iPresageCPI Day Options: A 5-Year Backtest Analysis
By iPresage Research · 9 min read · 2025-02-03
Backtesting 60 CPI releases across 5 years reveals post-release diamond signals win 63.7% of the time, with optimal entry 15-45 minutes after the 8:30 AM print.
Consumer Price Index releases have become the single most important recurring data point for options markets. Since 2020, when inflation became the dominant macro narrative, CPI day has consistently produced outsized moves, extreme volatility crushes, and predictable patterns in options flow. Using iPresage's dataset of 159 million historical options records, we backtested 60 CPI release days from January 2020 through December 2024 to quantify the opportunity and map it against our scanner's signal accuracy.
This report presents our findings across three dimensions: the structural impact of CPI on implied volatility, the directional performance of iPresage diamond signals on CPI day, and a set of practical strategies derived from the data.
THE CPI VOLATILITY PREMIUM
Every CPI release is preceded by a measurable volatility premium. We calculated the excess implied volatility -- defined as the difference between actual IV and what a non-event day IV model would predict -- for SPY at-the-money options in the five days leading up to each of the 60 CPI releases in our sample.
The average excess IV on the day before CPI was 2.4 volatility points. By comparison, the average excess IV before FOMC meetings was 1.8 points, before NFP was 1.1 points, and before GDP releases it was 0.6 points. CPI has the largest pre-event volatility premium of any recurring economic release, surpassing even the Fed.
This premium is not constant. It has grown over time. In 2020, the average pre-CPI excess IV was 1.3 points. By 2022, when inflation dominated markets, it reached 3.8 points. In 2024, as inflation stabilized, it moderated to 2.1 points. The premium tracks the market's perception of CPI's importance, and while it has come down from its peak, it remains elevated relative to the pre-2020 baseline of approximately 0.4 points.
The volatility crush on CPI day itself averaged 2.8 points across the full sample, with a range of 0.4 to 7.2 points. The largest crush occurred on November 10, 2022, when the cooler-than-expected October CPI print triggered a massive equity rally and one of the most dramatic single-day volatility unwinds in recent history.
CPI DAY DIRECTIONAL PATTERNS
CPI releases cluster into three categories based on the deviation from consensus expectations: hot (above consensus by 0.1% or more), in-line (within 0.1%), and cool (below consensus by 0.1% or more). Of our 60 releases, 22 were hot, 19 were in-line, and 19 were cool.
The directional patterns are stark. On hot CPI days, SPY closed lower 77.3% of the time (17 of 22) with an average move of -1.14%. On cool CPI days, SPY closed higher 84.2% of the time (16 of 19) with an average move of +1.37%. In-line prints produced modest moves with no directional bias: the average absolute move was 0.48% with a 52.6% positive close rate.
These directional tendencies are well-known. What is less well-known is how they interact with options flow and scanner signals.
IPRESAGE DIAMOND SIGNAL PERFORMANCE ON CPI DAY
We identified 847 diamond signals generated across the 60 CPI days in our sample. We split these into pre-release signals (generated before the 8:30 AM EST data release) and post-release signals (generated after 8:30 AM).
Pre-release signals, of which there were 312, had an overall win rate of 49.1% and an average EV score of +0.3. This is statistically indistinguishable from random. The information content of pre-release positioning on CPI day is effectively zero -- the data release overwhelms any signal generated before it.
Post-release signals tell a dramatically different story. The 535 post-release signals posted an overall win rate of 63.7% and an average EV score of +5.8. This represents the highest single-day signal performance in our entire database outside of the March 2020 crash recovery.
Breaking post-release signals down further by CPI outcome category:
After hot CPI prints, bearish diamond signals posted a 71.4% win rate across 148 signals with an average EV of +7.2. The scanner excels at identifying the capitulation selling that follows an above-consensus inflation print.
After cool CPI prints, bullish diamond signals posted a 73.1% win rate across 131 signals with an average EV of +8.1. The relief rally following a benign CPI print creates strong flow dislocations that the scanner identifies effectively.
