Bollinger Band settings for scalping: 2.0 vs 2.5 deviations and what you give up

[Fairman]

Bollinger Band settings for scalping: 2.0 vs 2.5 deviations and what you give up

The default Bollinger Bands are 20 periods and 2 standard deviations. Some scalpers push the deviation up to 2.5 or even 3 to catch only the more extreme moves, while others go down to 1.5 for more frequent signals. Each choice has a tradeoff.

At 2.0, price touches or pierces the bands fairly often on M1 and M5. If you fade band touches, you'll get a lot of opportunities, but many of them happen in trending moves where price walks along the band. Fading those is painful.

At 2.5, touches are rarer and more often mark a genuine stretch from the mean. You get fewer trades, but each one is more likely to be at an extreme. The cost is that you miss the moderate moves that might have been fine trades, and the bands are wider, so stops beyond them are further away.

At 1.5, the bands are tight and touches are frequent enough that they're more useful as a trend filter than a reversal signal. Price spending time outside a 1.5 band suggests directional pressure.

Whatever you pick, the band width itself is useful. Narrow bands on M5 often come before an expansion, so I watch for that as a heads up rather than a signal.

How do others here use the deviation setting, and does it change by session?

[Fairman]

Changing the period has an effect too, and it interacts with deviation. A shorter period with the same deviation makes the bands hug price more closely and react faster to changes in volatility. On M1, a 20-period band covers twenty minutes, which can be very different from the session's real context. Some people use a longer period on M1 to match a more meaningful window, like an hour. I'd test changes to period and deviation separately rather than both at once, so you can tell which one made the difference.

[Fairman]

I'd add that band touches near a known level, like the previous day's high or a session high, mean more than touches in the middle of nowhere. The bands are just a statistical envelope. They don't know where liquidity is. Combining a band extreme with a structural level gives you a reason for the price to react there, not just a sign that price is stretched. Without that, a fade at the upper band in a strong trend is mostly hoping the stretch ends right where you happened to enter.

[Fairman]

Band width itself can be useful separate from the deviation setting. When the bands are very narrow compared with their recent average, the market is compressed, and a breakout might be coming. When they're very wide, the market is already volatile. Some traders only fade band touches when width is normal or narrow, and avoid fading when the bands have just expanded sharply, because that's when touches turn into trends. A width percentile over the last hundred bars is an easy number to compute and compare across sessions. When width is in the bottom fifth of its recent range, I'm much less keen to fade a touch.

[FXS Prop Desk]

One thing that tends to get missed in the 2.0 vs 2.5 debate is how many pips the distance to the middle band actually is. In the late Asian session on EURUSD M1, the 20-period standard deviation can be so small that even a 2.5 touch sits only three or four pips from the midline. After spread and commission, a fade back to the middle has very little left in it, however clean the signal looks.

I'd compare the two settings by expectancy per day rather than win rate: log each touch, the pip distance to the midline at that moment, and whether price got there before moving another band width against you. Did you find the London open changes which setting comes out ahead?