Compression Breakout on S&P 500: Kaufman’s Best-Kept Secret

Most breakout strategies chase price above a recent high. Kaufman’s Compression Breakout is different — it first demands evidence that volatility has contracted, then waits for the subsequent expansion. Buy only after the market has gone quiet and then woken up. The premise: a market in compression is coiling energy; the breakout that follows tends to be directional and sustained rather than random noise.

Our backtest of 600 parameter combinations across S&P 500 data identified the best configuration: ATR Short=10, ATR Long=20, Compression Threshold=0.70, Breakout Lookback=10, Max Hold=20 days, no MA filter — CAGR 18.93%, Sharpe 0.98, Max Drawdown −5.73%. These are out-of-sample test results. Here is the full analysis.

What Is the Compression Breakout?

Perry Kaufman describes volatility compression as a period where short-term volatility falls to a fraction of long-term volatility — a market “holding its breath.” The Compression Breakout system, drawn from Kaufman’s systematic work, identifies this state and enters when price breaks out of the compressed range.

The four-component logic:

1. Compression detection: Divide short-term ATR by long-term ATR. If the ratio falls below the Compression Threshold (e.g., 0.70), the market is in compression — short-term volatility is 70% or less of its longer-term baseline.
2. Breakout confirmation: Once in a compression state, wait for price to close above the highest high of the last N days (Breakout Lookback). This is the expansion signal.
3. Entry: Buy on the close of the breakout bar.
4. Exit: Hold for a fixed number of days (Max Hold) or exit on MA filter signal if enabled.

The key insight: not all breakouts are equal. A breakout from a compressed market has higher directional conviction than a breakout from normal or elevated volatility. Compression acts as a filter — it selects for the subset of breakouts most likely to initiate genuine trends.

Backtest Setup

Walk-forward evaluation is the critical methodological choice here. The dataset is split into sequential train and test periods — parameters are selected on the training period, then applied to the unseen test period. This prevents look-ahead bias and tests whether the patterns discovered in-sample generalize to new data. Results reported below are out-of-sample (test period) only.

Top 10 Configurations by Out-of-Sample Sharpe Ratio

Green row = #1 ranked configuration by out-of-sample Sharpe. All results are out-of-sample (test period). CAGR and drawdown values are annualized on the test period only.

The Best Configuration: ATR 10/20, Threshold 0.70, Breakout 10, Hold 20

The top-ranked configuration combines the highest Sharpe Ratio (0.98) with the lowest maximum drawdown (−5.73%) in the entire result set — an unusual combination of strong returns and exceptional risk control.

A Sharpe of 0.98 approaches the 1.0 threshold that systematic traders typically consider excellent. The maximum drawdown of −5.73% is exceptional — the system takes small, frequent losses rather than catastrophic ones. Most professional hedge funds targeting equity-like returns accept drawdowns of 15–25%; this configuration’s worst period is just 5.73%.

The 15 out-of-sample trades reflect the strategy’s selectivity: it only fires when both conditions are met — compression must be confirmed (ATR 10/ATR 20 ratio below 0.70) and then a 10-day high must be breached. Most days, neither condition holds. The system waits.

The Compression Threshold: Why 0.70 Outperforms 0.80

The data reveals a meaningful difference between compression threshold = 0.70 and 0.80:

• Threshold 0.70 (stricter): Short-term ATR must be 70% or less of long-term ATR to qualify as compressed. This produces fewer but higher-quality signals. The best 0.70 result achieves Sharpe 0.98 with only 15 trades and −5.73% max drawdown.
• Threshold 0.80 (looser): Short-term ATR must only be 80% of long-term ATR. This triggers more frequently — the best 0.80 result produces 44 trades with CAGR 30.34% but Sharpe 0.81 and max drawdown −25.31%.

The tradeoff is clear: tighter compression thresholds produce fewer, cleaner signals with better drawdown profiles. Looser thresholds generate more trades with higher raw CAGR but substantially higher volatility and drawdown. For risk-adjusted performance (Sharpe), the 0.70 threshold wins.

