term trend filters such as the 200-day. Neither is better in every case.
What is a golden cross on gold?
A golden cross is when a shorter moving average, traditionally the 50-day, crosses above a longer one, traditionally the 200-day. A death cross is the opposite. Both are widely reported, but they lag price significantly and often appear after much of the move has happened.
Do moving averages act as support and resistance?
They can appear to, especially well-known ones like the 50 and 200 on the daily chart, because many traders watch them. They’re dynamic areas, not exact lines. Gold often trades a few dollars through an average before reacting, if it reacts at all.
What moving averages do scalpers use on gold?
Scalpers often use fast EMAs such as the 9 and 21 on M1 to M5 charts to judge short-term direction. In choppy conditions these averages flip constantly, so most scalpers combine them with structure, session timing and a spread check.
Why do moving average crossovers fail so often?
Because they need price to move a lot before they react. In ranges, price oscillates around the averages and produces repeated false crosses, often called whipsaws. Crossovers tend to look best on trending historical charts and worst in live sideways markets.
Sources and further reading
- CME Group — Gold futures overview — Contract specifications and trading hours for COMEX gold
- CFTC — Commitments of Traders — Weekly futures positioning, including gold
- Investor.gov — Stop order — How stop orders work, and their limits in fast markets
External sites are independent of XAUUSD Global. Check dates and data yourself.