Learn mean-reversion trading

Why does this work? Markets overshoot due to herd behavior and emotion. A bad earnings report causes panic selling below fair value. A hot product launch causes euphoric buying above fair value. The mean (average price) acts as gravity, pulling price back over time.

What is Mean-Reversion Trading?

Key Components:

• Mean (Average): Typically the moving average (MA) — e.g., 20-day MA, 50-day MA, or Bollinger Band midline.
• Deviation Measurement: How far price is from the mean (standard deviations, RSI, Bollinger Bands).
• Entry Signal: Price crosses below lower threshold → BUY. Price crosses above upper threshold → SELL/SHORT.
• Exit Signal: Price returns to the mean or crosses it.

WHY the mean? Historically, most stocks oscillate around a trend. If a stock normally trades at ₹100 and drops to ₹80 on no fundamental change, probability favors a return toward₹100.

Deriving the Mean-Reversion Logic

Step 1: Define the Mean

The mean is the expected value of price over a lookback period. For simplicity, use the Simple Moving Average (SMA):

where $P_t$ is price at time $t$, and $n$ is the lookback period (e.g., 20 days).

Why SMA? It smooths out noise and represents the "fair value" consensus over $n$ days.

Step 2: Measure Deviation

How far is current price $P_t$ from the mean?

But absolute deviation depends on asset price (₹10 deviation matters more for₹50 stock than ₹500 stock). So normalize:

where $\sigma_n$ is the standard deviation of price over $n$ days:

This $z$ is called the -score. It tells us how many standard deviations price is from the mean.

Why z-score?

• $z = 0$: Price at the mean (no trade).
• $z = -2$: Price is 2 std devs BELOW mean → oversold → BUY signal.
• $z = +2$: Price is 2 std devs ABOVE mean → overbought → SELL signal.

Step 3: Entry and Exit Rules

Entry: When $|z| > \text{threshold}$ (e.g., 2), enter opposite position:

• $z < -2$: BUY (expecting price to rise back to mean).
• $z > +2$: SELL/SHORT (expecting price to fall back to mean).

Exit: When price returns to mean ($z \approx 0$) or crosses it.

Why these thresholds? In a normal distribution, ~95% of data lies within 2 standard deviations. When $|z| > 2$, price is in the extreme 5%, statistically likely to revert.

Indicators for Mean-Reversion

1. Bollinger Bands

where $k = 2$ typically (can adjust to 2.5 or 3 for wider bands).

HOW to use:

• Price touches Lower Band → oversold → BUY.
• Price touches Upper Band → overbought → SELL.
• Price returns to Middle Band → EXIT.

WHY works? Bollinger Bands visualize the z-score concept. Bands expand during volatility (large $\sigma$) and contract during consolidation.

2. Relative Strength Index (RSI)

Typical $n = 14$ days.

Interpretation:

• RSI < 30 → oversold → BUY.
• RSI > 70 → overbought → SELL.
• RSI = 50 → neutral (no mean-reversion signal).

WHY? RSI compares upward vs downward momentum. Extreme RSI means one-sided pressure that exhausts itself, triggering reversal.

3. Z-Score (Direct Calculation)

As derived above. More flexible than RSI/Bollinger Bands because you control lookback period and threshold.

Worked Example 1: Bollinger Band Trade

Setup: Stock ABC trades at ₹100.20-day SMA = ₹100, 20-day $\sigma$ = ₹5. Bollinger Bands with $k=2$:

• Upper Band = 100 + 2(5) = ₹110.
• Lower Band = 100 - 2(5) = ₹90.

Day 1: Price drops to ₹88(below Lower Band).

• Why this step? Price at ₹88 is 2.4 std devs below mean (z = (88-100)/5 = -2.4). This is extreme. Historical data shows price returns to mean 90%+ of the time from here.
• Action: BUY 100 shares at ₹88. Risk: ₹2 per share (stop-loss at ₹86 if breakdown continues).

Day 5: Price rises to ₹98 (near Middle Band).

• Why exit? Mean-reversion assumes return to mean, not continuation. At ₹98, the rubber band has snapped back. Probability of further rise without new catalyst is lower.
• Action: SELL 100 shares at ₹98. Profit = (98-88) × 100 = ₹1,000.

Why not wait for₹110? Mean-reversion is not momentum. Target is the mean (₹100), not the opposite extreme. Overstaying invites reversal risk.

Worked Example 2: RSI Mean-Reversion

Setup: Stock XYZ at ₹200. RSI(14) = 25 (oversold).

Logic:

• RSI = 25 means recent losses outweigh gains heavily. Sellers exhausted.
• Why this step? When RSI < 30, historically 70% of the time price rebounds within 5-10 days (in ranging markets).
• Action: BUY at ₹200. Set target at ₹210 (where RSI returns to 50, the neutral zone).

