Skip to main content

Grid Trading and the Fragility of Stability

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

Grid trading looks like easy money.

Place buy orders below the current price. Place sell orders above. As price oscillates, orders fill. Profit accumulates.

It's mean reversion made mechanical.

Until it isn't.

What Grid Trading Assumes​

Grid trading is built on a simple assumption: prices will oscillate within a range.

You create a grid of orders:

  • Buy orders at regular intervals below the current price
  • Sell orders at regular intervals above the current price
  • As price moves up and down, orders fill
  • You profit from the oscillations

The strategy works beautifully when:

  • Volatility is stable
  • Prices mean-revert
  • The distribution doesn't change

Sound familiar?

This is exactly the kind of linear, stability-assuming strategy that breaks during volatility regime shifts.

Grid Trading is Concave, Not Convex​

Recall from Part II: mean reversion is concave.

Concave strategies:

  • Small wins most of the time
  • Occasional large losses
  • Optimized for stability
  • Break during regime shifts

Grid trading fits this perfectly.

Normal times:

  • Price oscillates within your grid
  • Orders fill on both sides
  • Small profits accumulate
  • High win rate feels good

Volatility spike:

  • Price moves in one direction
  • All buy orders fill (or all sell orders)
  • Price doesn't revert
  • Large unrealized losses accumulate
  • Grid breaks

This is the fragility we've been talking about.

The Volatility Problem​

Grid trading uses fixed intervals.

If you place orders every $10, you're assuming volatility stays constant. When volatility doubles, your grid spacing becomes half as effective.

Example:

You're trading a stock at $100. You place:

  • Buy orders at 95,95, 90, 85,85, 80
  • Sell orders at 105,105, 110, 115,115, 120
  • Grid spacing: $5 intervals

Volatility is stable at 2% daily moves. Your grid works.

Then volatility spikes to 8% daily moves.

Your $5 intervals are now too tight. Price gaps through your grid. All your buy orders fill in one day. Price keeps dropping. You're stuck with large positions and no way out.

This is Post I in action: Volatility spikes aren't "high volatility" - they're distribution changes. Your grid assumes stability. When the distribution changes, your grid breaks.

Why Grid Trading Feels Safe​

Grid trading appeals because it feels safe:

  1. High win rate - Most trades are small wins
  2. Mechanical - No emotion, just execute
  3. Backtests look great - In stable regimes, grids print money
  4. Seems like arbitrage - Buying low, selling high

But this is the trap.

High win rate strategies are usually concave. They feel good until they don't. During volatility spikes, they implode.

Backtests assume stability. They show beautiful equity curves right up until the day they don't. Grid trading backtests are particularly misleading because they optimize for the last regime.

It's not arbitrage. Arbitrage assumes no risk. Grid trading has massive tail risk - it just doesn't show up in normal times.

A Real Example: Grid Trading During March 2020​

Let's make this concrete.

The Setup:

You're grid trading SPY in early 2020:

  • Current price: $3,000
  • Grid spacing: $30 (1% intervals)
  • Buy orders: 2,970,2,970, 2,940, 2,910,2,910, 2,880
  • Sell orders: 3,030,3,030, 3,060, 3,090,3,090, 3,120
  • Position size: $10,000 per grid level

Volatility has been stable. Your grid has been profitable. Life is good.

March 9th, 2020:

SPY drops 7.6% in one day. From 3,000to3,000 to 2,772.

What happens to your grid:

  • All your buy orders fill: 2,970,2,970, 2,940, 2,910,2,910, 2,880
  • Price gaps through your entire grid
  • You're now long $40,000 worth of SPY
  • Price is at $2,772 - below your lowest buy order
  • Unrealized loss: ~$8,000

The problem:

Your grid assumed prices would oscillate. They didn't. They moved in one direction and kept going.

Your fixed intervals assumed stable volatility. Volatility spiked. Your grid spacing became useless.

Your mean reversion assumption broke. Prices didn't revert. They continued down.

This is Post I, II, and III combined:

  • Volatility spike = distribution change (Post I)
  • Concave strategy broke (Post II)
  • No convexity to protect you (Post III)

The Deeper Problem​

Grid trading optimizes for the last regime.

When volatility is stable, grids work. When prices mean-revert, grids work. When distributions don't change, grids work.

But markets don't stay stable forever.

The trap: Grid trading feels safe because it works most of the time. But "most of the time" isn't enough. You need strategies that survive regime shifts.

The reality: Grid trading is a high win rate, low Sharpe strategy. It prints money until it doesn't. When it breaks, it breaks hard.

Practical Implications​

If you're using grid trading, here's what you need to know:

Recognize the assumptions:

  1. You're assuming stable volatility
  2. You're assuming mean reversion
  3. You're assuming the distribution won't change
  4. You're optimizing for the last regime

Make grids volatility-adaptive:

Don't use fixed intervals. Scale your grid spacing with volatility:

  • Low volatility = tighter spacing
  • High volatility = wider spacing
  • Monitor for volatility spikes and widen spacing proactively

Add convexity:

Grid trading is concave. Add convexity to protect against regime shifts:

  • Combine grids with long options
  • Use options to hedge tail risk
  • Accept that pure grid trading is fragile

Monitor for distribution changes:

When volatility spikes, exit your grid. Don't assume reversion. Distribution changes mean your assumptions are broken. Recalibrate before re-entering.

Accept the fragility:

Grid trading works until it doesn't. During normal times, it underperforms convex strategies. During regime shifts, it breaks. This is the trade-off.

The Meta-Lesson​

Grid trading is a perfect example of optimizing for the last regime.

It feels safe because:

  • High win rate
  • Mechanical execution
  • Backtests look great
  • Seems like easy money

But it's fragile because:

  • Assumes stability
  • No convexity
  • Breaks during regime shifts
  • Large tail risk

Most traders love grid trading because it feels safe. It has a high win rate. It's mechanical. It seems like arbitrage.

But volatility spikes don't reward comfort. They reward convexity. They reward strategies that survive distribution changes.

Grid trading is the opposite of what survives volatility spikes.

What to Do Instead​

If you want to trade mean reversion, consider:

  1. Volatility-adaptive grids - Scale spacing with volatility
  2. Add convexity - Combine with long options for protection
  3. Monitor regime changes - Exit when volatility spikes
  4. Accept lower win rates - Convex strategies have lower win rates but survive regime shifts

Or, recognize that mean reversion is concave and accept that you're trading a fragile strategy. During normal times, you'll outperform. During regime shifts, you'll break.

The question is: Can you handle breaking when volatility spikes?

Most grid traders can't. They quit after the first big loss. The ones who stick around learn the hard way that stability is fragile.

The Thread That Connects Everything​

Grid trading demonstrates all four concepts:

  1. Volatility is not risk - Grid trading uses fixed intervals, assuming stable volatility. When volatility spikes, grids break.

  2. Convexity survives - Grid trading is concave. It works until volatility spikes, then it breaks hard.

  3. Options are volatility instruments - Grid trading focuses on price direction, ignoring volatility. Adding options adds convexity.

  4. Real examples matter - March 2020 showed exactly how grid trading breaks during volatility spikes.

The pattern is consistent: Strategies that assume stability break when distributions change.

Grid trading is just another example of this pattern.

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

Previous: Part IV - A Real Trading Example