This is consistent with the underlying theory: the rarer the compression state, the more meaningful the breakout signal. When the market has genuinely compressed to 70% of its normal volatility range, the subsequent breakout tends to be directional. At 80%, the signal is more common and more prone to false breaks.

Max Hold Period: The Patience Parameter

The Max Hold parameter determines how long the system stays in a position before exiting regardless of price action. The data shows a consistent pattern: longer hold periods improve performance.

• Hold=20 days (best): CAGR 18.93%, Sharpe 0.98 — allows winning trades to run
• Hold=15 days: CAGR 7.12%, Sharpe 0.74 — exits too early, leaves gains on table
• Hold=10 days: CAGR 17.76%, Sharpe 0.55 — more trades, more drawdown
• Hold=5 days: Low Sharpe across all variants — insufficient time for trends to develop

The 20-day hold period (one trading month) allows genuine compression breakouts to develop fully. Volatility compression preceding a breakout is a precursor to directional moves that typically last several days to weeks; cutting positions at 5 or 10 days systematically exits before the move completes.

Why Compression Precedes Major Moves

The academic and practitioner literature offers two explanations for why volatility compression predicts subsequent directional expansion:

1. Information Accumulation

During compression periods, informed market participants accumulate positions without revealing their hand — low volatility reflects disagreement resolution happening quietly in the order book. The eventual breakout reflects the accumulated directional pressure being released. This “quiet before the storm” dynamic has been documented in equity markets, FX, and futures.

2. Option Market Dynamics

Low volatility compresses option premiums. When options are cheap, large institutional players can position for directional moves with limited capital at risk, further building the directional overhang that eventually breaks the compression. This self-reinforcing mechanism is particularly relevant in S&P 500 markets where options markets are highly liquid and heavily traded.

3. Kaufman’s Efficiency Framework

Kaufman frames compression through the lens of market efficiency: a compressed market is one that has temporarily exhausted directional information. When new information arrives (earnings, macro data, policy change), the transition from compressed to directional is sharper and more sustained than a breakout from normal volatility — because there is no residual noise to absorb the move.

Walk-Forward Validation: Why This Matters

Many strategy backtests report in-sample results that collapse in real trading. Walk-forward evaluation (WFE) is the industry standard for avoiding this overfitting trap. The methodology used here:

1. Parameters are optimized on the training period only.
2. The single best parameter set from training is applied — without adjustment — to the unseen test period.
3. All results reported above are from the test period exclusively.

This means the 18.93% CAGR and 0.98 Sharpe were not seen when selecting parameters. They represent genuine out-of-sample performance. This does not guarantee future results, but it is far more predictive than in-sample optimization alone.

Of 600 total combinations tested, 75 achieved positive CAGR with at least 5 trades in the test period. The top configurations cluster around ATR 10/20 with compression threshold 0.70–0.80, suggesting the parameter relationships are robust rather than a single lucky combination.

Practical Implementation Notes

• Entry mechanics: Calculate 10-day ATR and 20-day ATR daily. When 10d/20d ratio drops below 0.70, enter compression watch mode. Track the 10-day rolling high. Buy on close when price exceeds that high.
• Exit mechanics: Exit at close on day 20 after entry (no earlier unless a hard stop is added independently).
• Signal frequency: Expect approximately 1–2 signals per month at most. Long periods of no signal are normal and expected — the system waits for the specific conjunction of compression + breakout.
• Position sizing: The system’s low drawdown (−5.73%) makes it amenable to higher position sizing than typical trend-following. Full position on each signal is defensible given the Sharpe profile.
• Instrument: Tested on S&P 500 index (^GSPC). Applicable to SPY, ES futures, or equivalent instruments with similar liquidity.

