1 The Edge — why it works
Buy the dip — but only inside an uptrend
The 2-period RSI strategy buys short-term oversold extremes within a confirmed uptrend. A 2-period RSI is so fast it routinely spikes into single digits on a normal pullback — and in an uptrend, those pullbacks tend to bounce.
It fades the panic, never the trend. The 200-day moving average is the gate that keeps you on the right side.
2 Where it works — and doesn't
Mean reversion needs a trend to revert to
Works best when…
- Liquid indices & large-caps in a clear uptrend (price > 200-day MA).
- Orderly pullbacks, not news-driven collapses.
- Markets with a mean-reverting character.
Fails / avoid when…
- Downtrends — buying RSI(2) lows is catching a falling knife.
- Strong gaps / earnings — the dip keeps going.
- Runaway momentum that never pulls back.
3 Setup checklist
All true before you act
- ✓Uptrend filter. Price is above its 200-day moving average.
- ✓Oversold trigger. 2-period RSI closes below 5 (below 10 is a softer variant).
- ✓No event risk. No earnings or major news pending.
- ✓Liquidity. Tight spreads; you can size and exit cleanly.
4 The process
Connors' rules, with risk added
Entry
On the close when RSI(2) < 5 and price is above the 200-day MA.
Stop (1R)
Connors' original uses an indicator exit, not a hard stop — but for risk control set a disaster stop (e.g., below the recent swing low or a fixed ATR multiple). Entry − stop = 1R.
Position size
Risk a small fixed % of the account; shares = risk ÷ 1R.
Exit & management
Exit when price closes above its 5-day moving average (the bounce has played out), or if price closes back below the 200-day MA.
5 Worked example (illustrative)
In R
| Account / risk | $25,000 · 1% = $250 |
| Entry (RSI2 < 5, uptrend) | $98.00 |
| Disaster stop — 1R | $94.00 · 1R = $4.00 |
| Size = $250 ÷ $4.00 | ≈ 62 shares |
| Exit (close > 5-day MA), +1.5R | $104.00 · +$375 |
| If stopped: −1R | − $250 |
6 Honest expectancy
High hit-rate, fat-tail risk
Connors' published backtests show high win rates for RSI(2) on indices — but two honest caveats matter. First, the original rules have no hard stop, so a minority of trades become large losers when a dip doesn't bounce; that's why we add one. Second, backtests are historical — mean reversion works until a trend regime breaks, and it fails badly in bear markets.
A genuine edge here comes from the trend filter, selectivity, and risk control — and it is an expectation, never a guarantee.
7 Make it yours
Test before you trade
A no-risk validation routine
Backtest the exact rules on your own market and timeframe. Compare RSI(2) < 5 vs < 10, test adding a hard stop, and confirm the 200-day filter actually improves results. Track everything in R before risking a cent.
8 Common mistakes
How traders blow this up
- Using it in downtrends. The single biggest killer — no 200-day filter.
- No stop. One un-stopped loser can erase many small wins.
- Oversizing. High win rate tempts over-leverage; the rare loss then hurts.
- Fighting strong news. Earnings and shocks override mean reversion.