RSI Trend Catcher Tested: 0 of 40 Runs Passed, and on EURUSD M5 the 50-Line Cross Did Worse Than a Coin

“RSI is not just for overbought and oversold, use it to catch the trend” is one of the first pieces of advice a beginner hears. The usual versions are simple: buy when RSI crosses above 50, or wait for a pullback to 40 in an uptrend and buy when RSI turns back up. They sound reasonable and every rule in them is exact, so they can be counted.

I already tested RSI as a swing tool on H1 and H4 and as a divergence signal. This test is about the trend use only, on the same bars, with the same costs.

What exactly did I test?

• Data: EURUSD, June 3 to August 26, 2026: 16,926 M5 bars and 5,642 M15 bars built from them. Gold (XAUUSD) over the same weeks: 13,736 M5 bars and 4,577 M15 bars.
• RSI: period 14, Wilder’s formula, the one MT4 and MT5 use. Trend side: close above or below the 200 EMA.
• Five rules: (1) RSI crosses 50; (2) the same, only in the direction of the 200 EMA; (3) pullback: above the 200 EMA, RSI dips under 40 and closes back above it (short side: over 60 and back under it); (4) momentum: RSI breaks above 60 to buy, below 40 to sell; (5) the 70/30 fade, the opposite idea, as a reference.
• Two exits: stop 1.5 ATR, target 3 ATR, close after 48 bars at the latest; or a plain close after 12 bars. One trade at a time.
• Cost: 1 pip per EURUSD trade from my spread and swap measurement, $0.30 an ounce per gold trade.
• Checks: the data split at July 15 into two halves, and a coin: random entries on the same bars with the same exit and the same number of trades, 300 runs. A rule passes only if it makes money overall, in both halves, and beats the coin’s best 5%.

Does the RSI 50 cross catch trends?

Per trade. Coin = random direction on random bars, same exit, same number of trades, 300 runs; the column is the result that only 5% of coin runs beat.

On EURUSD M5 the cross was not just unprofitable, it was worse than chance. A coin with the same stop and target averaged about -1.0 pip a trade, which is the spread and nothing else. The RSI cross averaged -1.65, below 95% of coin runs. It lost 0.65 pips a trade before the spread, so the signal itself pointed the wrong way: on the 5-minute chart a cross of 50 came near the end of a small move, not at the start of a trend.

Gold M5 is the only row in the black, +$0.29 an ounce, and positive in both halves. It still sits under the coin’s +$0.43, so luck explains it.

Does filtering with the 200 EMA help?

Less than you would expect. Taking only the crosses on the side of the 200 EMA cut the number of trades by 9% to 15% and changed little else:

After cost per trade, trades in brackets.

Two better, two worse. With the 12-bar exit the filtered version did best on gold M15: +$1.72 an ounce, the same in both halves (+$1.72 each), but a coin on 130 gold M15 trades reached +$2.96 in its best 5%. The same story as the higher-timeframe filter test: a trend filter removes trades, it does not turn a weak entry into a good one.

What about buying the pullback to 40?

This is the version most “trend catcher” guides teach, and it was the worst on EURUSD:

On EURUSD M15 one trade in five ended in profit. On gold M15 the rule made money overall only because the second half was strong: the first half lost $4.18 an ounce a trade. On 66 trades that swing is normal, and a coin with 66 trades reached +$4.13 in its best 5%.

Is RSI 60/40 momentum any better?

Buying when RSI breaks above 60 and selling under 40 lost on EURUSD on both timeframes and both exits, from -0.17 to -1.66 pips a trade. On gold M15 it made +$1.37 an ounce with the 12-bar exit, +$4.00 in the first half and -$0.12 in the second. On gold M5 it made +$0.18, positive in both halves, and still under the coin’s +$0.59.

How many of the 40 runs passed?

The single run that beat the coin was not a trend rule: it was the 70/30 fade on EURUSD M15 with the ATR exit, +0.20 pips a trade. It made +1.26 in the first half and lost 0.84 in the second. The full trend-versus-fade comparison is in RSI trend rules vs overbought and oversold.

Does RSI really stay between 40 and 80 in an uptrend?

Yes. This part of the trend-RSI idea, usually credited to Andrew Cardwell, held on every chart:

Below the 200 EMA the mirror image held: RSI spent 92.4% to 94.4% of bars between 20 and 60.

So the description is right, and it is useful in one way: it tells you that waiting for RSI 30 in an uptrend means waiting for something that happens on 3 bars in 1,000. What it does not give you is an entry. Knowing where RSI lives did not make the pullback or the cross profitable.

So should a beginner use RSI to catch trends?

• Not as the entry signal on M5 or M15. 0 of 32 trend-rule runs passed, and on EURUSD M5 the 50 cross did worse than random entries.
• Use the 40-80 band as a description, not a trigger. It tells you the trend is alive; it does not tell you when to get in.
• Do not trust a single good gold result. The gold M15 rows that looked good flipped between halves or stayed under the coin.
• Exits mattered more than the entry. The same rule swung by up to $2.92 an ounce between the ATR exit and the 12-bar exit on gold M15 (the pullback rule: +$1.51 against -$1.41). My trailing stop test found the same about exits.

Which periods, levels and filters did best is in RSI trend settings for MT4. If you want RSI on a slower chart, the RSI swing setup for MT4 is the one that held up best on this blog so far.

Frequently asked questions

Does the RSI 50 cross work as a trend signal?

Not in this test. On EURUSD M5 it lost 1.65 pips a trade over 726 trades after a 1-pip cost, worse than 95% of random entries with the same exit. On gold M5 it made +$0.29 an ounce, within what luck gives.

Is buying RSI 40 in an uptrend a good strategy?

On EURUSD M15 it won 19.6% of 102 trades and lost 2.63 pips each. On gold M15 it made +$1.51 an ounce, but lost $4.18 a trade in the first half of the data.

Does the 200 EMA filter improve RSI trend entries?

Barely. It removed 9% to 15% of the trades, improved the result on 2 of 4 charts and made it worse on the other 2.

What RSI range shows an uptrend?

Above the 200 EMA, RSI 14 stayed between 40 and 80 on 90.3% to 91.4% of bars and went under 30 on 0.3% to 0.6% of them, on EURUSD and gold, M5 and M15.

Which RSI trend rule did best?

The RSI 50 cross with the 200 EMA on gold M15 with a 12-bar exit: +$1.72 an ounce a trade over 130 trades, the same in both halves, but below the +$2.96 that the best 5% of coin runs reached.