3Days Later: RSI climbs to 52, price at ₹208.

• Why exit? RSI back to neutral = mean-reversion complete. No edge to hold longer.
• Action: SELL at ₹208. Profit = ₹8 per share.

Risk Management: If RSI drops further (e.g., to 15), it signals a breakdown, not mean-reversion. Stop-loss at ₹195 (if price makes new low despite oversold RSI).

When Mean-Reversion FAILS

Why this feels right: In a ranging market, oversold = opportunity. So you trained yourself: "Oversold = buy."

The Fix: Mean-reversion works in range-bound markets (sideways price action). In trending markets (strong uptrend or downtrend), "oversold" can stay oversold for months. Example: A stock drops from ₹500 to ₹400(RSI = 25). You buy. It drops to ₹300 (RSI still 25). You buy more. It drops to ₹200. You're wiped out.

How to avoid:

1. Check the trend first. Use 200-day MA: If price is below200-day MA and falling, trend is DOWN. Don't fight it with mean-reversion longs.
2. Use ADX (Average Directional Index). ADX > 25 = trending market. Skip mean-reversion. ADX < 20 = ranging market, safe for mean-reversion.
3. Set hard stop-losses. If price makes a new extreme LOW after you bought an oversold signal, exit. Don't double down.

When to Use Mean-Reversion Trading

Best Conditions:

• Range-bound markets: Price oscillates between support and resistance with no strong trend.
• High volatility: Large swings create frequent overbought/oversold extremes.
• Liquid stocks/indices: Tight spreads, low slippage. Examples: Nifty 50 stocks, large-cap US stocks.

Avoid When:

• Strong trends (up or down): Trend traders make money; mean-reversion traders lose.
• Low volatility: Small swings → infrequent signals → opportunity cost.
• News-driven moves: Fundamental change (merger, bankruptcy) → new mean; old mean irrelevant.

Mean-Reversion vs. Momentum Trading

Key Insight: They are opposite strategies. Use the right one for the market regime. In a bull market (momentum regime), don't mean-revert; ride the trend. In a chopy market (range regime), don't chase momentum; trade the bounces.

Example Trade Setup Checklist

1. Identify the mean: Plot20-day SMA or Bollinger Bands.
2. Measure extremes: Calculate z-score or check RSI.
3. Confirm range-bound market: Price between stable support/resistance, no strong trend, ADX < 20.
4. Entry: Price at extreme (z< -2, RSI < 30, or at lower Bollinger Band).
5. Position size: Risk 1-2% of capital (calculate from stop-loss distance).
6. Stop-loss: Below recent low (for longs) or above recent high (for shorts). If price breaks to new extreme, exit.
7. Target: The mean (SMA, middle Bollinger Band, or RSI = 50).
8. Exit: At target or if market regime changes (trend emerges).

Stock prices are like the yo-yo. Sometimes people panic and sell stock way too cheap (yo-yo goes too low). Sometimes people get excited and buy it way too expensive (yo-yo flies too high). But the "string" is the average price, and the stock usually comes back to that average.

Mean-reversion traders are like kids who catch the yo-yo at the bottom and sell it when it comes back to the middle. They don't wait for it to go to the top — that's risky! They just want the easy bounce back to normal.

The trick? Make sure the yo-yo is attached to a string! If the company is actually going bankrupt, there's no string — the yo-yo just falls and never comes back. So you check: Is this a normal bounce, or is something really broken?

"Don't Fight the Trend, Trade the Range"

• If ADX > 25, trend is boss. No mean-reversion.
• If ADX < 20, range is king. Mean-revert away!

Connections

Related Notes:

• 4.1.01-Trading-vs-investing-core-difference — Mean-reversion is a trading strategy (short-term), not investing.
• 4.1.05-Learn-momentum-trading — The opposite of mean-reversion; know when to switch strategies.
• 4.1.08-Learn-technical-indicators — RSI, Bollinger Bands, ADX all used in mean-reversion.
• 4.2.01-Risk-management-position-sizing — Crucial for mean-reversion: tight stops because you're betting against the current move.
• 3.2.04-Support-and-resistance-levels — Mean-reversion works best when price bounces off support/resistance in a range.

Broader Context:

• Mean-reversion is a statistical edge, not a certainty. Markets are mean-reverting 60-70% of the time in ranges, not 100%.
• Combines well with options strategies (e.g., sell puts when stock oversold, collect premium as it reverts).

#flashcards/stock-market

What is the core assumption of mean-reversion trading? :: Asset prices fluctuate around a long-term average (mean), and extreme deviations are temporary. Prices tend to return to the mean over time.

What indicators are best for mean-reversion trading? :: Bollinger Bands, RSI, z-score (direct calculation), and ADX (to confirm ranging market). All measure deviation from mean and identify extremes.

Concept Map

Hinglish (regional understanding)