The RSI (Relative Strength Index) measures how fast and how far price has moved recently, on a 0–100 scale. The classic reading: above 70 = overbought, below 30 = oversold. The classic beginner mistake: treating those levels as automatic buy/sell signals. In a strong downtrend, RSI can sit below 30 for weeks while price keeps falling.

What makes RSI genuinely useful is combining it with a trend filter: buy oversold dips in uptrends only. Here's that strategy, in testable form.

The strategy rules

• Timeframe: 1H or 4H (works on 15m for experienced traders)
• Indicators: RSI (14), EMA (200)
• Setup (long): price is above the 200 EMA — uptrend confirmed
• Entry trigger: RSI (14) dips below 40, then closes back above 40
• Stop-loss: below the most recent swing low
• Take profit: 2× your risk (2R), or exit when RSI reaches 70
• Short version: mirror everything below the 200 EMA (RSI rallies to 60, rolls over)

Note the tweak: in real uptrends RSI rarely reaches 30 — pullbacks usually bottom in the 35–45 zone. Waiting for a textbook "30" makes you miss most of the trend's dips. This is exactly the kind of assumption you should verify yourself in replay.

Why it works

An uptrend is a series of impulses and pullbacks. The 200 EMA filter keeps you on the side of the dominant flow; the RSI dip identifies the pullback; the close back above 40 signals the pullback is ending. You're not predicting a reversal — you're joining an existing trend at a discount.

Variations worth testing

• RSI divergence: price makes a lower low while RSI makes a higher low — a reversal signal, stronger at major support levels. Powerful but rarer.
• RSI + MACD: require a MACD bullish cross as confirmation. Fewer trades, higher quality.
• RSI + Bollinger Bands: RSI oversold and price at the lower Bollinger Band for mean-reversion in ranges.
• Settings: RSI 14 is standard; RSI 7 reacts faster (more signals, more noise), RSI 21 smoother. Don't optimize to perfection — that's overfitting.

How to backtest this strategy

1. Open the free chart replay tool and load a random historical chart on your timeframe.
2. Check trend (price vs. 200 EMA area), then advance candle by candle.
3. Take every valid RSI signal — enter, set the stop, target 2R, and let it resolve.
4. Log 50 trades and compute expectancy. Most traders find win rates near 45–55% with 2R winners — profitable, but only if losers are cut at 1R every time.

Full method: how to backtest a trading strategy.

Common mistakes

1. Counter-trend "oversold" buying — RSI below 30 in a downtrend is a chainsaw, not a discount.
2. No stop-loss because "it's already oversold" — it can get more oversold.
3. Exiting winners at +0.5R out of fear — the math needs your 2R winners.
4. Trading every timeframe at once — pick one, build the sample, then expand.

FAQ

What is the best RSI setting for trading?

RSI 14 on 1H–4H charts is the standard and a fine default. Shorter (7) suits scalping but produces more false signals. The honest answer: backtest 7 vs. 14 on your market in replay and use your own data.

What win rate does an RSI strategy have?

With a 200 EMA trend filter and 2R targets, replayed backtests typically land around 45–55% — profitable through the win/loss size ratio, not the hit rate. Unfiltered "buy below 30" performs far worse in trends.

Is RSI good for day trading?

Yes, on 5–15 minute charts with the same trend-filter logic — but expect more noise and tighter fee margins. Test it in replay on minute data before trading it live.