RSI Mean Reversion Credit Spreads

RSI Mean-Reversion — Short Spreads (overview)

When an instrument becomes overbought (high RSI) or oversold (low RSI) AND other filters agree (price vs moving average, volatility context, liquidity), we will sell defined-risk credit spreads that profit from mean reversion (RSI returning toward neutral).

---

Strategy components (defaults you can tune)

• Universe: liquid index & ETFs (e.g., SPX/SPY, QQQ, IWM) or highly liquid single names.

• Bar timeframe: 15-min intraday or daily (choose one per implementation). Defaults below assume daily (smoother signals); I’ll note intraday adjustments.

• RSI: Period = 14 (alt: 9 or 21 for faster/slower). Use Wilder-style RSI.

• RSI thresholds:

  • Overbought: RSI ≥ 75 (aggressive) / 70 (standard).

  • Oversold: RSI ≤ 25 (aggressive) / 30 (standard).

• Price trend filter: 20-period EMA (daily) — check how far price is from EMA (z-score or %).

• Volatility filter: IV Rank / IV Percentile over last 252 trading days. Prefer IVRank ≥ 40–50 to sell premium (you get richer premium).

• Liquidity: min option open interest & volume (e.g., OI ≥ 200, volume ≥ 50), bid-ask spread < 0.25×midpoint (or spread < $0.50 on index options).

• Strike selection: target short option delta ~0.10–0.30 (conservative 0.10–0.20). Long leg width 5–20 strikes (or set width to risk budget).

• DTE: 20–60 days to expiry (avoid < 10 DTE unless you have special intraday rules). For intraday implementation use 5–20 hours.

• Position sizing: risk-per-trade 1–2% of account equity (max), or use volatility-adjusted sizing (notional / ATR).

• Max concurrent trades: e.g., 3–6 depending on capital.

---

Entry Criteria (all must be checked — logical ANDs and ORs shown)

Primary directional decision:

• Sell call credit spread (bearish spread) when:

  1. RSI ≥ Overbought threshold (default 75).

  2. Price is above 20-EMA and has diverged from EMA by X% or z-score > +1 (shows extreme).

  3. IVRank ≥ 40 (preferably) — premium rich.

  4. No scheduled company-specific catalyst / earnings in next 10–14 days.

  5. Liquidity checks pass (OI/volume/spread).

  6. Market breadth or index-level confirmation (optional): e.g., sector index also extended or market internals weak.

• Sell put credit spread (bullish spread) when:

  1. RSI ≤ Oversold threshold (default 25).

  2. Price is below 20-EMA and diverged by X% or z-score < −1.

  3. IVRank ≥ 40.

  4. No earnings/catalyst.

  5. Liquidity checks pass.

  6. Market breadth/sector confirmation supports mean reversion.

Additional optional confirmation (useful to avoid false signals):

• RSI divergence: price makes new low but RSI does not (bullish divergence) before selling put spread; or price makes new high but RSI fails to confirm (bearish divergence) before selling call spread.

• Candlestick rejection: e.g., rejection wick away from extremes, hammer/shooting star on daily timeframe.

• Volume confirmation: extreme RSI accompanied by declining volume on the extreme move can help.

Entry execution:

• Prefer selling one-lot initially, scale if mean reversion begins and risk allows.

• Use limit orders for spreads at a target credit (e.g., 40–60% of theoretical mid) to avoid poor fills.

---

Exit Criteria (prioritize — execute as soon as any rule triggers)

A. Profit targets

• First target: 30–50% of max possible premium collected (common).

• Conservative: close at 25–30% early to lock profits; scale out (close half @ 30–40%, rest @ 50%).

B. Stop-loss / Hard exit

• Close if the position loses max_loss_per_trade (e.g., 50–100% of collected premium equivalent to X% account risk). Alternatively:

  • If short strike breached by ≤ 0.25 × width (i.e., underlying moves near/through short strike) — tighten and consider closing/rolling.

  • If loss reaches max_dollar_loss (e.g., 1–2% of equity) close trade.

C. Time-based exits

• If remaining DTE ≤ 7 days, consider closing to avoid gamma/assignment risk (unless position is comfortable and hedged).

• If days-in-trade > threshold (e.g., you planned 30 DTE and now 5 DTE left) close or manage.

D. IV & volatility exits

• If IV spikes suddenly (IVRank jump +20 pts), consider closing/hedging (short spreads hurt from IV spike).

• If realized volatility > expected and underlying moves strongly against you, cut losses.

E. Price action exit

• If price closes beyond long leg (i.e., spread is in the money and close to max loss) — consider closing or rolling.

F. Roll rules

• If trade moves against you but you still believe in mean reversion and maximum risk budget allows:

  • Roll down/up (add credit) to widen spread, or roll to later expiry for debit — but only if roll reduces risk or results in acceptable new R/R. Keep roll count limited (e.g., max 1–2 rolls).

• If underlying breaks structural support/resistance or a new trend emerges, avoid rolling — exit.

G. Portfolio-level risk kill-switch

• If portfolio loss in a day > X% (e.g., 3–5% equity), close new entries or close most risk to preserve capital.

