Shared report
Free viewBollinger Band Mean Reversion
We tested 358 variations on older data and kept 2H candles instead of 1H, stop 3×ATR(14) and removed exit rule (close_vs_ma). Chosen on the window shown, from 262 candidates checked there.
On the 2-hour chart of Euro / US Dollar it buys when close crosses below the lower Bollinger band (20, 2σ), and sells short when close crosses above the upper Bollinger band (20, 2σ). Stop 3× ATR(14) against the trade, 1% of the account per trade.
2 differences from what you wrotethe very next candle closes back inside the bands · Exit rule (close_vs_ma)
- the very next candle closes back inside the bandsbb_break (cross: true) followed by a candle closing inside
The grammar does not have a 'next candle' sequence for BBs; approximated as a BB break followed by a return.
- Exit rule (close_vs_ma)Not tested
Backtestify can't test this exit yet. Trades close on the stop, the target or the end of the data instead.
Notes
- 40 trades: the most these rules produce in 2 years of 1h candles, short of the 100 a confident read wants.
- 45 trades: the most these rules produce in 1.9 years of 1h candles, short of the 100 a confident read wants.
Market
Timeframe
Account
History
Requested 2025-08-30 → 2026-10-18 · traded from 2025-08-31 (warm-up 108 bars)
Verdict
RECOMMENDEDwin rate · 59 trades
- Ended above the starting budget+6.4% · 0.111R
- Win rate 50% or above61.0%
- More than 50 trades59 trades
- Verdict rules v9
- Trades
- 59
- Timeframe
- 2H
- Selection window
- 206 days
- Time in market
- 76.8%
high win rate · 59 trades
What to watch
- The window shown was used to choose these rules (improve: the shipped variant was chosen among 262 candidates measured on this window), so no untouched window is left to check them on
- Removed: exit rule (close_vs_ma) — the engine cannot execute it
- Tested an approximation of your rules
- Could not be checked with doubled costs
2 moreShow less
- The unseen window was used to choose these rules (improve: the shipped variant was chosen among 262 candidates measured on this window), so no untouched window is left to check them on
- Chosen from 263 candidates: not significant after the selection penalty (p 1.000)
What the data supports
Short side held up
+0.319R average trade · 69.2% win rate
short only · 26 of 59 trades
Best entry hours (analysis, not a rule)
+0.296R average trade · 60.0% win rate
entries 18:00–00:00 NY · 10 of 59 trades
Beat buy & hold
+6.4% vs -3.1%
same symbol, same window · 59 closed trades
Months in profit
50% of months closed up
4 of 8 months in the tested window
Win rate · original vs improved
Original
Improved
Both measured on the same window. The improved rules were chosen on it.
What changed
3 upgrades applied
Timeframe
2H candles instead of 1H
Stop
3×ATR(14)
Removed
Exit rule (close_vs_ma) — the engine cannot execute it
Profit factor
0.25×Avg per trade
0.11RWin rate
27.91 ppReturn
5.63 ppMax drawdown
3.65 ppTrades
89Evidence window
59 trades
chosen on this window
Best of ≥ 263 candidates on this window
Failed
t 0.84 · p 1.000 · best of ≥ 263 (Bonferroni)
Variations tested
358
262 checked on this window
Account balance
Selection window only (25 Mar 2026 – 16 Oct 2026)
$10,000
Starting balance
$10,079
Final · original
$10,671
With the open position marked
$9,692
Buy & hold
- Original
- Improved
- Buy & hold
Trust scoreThe fixed factors behind the trust score, each with its own measurement.
C
Grade
RECOMMENDED
Verdict
Strategy replay
- BB 20, 2
Entry
1.1184
Buy · 12 Oct, 07:00 NY · 25 bars
Stop loss
1.1143
0.37% away
Take profit
None
Exits on its rule
Result
-1.03R
Stop loss at 1.1142 · 14 Oct, 07:00 NY
Trace4 steps
Signal
candle close
1.1184
12 Oct, 05:00 NY
Filled
next open
1.1184
12 Oct, 07:00 NY
Stop
ATR distance
1.1143
12 Oct, 07:00 NY
Exit · Stop loss · -1.03R
1.1143
14 Oct, 07:00 NY
Monthly returns
Original beside improved, month by month
- Original
- Improved
Wins and losses
Original
- Wins49
- Losses99
Improved
- Wins36
- Losses23
Improved rules
Changed from the original: Timeframe: 2H candles instead of 1H · Stop: 3×ATR(14) · Removed: exit rule (close_vs_ma) — the engine cannot execute it
Buys when
close crosses below the lower Bollinger band (20, 2σ)
Stop
3 × ATR(14) from entry
Target / exit
no fixed target
Risk per trade
1% of equity risked to the stop
- After an exit
- The next entry signal is read from the candle after the exit, not on the exit candle
- Opposite signal
- Closes the open trade and opens the reverse at the next open
- Not modelled
- Margin calls, liquidation and CFD overnight financing; perpetual funding is charged when the data carries it
Latest trades
59 closed trades · 1 open at the end of the data, marked at the last close
| Date | Direction | Entry price | Exit price | Result |
|---|---|---|---|---|
| 12 Oct 2026 | long | 1.1184 | 1.1142 | Loss |
| 8 Oct 2026 | short | 1.1230 | 1.1184 | Win |
| 4 Oct 2026 | long | 1.1182 | 1.1230 | Win |
| 1 Oct 2026 | long | 1.1295 | 1.1248 | Loss |
| 29 Sept 2026 | long | 1.1352 | 1.1313 | Loss |
| 25 Sept 2026 | short | 1.1394 | 1.1352 | Win |
| 23 Sept 2026 | long | 1.1382 | 1.1394 | Win |
| 23 Sept 2026 | long | 1.1431 | 1.1394 | Loss |
| 16 Sept 2026 | long | 1.1481 | 1.1436 | Loss |
| 13 Sept 2026 | long | 1.1586 | 1.1550 | Loss |
Monthly heatmap
Original
| Year | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2026 | — | — | -5.4% | 2.2% | 0.0% | -2.9% | 1.6% | -1.1% | -1.8% | 8.8% | — | — |
Improved
| Year | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2026 | — | — | 0.0% | 1.9% | -2.6% | 1.3% | 7.1% | 1.9% | -1.6% | -1.5% | — | — |
Advanced AnalysisPro