Skip to main content

A Real Trading Example: When Theory Meets Reality

· 5 min read
Calvin Cheng
Calvin Cheng
Financial Markets in an AI World

Theory is clean. Markets are messy.

Let's walk through a real example that ties together everything from the first three posts.

The Setup: March 2020​

It's early March 2020. The S&P 500 has been trading in a tight range for months. Realized volatility is low - around 12-15%. ATR suggests daily moves of 1-2%. Everything looks stable.

Two traders approach this differently:

Trader A (Linear Thinking):

  • Uses ATR-based position sizing
  • Fixed 2% stops
  • Mean reversion strategy
  • High win rate, small wins

Trader B (Convex Thinking):

  • Volatility-scaled position sizing
  • Long options for convexity
  • Trend-following bias
  • Lower win rate, asymmetric payoffs

Both see the same market. Both have the same information. But their frameworks are fundamentally different.

The Distribution Change​

On March 9th, 2020, the S&P 500 drops 7.6% in a single day.

This isn't "high volatility" - this is a distribution change. The market structure has shifted. What worked yesterday won't work today.

Trader A's Framework Breaks:

  • ATR said daily moves should be 1-2%. Today's move is 7.6%. The model is useless.
  • Fixed 2% stops get hit immediately. Positions are closed at the worst possible time.
  • Mean reversion strategy assumes prices will revert. They don't. They keep going.
  • High win rate strategy becomes a high loss rate strategy overnight.

Trader B's Framework Adapts:

  • Volatility-scaled sizing means smaller positions as volatility increases. Risk is contained.
  • Long options provide convexity. The 7.6% move creates asymmetric gains.
  • Trend-following bias means the strategy is positioned for continuation, not reversion.
  • Lower win rate doesn't matter - the asymmetric payoff compensates.

The Numbers​

Let's make this concrete.

Trader A: Mean Reversion Strategy

  • Position: Short SPY futures, sized based on ATR
  • Entry: $3,000 (SPY price)
  • Stop: $3,060 (2% stop)
  • Expected move: 1-2% based on ATR

When SPY drops to $2,772 (7.6% down):

  • Stop gets hit immediately (if not before)
  • Position closed at worst possible time
  • Loss: 2% of position size
  • But the real loss: missing the continued move down

Trader B: Convex Options Strategy

  • Position: Long SPY put options, sized for volatility
  • Entry: SPY 3,000,buying3,000, buying 2,900 puts (3% OTM)
  • IV: 15% (low, relative to what's coming)
  • Cost: $20 per contract

When SPY drops to $2,772:

  • Options are now deep ITM
  • Delta increases (convexity working)
  • Value: ~$128 per contract
  • Gain: 540% on the position
  • Risk was limited to premium paid

Why This Matters​

This isn't about being right or wrong. It's about surviving when the math changes.

Trader A's mistakes:

  1. Used historical volatility as risk measure - ATR said 1-2% moves. Reality said 7.6%. The framework broke.
  2. Fixed stops assumed stability - 2% stops work until they don't. During distribution changes, they fail.
  3. Mean reversion assumed reversion - But distributions can shift. What reverts in normal times doesn't revert during regime shifts.
  4. Optimized for win rate - High win rate feels good until volatility spikes. Then it becomes a liability.

Trader B's advantages:

  1. Traded volatility, not direction - Bought options when IV was low relative to expected realized vol. Direction was secondary.
  2. Used convexity - Long options provide asymmetric payoffs. Small losses most of the time, large gains during spikes.
  3. Accepted insurance premium - Lost money on options most days. But when volatility spiked, the payoff was asymmetric.
  4. Focused on payoff geometry - Lower win rate, but survived the regime shift.

The Aftermath​

Over the next two weeks, SPY continued to drop. Volatility spiked to 80%+.

Trader A:

  • Stopped out repeatedly
  • Mean reversion trades kept losing
  • Framework completely broken
  • Forced to sit out or change strategy mid-crisis

Trader B:

  • Options positions continued to gain
  • Convexity provided protection
  • Framework adapted to new regime
  • Could continue trading through the crisis

The Real Lesson​

This isn't about predicting March 2020. It's about building a framework that survives distribution changes.

Most traders optimize for the last regime. They use ATR, fixed stops, mean reversion - all assuming stability. When distributions shift, these frameworks break.

Convex traders accept that:

  • They'll lose money most of the time
  • Win rates will be lower
  • The framework feels wrong psychologically

But when volatility spikes - and it always does - they're the ones still standing.

Practical Takeaways​

If you're using linear frameworks:

  1. Recognize the assumptions - you're assuming stability
  2. Widen stops or make them volatility-adaptive
  3. Reduce position sizes during volatility spikes
  4. Don't trust backtests that assume stable distributions

If you're building convexity:

  1. Accept the insurance premium - you'll lose money most days
  2. Size positions for volatility, not fixed amounts
  3. Use options for convexity, not leverage
  4. Monitor IV vs realized vol, not just price direction

The meta-lesson:

Most traders quit convex strategies before the regime shift happens. They can't handle losing money repeatedly. The ones who stick around get paid when volatility spikes.

But here's the thing: you don't know when the spike is coming. So you have to pay the insurance premium every day, knowing most days you'll lose.

Most people can't do this. That's why the edge persists.

Part IV of the series: Trading What Breaks Linear Thinking.

Previous: Part III - Options Are Volatility Instruments | Next: Part V - Grid Trading and the Fragility of Stability