The setup
Twenty-period Bollinger Bands at two standard deviations, RSI at fourteen. A long condition is price touching or piercing the lower band while RSI is under 30; a short is the upper band with RSI over 70. Enter on the close of the candle that closes back inside the band, not while it is still outside — outside the band is where a trend spends its time, and buying it is the classic way this method fails.
It only works in a range
Bands measure how far price is from its own recent average, so in a trend price rides the band for a long time and every touch is a loss. Before taking any of these signals, confirm the higher timeframe is going sideways between two levels. If it is trending, this strategy is off — not adjusted, off. That single condition is the difference between the method working and the method being a slow way to pay spread.
Stop and target
The stop goes beyond the recent extreme, which on a short timeframe is close by. The target is the middle band, the twenty-period average, because that is what mean reversion actually claims: a return to the average, not a move to the opposite side. Taking the middle band and leaving the rest is what keeps the win rate high enough for the small targets to add up.
The spread arithmetic
If the target is ten pips and the spread is two, a fifth of every winning trade is gone before it starts, and the same two pips are added to every loss. Do that arithmetic before trading this at all: with a spread of two pips and a ten-pip target you need roughly a 55% win rate merely to break even. On a pair with a three-pip spread, or during the quiet hours when spreads widen, the method cannot be profitable no matter how good the signals are.
Who this is not for
Scalping needs many trades, a very low cost per trade, and complete indifference to individual outcomes. On a phone, on a retail spread, taken a few times a day around a job, the costs dominate and it does not work. It is included here because people ask about it constantly, and the honest version — including the arithmetic that usually rules it out — is more useful than the version that leaves the spread out.
In short
| Bands | 20 period, 2 standard deviations |
|---|---|
| Long | Lower band touch, RSI under 30, close back inside |
| Required | A ranging higher timeframe |
| Target | The middle band |
| Decided by | Spread against target size |
Common questions
Why does this lose in a trend?
Because price riding the band is normal in a trend, so the touch is not a stretch — it is the trend. The band is a measure of distance from the average, not a barrier.
Can I aim for the opposite band?
You can, and the win rate falls a long way. Mean reversion claims a return to the mean; anything past it is a different trade with different odds.
What spread makes this unworkable?
Run the numbers on your own pair: spread divided by target, doubled, is roughly how much of the edge the cost takes. If that is a large fraction, no amount of signal quality fixes it.
Trading forex and CFDs on margin carries a high level of risk. Most retail accounts lose money. Nothing on this page is financial advice or a recommendation to trade.