---

Risk Management & Position Sizing

• Use fixed % of equity per trade (1–2% risk per trade). Calculate dollar risk = (width of spread − credit received) × contract size × #contracts.

• Limit total margin exposure to e.g., 10–15% of equity.

• Correlation control: avoid concentrated exposure to same sector/direction; control portfolio delta near neutral.

• Max simultaneous losers: e.g., if 3 trades lose beyond stop, pause strategy for the day.

• Use alerts & automated order cancels for expiries/assignments.

---

Pseudocode (high-level)

for each tradable_symbol in universe:
  if not passes_liquidity(tradable_symbol): continue
  compute RSI14, EMA20, price_zscore, IVRank, days_to_earnings

  if days_to_earnings <= 14: skip

  if RSI >= RSI_overbought:
    if price_above_EMA and price_zscore >= Z_high and IVRank >= IV_min:
      short_leg = find_strike(delta=target_short_delta_call)
      long_leg = short_leg + width    # credit_spread
      if spread_liquid_and_credit_acceptable:
        enter_call_credit_spread(symbol, short_leg, long_leg, contracts_per_risk)

  if RSI <= RSI_oversold:
    if price_below_EMA and price_zscore <= Z_low and IVRank >= IV_min:
      short_leg = find_strike(delta=target_short_delta_put)
      long_leg = short_leg - width
      enter_put_credit_spread(...)

# Manage open positions every bar:
for each open_position:
  if profit >= profit_target_pct * max_profit: close_position()
  if loss >= loss_stop_pct * max_loss: close_position()
  if imminent_expiry or IV_spike or price_touch_long_leg: close_position_or_roll()
  enforce_portfolio_risk_kill_switches()

---

Decision Matrix Diagram (simple, actionable)

Below is a compact decision table you can implement as a hard rule-check before order placement.

Condition	RSI ≥ 75	RSI 70–75 (weak)	RSI 30–70 (neutral)	RSI 25–30 (weak oversold)	RSI ≤ 25
Price vs 20EMA	Above by > Z_high	Above moderately	Around EMA	Below moderately	Below by > Z_high
IVRank ≥ 40	SELL CALL SPREAD (if price & liquidity OK)	Consider small size / skip	NO TRADE	Consider small PUT spread only w/ divergence	SELL PUT SPREAD
IVRank < 40	Consider skipping (premium poor)	Skip	NO TRADE	Skip	Avoid (premium poor)
Earnings within 14d	Abort / Skip	Skip	Skip	Skip	Skip
Liquidity fail	Abort	Abort	Abort	Abort	Abort

Interpretation:

• Bold cell indicates primary action. Weak ranges (70–75 or 25–30) may be allowed only if divergence / candlestick rejection / extra confirmation present and position size reduced.

(If you prefer a graphical flowchart I can render a mermaid flowchart or canvas.)

---

Example (hypothetical trade)

• Underlying: XYZ at $100, RSI14 = 78, price = $106 (EMA20 = $98), price_zscore = +1.4, IVRank = 65, no earnings.

• Action: Sell 30-delta call credit spread: Short 1x 110 call, Long 1x 115 call (width $5). Credit received = $1.10.

• Max risk = $5 − $1.10 = $3.90 × 100 = $390.

• Profit target = 40% of max profit (max profit = $110) → target ≈ $44 (close when spread value ≤ $66).

• Stop loss = 100% of premium or when spread value ≥ $330 (or price closes above short strike by > $1.50 for 2 consecutive days).

---

Backtest & eval checklist

• Backtest on daily & intraday (15m) separately.

• Metrics: Annualized return, Sharpe, Sortino, max drawdown, win rate, average win/loss, return per unit of risk (e.g., return per max drawdown), avg days per trade, roll frequency, commissions/fees impact, slippage.

• Examine behavior across regimes (2008-like vol spike, 2017 low vol, 2020 pandemic) — short premium strategies behave differently in regime shifts.

• Include transaction costs, realistic fills (mid vs. bid/ask), and worst-case slippage scenarios.

• Forward-test (paper) live for at least 3 months and compare PnL drivers (IV changes vs. delta moves).

---

Implementation notes & practical tips

• Prefer automated entry/exit (API) with limit orders and pre-programmed stop/targets to avoid emotional errors.

• Use monitoring for IVRank & earnings calendar (automation to screen earnings prevents large losses).

• For intraday you can compress RSI period (9) and reduce DTE to same day or next-day options; increase profit targets because activity is noisier.

• Keep a log of every trade reason-code (which filter tripped) for later analysis.

• Consider hedging portfolio-level tail risk with a small allocation to long-dated puts or buying cheap protection when IV is very low and risk of crash increases.

---

Quick implementation checklist

1. Data: daily OHLCV, option chains (IV, delta, OI, vol rank), earnings calendar.

2. Precompute RSI14, EMA20, price_zscore, IVRank.

3. Candidate signals: mark symbols passing entry filters.

4. Find strikes by target delta or nearest strike and check credit.

5. Place limit spread order; set automated OCO rules (profit target & stop).

6. Monitor positions nightly (daily) or intraday loop.

7. Risk & portfolio checks before adding new trades.

---