Does the Bollinger Bands + stochastic strategy work on USD/CAD?

Key results

Total return

-42.5%

Buy & hold: +8.5%

Annual return (CAGR)

-5.4%

Buy & hold: +0.8%

Max drawdown

-45.1%

Buy & hold: -18%

Win rate

55.2%

Profit factor

0.77

Trades

1,913

Longest losing streak

11 trades

Period

Oct 2, 2016 to Oct 1, 2026

10 years

Market

USD/CAD

1-hour

Equity curve vs buy & hold

Returns year by year

Calendar years, compounded from monthly results. The first and last years can be partial.

The Bollinger Band bounce with a stochastic confirmation is a staple of forex education, including BabyPips' Bollinger Bands lesson: when price stretches to an outer band and the stochastic says the move is overdone, bet on a return to the middle. We tested it as written, with the middle band as the exit, on ten years of hourly USD/CAD, a pair often cited as range-friendly.

The rules we tested

• Market and timeframe: USD/CAD, 1-hour candles.
• Period: 2 October 2016 to 1 October 2026.
• Long entry: the close is below the lower Bollinger Band (20 periods) and stochastic %K (14) is below 20.
• Short entry: the close is above the upper Bollinger Band and stochastic %K is above 80.
• Exit: longs close when the price crosses above the middle band, shorts when it crosses below it. An opposite entry signal also closes and reverses the position.
• Stop-loss: 45 pips. There is no separate profit target: the middle band is the target.
• Costs and sizing: 0.02% commission per side, full position size on a 10,000 starting balance.

What the backtest shows

The headline contradiction is the point of this page. 55% of trades won, more than half, yet the account lost 42.5% (-5.4% a year) with a maximum drawdown of 45.1%. Holding USD/CAD over the same period returned +8.5% with an 18% drawdown.

The exits tell you why. 1,375 trades ended on a signal, mostly the return to the middle band, and 537 hit the 45-pip stop. A move from an outer band back to the middle on an hourly chart is usually short, so the typical win is small. The stop, by contrast, is a full 45 pips. A profit factor of 0.77 means gross losses were about 30% larger than gross wins.

Then there is frequency. Close to 200 trades a year at 0.04% per round trip adds up to roughly 77% of the position size over ten years. A strategy built on small wins cannot absorb that.

The yearly table shows a steady decline rather than one bad period. Only the partial 2016 (+0.27%) and 2017 (+2.78%) were positive. Every year since has lost between 1.85% and 9.53%, the worst full years being 2019 (-9.53%) and 2024 (-8.01%).

Why it doesn't work here

The bounce assumes that a close outside the band is an excess that will be corrected. Sometimes it is. But a close outside the band is also what the start of a trend looks like, and the stochastic stays oversold or overbought for long stretches when a trend is real. In those cases the entry fires repeatedly into a move that keeps going, and the 45-pip stop pays for it. Meanwhile the successful trades collect only the distance to the middle band, which shrinks as the bands narrow. Small wins, full-size losses and heavy costs is a combination that bleeds slowly, and that is exactly the shape of the equity curve.

What you could test next

• Add a trend filter such as a 200-period EMA and only take bounces in its direction, or only when price is flat around it.
• Require the stochastic to cross back up through 20 (or down through 80) before entering, instead of simply being in the zone.
• Move to the 4-hour chart, where moves back to the middle band are longer and costs weigh less.
• Use the opposite band as the target to see whether a larger payoff changes the balance.

Methodology and assumptions

Frequently asked questions

Reproduce it, then change it

Every number on this page comes from the Backtrex engine. Rebuild the strategy in the app, then change the market, the timeframe or a parameter and see whether the result holds.

Other backtests