Comparison to Other Strategies Tested

The Compression Breakout stands apart from other strategies in our S&P 500 series — it is the only one to combine above-market CAGR with a Sharpe above 0.90:

• Compression Breakout (ATR 10/20, 0.70, N=10, Hold=20): CAGR 18.93%, Sharpe 0.98 — top performer
• Dual Momentum: CAGR 9.29%, Sharpe 0.598 — strong but lower Sharpe
• Stochastic Oscillator (K=5/25/85): CAGR 7.23%, Sharpe 0.532
• Keltner Channel (ATR×2.0): CAGR 5.84%, Sharpe 0.401
• Parabolic SAR (AF 0.01/0.10): CAGR 4.30%, Sharpe 0.420

The caveat: the Compression Breakout’s test period produced only 15 trades. A Sharpe ratio calculated on 15 trades has wider confidence intervals than one calculated on 192 trades (Parabolic SAR) or 1 trade (buy-and-hold [2]). More trades in the test period would narrow the uncertainty around these estimates.

Important Caveats

1. Small out-of-sample sample size. 15 trades in the test period limits statistical confidence. The Sharpe ratio of 0.98 is a point estimate with meaningful uncertainty bands. A strategy with 100+ trades would provide more robust statistics.
2. Walk-forward does not eliminate all bias. If the training and test periods share common market regimes, the apparent out-of-sample result may still reflect shared structural features. A truly independent holdout period (e.g., post-2024 data) would provide the strongest validation.
3. No transaction costs or slippage. The strategy generates approximately 15 round-trip trades in the test period. At $5–10/trade on a standard retail account, costs are minimal. For smaller accounts or illiquid instruments, costs matter more.
4. Fixed hold exit only. The strategy exits on day 20 regardless of price action. Incorporating a dynamic exit (trailing stop, volatility-adjusted) could improve or worsen performance depending on implementation.
5. Parameter sensitivity not fully explored. The grid tested ATR periods of 5 and 10 for short-term, and 20 for long-term. Other values (e.g., ATR 7/14, ATR 10/30) were not tested and may yield different results.
6. Index-only test. Performance on individual equities, sector ETFs, international indices, or commodities may differ substantially from these S&P 500 results.

Practical Takeaway

Kaufman’s Compression Breakout identifies a specific, rare market state — volatility compressed to 70% of its normal range — and enters only when that state resolves into a directional breakout. The out-of-sample results from this parameter study are the strongest we have found in the S&P 500 strategy series: CAGR 18.93%, Sharpe 0.98, maximum drawdown −5.73%.

The strategy’s power comes from its patience. It does not trade constantly; it waits for the conjunction of two specific conditions. Most days, it does nothing. When it fires, 60% of those entries are profitable with a max hold of 20 days. The result is a system that earns high returns per unit of risk — the definition of a high-quality systematic strategy.

The primary uncertainty is sample size: 15 out-of-sample trades is a thin statistical base. Traders interested in this approach should paper-trade or test with small size across multiple instruments to accumulate more signal observations before committing capital at scale.

References

1. RakoQuant (2023). KAMA SD Bands. TradingView. Link
2. Kavout Team (2024). Introducing Kavout’s AI Trade Spotter: Uncover Hidden Opportunities in Real Time. Kavout Market Lens. Link
3. NASA (2026). CALIPSO Final Report. NASA Technical Reports Server. Link
4. McKinsey & Company (2025). McKinsey on Investing Issue 11. McKinsey. Link
5. Nguyen et al. (2024). Reinforcement Learning from Human Feedback. RLHF Book. Link

Related Reading

• DCA Strategy for Beginners [2026]
• Basic Car Maintenance Everyone Should Know: Beginner Guide [2026]
• Mel Robbins 5-Second Rule: 3 Studies Prove Why It Works [2026]

Frequently Asked Questions

What is Compression Breakout: Kaufman Best-Kept Secret?

This article covers the evidence-based aspects of Compression Breakout: Kaufman Best-Kept Secret.

Why does this matter?

Understanding the topic helps make informed decisions backed by research.

What does the research say?

See the References section above for peer-reviewed sources.

Related Posts

• Parabolic SAR on S&P 500: 9 Combinations Tested
• The Dunbar Number: Why You Can Only Maintain 150 Friendships
• Hedonic Adaptation: Why Getting What You Want Never Makes