Best RSI Settings for Gold (XAUUSD): Tested on 2 Years of Data

The short answer

The most common RSI setting for gold is RSI 14 with 30/70 levels, and many scalpers shorten it to 7 or 9 with 20/80 levels. We tested all of them on two years of hourly XAUUSD data with a realistic cost per trade. Every one of the five standard settings lost money when used as a simple “buy when RSI leaves oversold, sell when it leaves overbought” rule.

That does not mean RSI is useless. It means the period you pick matters far less than how you use the indicator. On a trending market like gold, RSI works better as a momentum and context tool, especially for divergence and pullbacks in the direction of the trend, than as an automatic reversal trigger.

What RSI measures

The Relative Strength Index compares the size of recent up-closes with recent down-closes and scales the result from 0 to 100. With a 14-period setting on an hourly chart, it looks at the last 14 hours. Readings above 70 are called overbought and below 30 oversold, but those words only describe speed, not value. A market can be “overbought” and keep rising for a long time.

If you want the full theory, including divergence types, failure swings and range shifts, our lesson on RSI divergence on gold covers it in depth. This page is about settings and how they behaved on real data.

The RSI settings gold traders use

Why gold’s volatility changes the picture

Gold’s daily range has grown a lot. On COMEX futures data, the average daily high-to-low range was about $34 in 2024, $64 in 2025 and about $123 so far in 2026. Over our two-year test window the hourly ATR(14) averaged $17.49.

That matters for RSI in two ways. First, bigger, faster moves push RSI into extreme zones more often, so 30/70 levels fire constantly. Second, when gold trends, it can hold above 70 or below 30 for long stretches. A reversal rule then keeps fading a move that has not finished. Wider levels such as 20/80 and a direction filter are the two most common responses, and we tested both.

We tested 5 RSI settings on 2 years of hourly gold data

For context, simply holding gold gained about $1,194 an ounce in year 1 and only about $14 in year 2, even though year 2 included a run to roughly $5,627 in late January 2026 and a fall back to around $4,000.

Hourly (H1) results

Daily (D1) results since 2015

What the results say, in plain English

• No standard setting worked as a standalone reversal rule. All five lost money on the hourly chart after costs, and four of the five lost in both years. With a target twice the size of the stop, a rule needs to win about a third of its trades just to break even before costs. These settings won 28% to 33%, and gold’s trends kept them on the wrong side.
• Fewer signals lost less. RSI 14 with 20/80 took 65 trades instead of 330, and its loss was about an eighth of the size. Wider levels did not create an edge; they just kept the rule out of more counter-trend trades.
• The direction filter mattered more than the period. Only taking longs above the 200 EMA and shorts below it, with 40/60 levels used as a pullback trigger, brought the result to roughly break-even (+$35 over 163 trades). That is not good enough to trade, but it points to the real problem: fading a trending market.
• On the daily chart, shorts did the damage. For RSI 14 with 30/70, short trades lost about $1,239 an ounce while long trades made about $186. In the 2021–26 bull market, “sell overbought” kept selling a rising market. In the quieter 2015–20 period, the same rules were roughly flat.

The honest conclusion: if you are searching for the setting that makes RSI reversals profitable on gold, our data says it is not there.

How to use RSI on different timeframes

Scalping (M1 to M5). RSI 7 or 9 with 20/80 levels gives readings fast enough for short holds. At this speed the spread is a much bigger share of each move than in our hourly test, so treat RSI as a timing aid inside a level you already chose, not as the reason to trade. Check what a move is worth at your size with the gold pip calculator.

Intraday (M15 to H1). RSI 14 is fine. Use the range-shift idea: in an uptrend, RSI tends to hold between about 40 and 80, and dips to the 40–50 zone are pullbacks, not sell signals. Use RSI to confirm that momentum is returning in the trend direction, after price has reached a level.

Swing (H4 to D1). RSI 14 or 21. Our daily results show how costly it is to sell overbought readings in a bull market. On higher timeframes RSI is most useful for spotting divergence near major levels and for noticing when momentum is fading after a long run.

A combination that makes more sense

Combine RSI with structure and a trend filter rather than using it alone:

1. Decide the trend on a higher timeframe, for example with the 200 EMA or market structure.
2. Wait for price to pull back into a level you marked in advance, such as prior support.
3. Use RSI to time the turn: in an uptrend, RSI dipping toward 40 and turning back up.
4. Put the stop beyond the level, not at a fixed RSI value, and size it with the position size calculator.

This is the logic behind the filtered row in our table. It is also how the setups in the RSI divergence lesson are built: divergence plus a break in structure, not divergence alone. For a second momentum view, compare RSI with MACD on gold, and see how a similar oscillator behaved in our stochastic settings test.

Common mistakes with RSI on gold

• Selling just because RSI is above 70. In our daily test, this was where most of the losses came from.
• Changing the period after every losing trade. Our table shows the period did not fix the core problem; only the context did.
• Ignoring costs on fast settings. RSI 9 made 574 trades in two years. At $0.30 an ounce each, costs alone came to about $172 an ounce.
• Using the same levels in every market. 30/70 behaves very differently in a range than in a trend.
• Treating an indicator reading as a stop level. Stops belong beyond price structure, sized to current volatility. Our gold volatility tool shows today’s ATR.

Test it yourself

You do not need to take our numbers on trust. Pick one RSI rule, write it down precisely (entry, stop, target, exit), and test it on your own platform’s gold data:

1. Use at least a year of data, and check that the sample covers both trending and ranging months.
2. Include your real spread and commission per trade.
3. Split the period in two and compare the halves, as we did. A rule that only works in one half is fragile.
4. Record trades, win rate, average result per trade and the worst drawdown, not only the final profit.

Our lessons on backtesting a gold strategy and backtesting with free data walk through the process, and the gold indicator settings hub shows how 11 other indicators did under the same test.

Test it with other traders

Numbers are more useful when other people try to break them. Inside the XAUUSD Global community traders test settings like these on their own charts, share their journals and compare results across sessions. The first month is free, then $19.99 a month. Bring your RSI rule and see whether it survives other people’s data.

Educational content only, not financial advice. Trading gold with leverage carries a high risk of loss.

Frequently asked questions