RSI Trend Rules vs Overbought and Oversold: The Trend Side Won 12 of 16 Match-Ups on Gold and 2 of 16 on EURUSD

Beginners usually learn RSI one of two ways. One teacher says 70 is overbought and 30 is oversold, so you sell the first and buy the second. Another says that in a trend RSI stays high, so selling 70 is how you get run over, and you should use RSI to join the move instead. Both can be written as exact rules, so I ran both on the same bars with the same costs.

The trend rules and how each of them did on its own are in the RSI trend catcher test. This article is the head-to-head.

How was the comparison set up?

• Data: EURUSD, June 3 to August 26, 2026, 16,926 M5 bars and 5,642 M15 bars; gold over the same weeks, 13,736 M5 bars and 4,577 M15 bars.
• Fade: RSI 14 crosses back under 70: sell. Crosses back over 30: buy. No trend filter, as it is usually taught.
• Trend: RSI 14 crosses 50; the same with the 200 EMA; pullback to 40 (60 for shorts) on the 200 EMA side; break above 60 or below 40.
• Exits: stop 1.5 ATR, target 3 ATR, 48 bars at most; or close after 12 bars. One trade at a time.
• Cost: 1 pip on EURUSD, $0.30 an ounce on gold, from my spread measurement.
• A match-up is one trend rule against the fade on the same chart with the same exit: 4 rules, 4 charts, 2 exits, 32 match-ups.

Which side won on each chart?

After cost per trade: pips on EURUSD, dollars an ounce on gold.

Read it by market, not by row. On EURUSD the fade won 14 of 16 match-ups, mostly by losing less. On gold the trend rules won 12 of 16, and on gold M15 the fade was the worst rule in the whole test: -$3.31 an ounce a trade with the ATR exit, 28.2% of trades won.

Gold and EURUSD have answered the same code differently before, in the one rule on gold and EURUSD test. Here the pattern was simple: on EURUSD a move in RSI tended to fade, on gold it tended to carry on. When RSI meets a band indicator on the same bars, see Bollinger Bands vs RSI.

Is the overbought and oversold fade any good on its own?

EURUSD M15 is the only run in the whole RSI test that beat the coin’s best 5%, by one hundredth of a pip, and it lost in the second half. On EURUSD M5 the fade sat right at the coin’s edge. On gold it lost on both timeframes. The fade made money before July 15 on two charts of four, EURUSD M15 and gold M5, and lost on both after it. My earlier divergence vs overbought test also found the plain 70/30 signal fragile.

Does “overbought” in an uptrend mean the price will fall?

This is the heart of the argument, so I measured it directly. Every time RSI 14 crossed above 70 while price was above the 200 EMA, I looked at where the close was 12 bars later:

The rest of the 187 and 66 EURUSD cases closed exactly flat.

On three charts it was a coin: half higher, half lower, a median move of zero. On EURUSD M15 the price was higher more often, 39 against 26, the opposite of what “overbought” promises. So both teachers are half right. RSI 70 in an uptrend is not a sell signal. It is not a buy signal either.

Why did neither side win outright?

Because both sides trade the same thing, a move in RSI, and an RSI move on M5 or M15 is mostly a few bars of noise. Above the 200 EMA, RSI 14 sat between 40 and 80 on about 91% of bars on every chart, and above 70 on 8.4% to 11.0%. A rule that sells every one of those visits is selling into the trend; a rule that buys every cross of 50 is buying the middle of a range. The market chose which mistake cost less.

The Williams %R vs Stochastic vs RSI comparison shows how close RSI is to the other two popular oscillators on these same bars, so swapping RSI for one of them is not a way out.

So which RSI approach should a beginner use?

• Check the market before the method. On EURUSD the fade lost less in 14 of 16 match-ups, on gold the trend side won 12 of 16. The same rule set gave opposite answers.
• Do not sell RSI 70 in an uptrend because it is “overbought”. Price was as likely to be higher 12 bars later as lower.
• Do not expect the trend rules to save you either. Even on gold, where they won the match-ups, no trend rule beat the coin’s best 5% and also made money in both halves.
• If you test one idea, test the trend side on gold M15. Its best version, the 50 cross with a 12-bar exit, made +$1.74 an ounce over 160 trades and +$1.39 and +$1.92 in the two halves. The settings that did best are in RSI trend settings for MT4.

Frequently asked questions

Is it better to trade RSI with the trend or overbought and oversold?

It depended on the market. On gold the trend rules beat the 70/30 fade in 12 of 16 match-ups; on EURUSD the fade lost less in 14 of 16. Neither side produced a rule that passed every check.

Does RSI above 70 mean the price will drop?

Not in an uptrend. On EURUSD M5, after RSI crossed above 70 with price over the 200 EMA, the close 12 bars later was higher 86 times and lower 81 times. On gold M5 it was 85 and 86.

Does the RSI 70/30 strategy work on EURUSD?

On EURUSD M15 with an ATR stop and target it made +0.20 pips a trade over 123 trades, the best fade result in the test, but +1.26 in the first half and -0.84 in the second. On M5 it lost 0.54 pips a trade.

Why does the RSI fade lose on gold?

Gold trended over these weeks. On gold M15 the fade lost $3.31 an ounce a trade and won 28.2% of trades, and its sells did worst: -$4.14 an ounce a trade against -$2.42 for its buys.

Can RSI tell you the trend?

It can describe it: above the 200 EMA RSI 14 spent about 91% of bars between 40 and 80. It did not give a profitable entry inside that trend on M5 or M15.