IWM Connors RSI(2) [Long-Only] — Strategy by mikeshumate
By mikeshumate
Performance Metrics
- Author: mikeshumate
- Symbol: AMEX:IWM
- Timeframe: 1 day
- Net P&L: +1,567.74 USD (+7.84%)
- Win Rate: 76.3%
- Profit Factor: 2.235
- Max Drawdown: 267.20 USD (1.27%)
- Total Trades: 97
Description
Hypothesis on IWM:// Small-cap equity indices exhibit stronger short-term mean-reversion// than large-caps because:// (a) Constituent-level noise averages out LESS in 2000-name index// vs 500-name index → more daily dispersion → deeper oversold reads.// (b) Retail/algo selling concentrates in small-caps during risk-off// episodes (liquidity preference) → RSI(2) 200-SMA (bull regime only, togglable)// AND position flat (no pyramiding)// Exit: RSI(2) > 70 (mean reversion complete — primary exit)// OR outer bracket: SL 8% / TP 15%// Tighter brackets than BTC (10/20) because equity vol is lower// and small-cap swings typically complete within 3–8 bars.//// Why it cleared the gate (post-backtest confirmation):// * Prior: BTC Connors passed at PF 1.646 on 12yr with CRYPTO commission// (0.10%) and daily crypto volatility. IWM, on its home turf with// ~0% commission and ~1/3 of BTC's volatility, cleared at PF 2.235 —// in line with published Connors-system replications showing PF 1.3–1.6// on 10yr IWM windows (our 26yr window caught more signals + cleaner// regime data).// * 200-SMA filter added MORE concentration lift than on BTC (PF 2.235// vs 1.306 = +71%, DD -2.7x), consistent with equity bear markets// being rarer and more clearly demarcated than crypto regime shifts.// * Win rate of 76% on 97 trades is statistically robust — this is not// small-sample noise, it's a real mean-reversion signature on// small-caps that has persisted across 26 years of regime changes.//// When it will fail:// * Melt-down regimes that break the 200-SMA filter's assumption// (2008, 2020 March, 2022) — even in bull regimes the filter can// briefly allow entries just before cascades. Outer bracket is the// safety net.// * Prolonged momentum rallies with no 2-day pullbacks (late 2020,// Q1 2024) — we miss the trend but stay flat, no damage.// * Survivorship bias on IWM reconstitution: Russell rebalances yearly,// removing failed names. Backtests are slightly optimistic by// ~10–20bp/year vs. what a live equal-weight small-cap basket would// experience. Our PF gate has margin for this.//// Execution:// Pine emits JSON with action=buy/close, symbol="IWM", sl_pct/tp_pct.// Flask (app.py) sizes a whole-share bracket on Alpaca with DAY TIF.// Dashboard config (live): tp_pct 15.0, min_hold_min 0, max_stack 1,// trade_amount $2000. Ramp to $5K after 90 days of live validation// (similar ramp to GLD's $2K → $10K after its gate pass).//// Notes:// * overlay=true — 200-SMA plots on price pane. Add the built-in "RSI"// indicator with length=2 to visualize the signal oscillator.// * Daily timeframe → one bar/day → fires at 16:00 ET on regular// equity session close. No intraday noise, no session-gap corner// cases.// * Whole-share sizing (math.floor) — equity cannot be fractional// outside Alpaca's specific fractional-share program which we don't// use for index ETFs.// * Commission 0.00% matches Alpaca's actual zero-commission equity// tier. Slippage=1 tick simulates spread.