Larry Connors' classic equity edge — buy short-term oversold dips inside an uptrend, exit fast on the bounce.
Basis: 1928 user backtests · BTCUSDT · 1d · 4 years
RSI(2) Mean Reversion is the best-documented mean-reversion edge in equity research, popularized by Larry Connors and Cesar Alvarez in Short Term Trading Strategies That Work. It combines a long-term trend filter with a very short-term momentum extreme.
Trend filter: only take longs when the close is above the 200-period SMA — you only buy dips inside an established uptrend, never falling knives.
Entry: when the 2-period RSI drops below 10, the asset is sharply oversold on a short horizon. Inside an uptrend, that dip statistically tends to revert.
Exit: close the position when RSI(2) climbs back above 70 (momentum normalised) OR when the close rises above the 5-period SMA (the bounce played out). Whichever comes first.
The edge is short-lived and high-turnover — holds typically last a few bars. It is built for stocks and ETFs, which mean-revert on daily/weekly horizons, but also works on crypto. The 200-SMA here is computed on the candle series itself (on a daily chart = 200 days) — it is NOT the platform-wide 200-week filter.
| Name | Default | Range | Description |
|---|---|---|---|
| RSI Period | 2 | 2–50 | Lookback for the RSI. Connors uses 2 — a very fast oscillator that swings to extremes within a couple of bars. |
| Entry Threshold | 10 | 1–49 | RSI value below which a buy fires (inside an uptrend). Lower = stricter, fewer trades. Connors variants use 5 or 10. |
| Exit Threshold | 70 | 51–95 | RSI value above which the position exits. Default 70. |
| Trend SMA Period | 200 | 20–400 | Long-term trend filter. Longs only when close > this SMA. Computed on the candle series (daily chart = 200 days). |
| Exit SMA Period | 5 | 2–50 | Short SMA for the bounce exit. Position closes when close rises above this SMA. |
The pre-baked mini-backtest is refreshed daily — check back soon or start a live run in the Arena.
Run in Arena →Top-10 assets by average CAGR (1d interval), aggregated from community + platform backtests. Actual results depend on parameters and period.
| Asset | CAGR | vs B&H | ★ | Runs |
|---|---|---|---|---|
| AGTUSDT | +100.6% | -156.8pp | 57 | 1 |
| VVVUSDT | +85.8% | -378.8pp | 45 | 1 |
| FTMUSDT | +41.4% | -14.8pp | 56 | 12 |
| ARCUSDT | +38.5% | -22.6pp | 49 | 5 |
| STXUSDT | +35.4% | +66.6pp | 85 | 1 |
| VELVETUSDT | +33.1% | +62.5pp | 76 | 1 |
| XCNUSDT | +26.2% | +31.1pp | 83 | 1 |
| RUNEUSDT | +25.5% | +69.5pp | 79 | 11 |
| SUIUSDT | +24.1% | +52.8pp | 71 | 11 |
| ESUSDT | +19.2% | +118.9pp | 87 | 1 |
// Entry
if close > SMA(trendSmaPeriod) and RSI(rsiPeriod) < entryThreshold:
if position.is_flat:
BUY
// Exit
if RSI(rsiPeriod) > exitThreshold or close > SMA(exitSmaPeriod):
if position.is_long:
SELLA 2-period RSI is extremely sensitive — it swings between 0 and 100 within one or two bars. Connors found that this short horizon captures sharp, tradeable oversold extremes far better than the standard 14-period RSI, which is too smooth to reach sub-10 readings often. The trade-off is more noise, which the 200-SMA trend filter tames.
The strategy only fires inside an uptrend (close > 200-SMA). Assets in long bull runs produce many dip-buys; assets below their 200-SMA produce none. That is by design — it refuses to fight the trend. For statistical meaning, aggregate across multiple assets.
No. The trend SMA here is computed strategy-internally on the candle series you backtest (on a daily chart that is a 200-day SMA). The platform's separate 200-WMA filter always works on weekly closes. They are different tools — do not confuse them.
Long-only. Entries open a long on an oversold dip in an uptrend; exits close it. There is no shorting — the platform is long-only by design.
Check out our Strategy Insights Reports — pre-baked deep-dives with historical results, comparisons, and market context.