Sprint 2 starts: Bollinger Bands + RSI is live. Pro+ available across all asset classes, now part of the daily snapshot pipeline.
This gives us the first two real mean-reversion strategies in the new wave (rsi_ob_os was already there). The point of BB+RSI: strict on entry, generous on exit.
The mechanics
Both indicators are computed in parallel, then their signals are combined:
Entry (BUY) requires BOTH conditions simultaneously:
- Close at or below the lower Bollinger band (price stretched downward)
- RSI below the oversold threshold (momentum exhausted downward)
If only one is true — e.g. price drops to the lower band but RSI is at 50 — no trade. Both must match at the same time.
Exit (SELL) requires only ONE condition:
- Close reaches/breaks the upper Bollinger band, OR
- RSI rises above the overbought threshold
Whichever of the two fires first triggers the SELL.
Why this asymmetry
Classical mean-reversion theory: high conviction on entry, fast profit-taking on exit.
On entry you want to be sure the market is really stretched — so you wait for two independent confirmations. On exit you want to lock in profits fast as soon as the market recovers — regardless of which indicator says "overbought" first.
Symmetric AND-AND would be more conservative but would give back profits more often. Asymmetric AND-OR is the standard literature variant — we use it.
Comparison to RSI Overbought/Oversold
RSI/OB-OS has been in the platform for a while. Both strategies share the RSI logic on SELL, differ on BUY:
- RSI/OB-OS: BUY when
RSI < oversold. Period. - BB+RSI: BUY when
RSI < oversoldANDclose <= lower band. Both.
On daily crypto charts I'd expect ~30-50% fewer trades with BB+RSI vs RSI/OB-OS at identical default params. Per trade likely higher win rate (because setups are stronger), but fewer opportunities overall.
Which variant is better depends on market regime. In quiet sideways phases where RSI doesn't often hit extreme values, both are similar. In volatile phases with frequent wash-outs, BB+RSI plays its strength: the lower band adds confirmation that the move was really overdone.
The daily snapshot pipeline runs both. In ~3 weeks we'll have data for a direct comparison.
Where BB+RSI performs especially poorly
An honest weakness: strong downtrends with persistent oversold.
Example BTC mid-2022 — price spent months below the lower BB with RSI under 30. BB+RSI would have bought multiple times and gotten stopped out at each bounce-then-fail. That's the classic "catching a falling knife" problem of all mean-reversion strategies.
Mitigation from the platform: combine with the 200 WMA filter or the Bullmarket Gauge filter. Both block entries when the long-term trend is bearish — so BB+RSI only fires when the broader regime is bullish AND the local mean-reversion conditions align. Drops false-knife-catches drastically at the cost of skipping some real reversal bottoms.
Where available
Pro+ feature, live now at /dashboard/crypto:
- 5 input fields: BB period, BB std dev, RSI period, RSI oversold, RSI overbought
- All standard filters (200 WMA, ATR Volatility, Altcoin Season, Bullmarket Gauge, Min Profit Guard) combinable
- Works as Live Alert and in the Gauge
- Strategy Library with pseudo-code + FAQ at /strategies/bb-rsi