"RSI is oversold! Time to buy!"
Famous last words.
I've lost more money buying "oversold" conditions than I care to admit. RSI at 25. Must be a bottom, right?
Then RSI goes to 20. Then 15. Then 10. Price keeps falling. My account keeps bleeding.
The indicator was oversold. The market didn't care.
Let me bust a myth that cost me thousands.
Oversold doesn't mean "buy." Overbought doesn't mean "sell."
RSI measures momentum. When RSI is low, it means price has been falling. That's it. It doesn't mean price will stop falling.
In a strong downtrend, RSI can stay "oversold" for weeks. Months even. Buying every RSI 30 reading in a bear market is a fast way to go broke.
I learned this the hard way. Multiple times. I'm a slow learner.
RSI (Relative Strength Index) measures the speed and magnitude of recent price changes.
That's useful information. But it's not a trading signal.
The signal comes from context. What's the trend? Where's the level? What's the price action saying?
After years of getting burned, here's my RSI playbook.
This is the one RSI signal that actually works for me.
Bullish divergence: Price makes a lower low, but RSI makes a higher low.
Translation: Price is falling, but momentum is weakening. Sellers are losing steam.
Bearish divergence: Price makes a higher high, but RSI makes a lower high.
Translation: Price is rising, but momentum is weakening. Buyers are losing steam.
Divergences don't guarantee reversals. But they're a warning sign. Combined with a key level and price action confirmation, they're powerful.
Here's a simple framework:
In an uptrend:
In a downtrend:
Notice: I'm not buying oversold in downtrends or selling overbought in uptrends. That's fighting the trend. Bad idea.
RSI alone means nothing. RSI at a key level means something.
Price at major support + RSI divergence + bullish candle = high probability long.
Price in the middle of nowhere + RSI oversold = probably a trap.
The level matters more than the indicator.
Here's a real strategy I use.
The Divergence Reversal
Conditions:
Entry: On close of confirmation candle Stop: Beyond the recent extreme Target: Next major level or 2:1 reward
I set this up in dashpull as a conditional order. The system watches for divergence + level + confirmation. When all conditions align, it executes.
"Should I use RSI 7 or RSI 14 or RSI 21?"
Honestly? It doesn't matter much.
RSI 14 is the default. It works fine. Shorter periods (7) are more sensitive. Longer periods (21) are smoother.
I use RSI 14. I've tested others. The difference is marginal.
Don't waste time optimizing RSI settings. Focus on how you use it.
I've said it three times already. I'll say it again.
Oversold in a downtrend = price keeps falling.
Don't buy just because RSI is low. Wait for trend change or at least a key level with confirmation.
RSI works differently in trends vs. ranges.
In a range, overbought/oversold extremes are meaningful. Price tends to reverse at range boundaries.
In a trend, extremes just mean the trend is strong. Fighting them is painful.
Know your context.
Textbook divergences are rare. Real divergences are messy.
The lows don't have to be exactly aligned. The RSI reading doesn't have to be exactly higher. Close enough is good enough.
Don't miss good setups waiting for perfection.
RSI + nothing = bad trades.
RSI + trend + level + price action = good trades.
Never use any indicator in isolation. Always combine with other factors.
Here's something useful:
Check RSI on multiple timeframes.
If daily RSI is oversold AND 4H RSI is showing divergence AND 1H RSI is turning up... that's confluence. Multiple timeframes agreeing.
If daily RSI is oversold but 4H and 1H are still falling... the bounce might not be ready yet.
I use higher timeframe RSI for direction, lower timeframe RSI for timing.
The beauty of RSI is that it's quantifiable. A number. Easy to code into conditions.
In dashpull, I can set up:
Condition set 1:
Condition set 2:
When conditions are met, the order executes. I don't need to watch RSI all day.
Let me give you some honest numbers.
RSI oversold alone: ~40% win rate (basically random) RSI oversold + uptrend: ~55% win rate RSI divergence + key level: ~60% win rate RSI divergence + key level + confirmation: ~65% win rate
These are rough estimates from my own trading. Your results may vary.
The point: RSI alone is useless. RSI combined with other factors is useful.
RSI is a tool. A useful tool when used correctly.
Stop using it as a buy/sell signal. Start using it as one piece of a larger puzzle.
The puzzle includes:
When all pieces align, you have a trade. When only RSI is signaling, you have a trap.
dashpull helps me combine all these pieces into conditional orders. RSI condition + level condition + pattern condition = execute. No emotional interference. No "but RSI says..."
Just systematic execution of a complete strategy.
Ready to build RSI-based conditional orders? Try dashpull →
Trading Indicators: The Brutal Truth About RSI, MACD, and Everything Else
Most traders use indicators wrong. Here's what indicators actually tell you, what they don't, and how to use them without getting destroyed.
MACD Trading Strategy: Why the Golden Cross Keeps Losing You Money
MACD crossovers are the most popular signal in trading. They're also one of the worst. Here's how to actually use MACD.