After in-line prints, signal performance was more modest: 55.8% win rate and +2.3 average EV across 256 signals. In-line prints produce less dramatic re-pricing, which means fewer dislocations for the scanner to exploit.
SECTOR-LEVEL ANALYSIS
CPI does not affect all sectors equally. We tracked diamond signal performance across the 11 GICS sectors on CPI day and found significant dispersion.
Technology stocks (represented by XLK) showed the highest sensitivity. Post-release diamond signals on tech names posted a 67.4% win rate on CPI days compared to 53.8% on non-CPI days. The excess performance of 13.6 percentage points was the largest of any sector.
Financials (XLF) were the second most responsive, with a 64.9% CPI-day win rate versus 54.2% on normal days. Banks and financial firms benefit from the direct link between CPI, rate expectations, and yield curve dynamics.
Utilities (XLU) and Consumer Staples (XLP), traditionally defensive sectors, showed the least CPI sensitivity. Their post-release signal win rates of 56.1% and 55.7% were only marginally above non-CPI baselines.
For individual stocks tracked by iPresage, the highest CPI sensitivity was found in MSFT (69.2% post-release win rate, 38 signals), AAPL (67.8%, 42 signals), and JPM (66.4%, 31 signals). These names combine high liquidity with direct sensitivity to the rate-expectation channel through which CPI affects equity valuations.
THE STRADDLE SELL STRATEGY
Given the predictable volatility crush that follows CPI, a natural strategy is selling straddles before the release and closing them after. We backtested selling SPY at-the-money straddles at the close on T-1 (the day before CPI) and closing at the close on CPI day.
Over 60 instances, this strategy produced a positive return 41 times (68.3% win rate) with an average return of +1.8% on notional and a median return of +1.4%. The maximum loss was -6.3% (June 2022, when a surprise hot print triggered a 2.9% SPY decline), and the maximum gain was +5.1% (November 2022, the landmark cool print).
When filtered by the iPresage regime classification at the time of the trade, the results improve. In SURGING regimes (28 instances), the win rate was 75.0% with an average return of +2.4%. In NEUTRAL regimes (18 instances), it was 66.7% and +1.5%. In DRAINING regimes (14 instances), it was 57.1% and +0.6%.
The regime filter is particularly valuable for risk management. DRAINING regimes are associated with wider distributions of CPI-day outcomes and larger absolute moves, making the straddle sell strategy riskier even though the volatility premium is typically larger.
TIMING THE POST-RELEASE ENTRY
We tested multiple entry windows for post-CPI directional trades triggered by diamond signals. The results suggest that the optimal entry window is 15 to 45 minutes after the 8:30 AM release.
Signals generated within the first 15 minutes after CPI had a win rate of 58.2%, below the 63.7% average for all post-release signals. The initial reaction is noisy and subject to reversal. Signals generated 15 to 45 minutes post-release had the highest win rate at 68.9% and the highest average EV at +7.1. By this point, the market has digested the data, institutional flow has established a direction, and the scanner can identify high-conviction dislocations.
Signals generated more than 45 minutes after the release but still within the first trading hour maintained a solid 62.3% win rate. Beyond the first hour, performance decayed toward the non-CPI baseline.
This timing dynamic suggests that patience pays on CPI day. Waiting for the initial volatility to settle before acting on scanner signals materially improves outcomes.
INTERACTION WITH THE FED CYCLE
We examined whether CPI's impact on options markets varies depending on proximity to the next FOMC meeting. When CPI falls within 10 trading days of an FOMC announcement (which occurred 14 times in our sample), the volatility premium is amplified and the post-release signal performance is enhanced.
In these CPI-near-FOMC instances, the post-release diamond signal win rate was 69.8% with an average EV of +7.4, compared to 61.9% and +5.2 for CPI releases that were more than 10 days from the nearest FOMC meeting. The compounding of event risk creates larger dislocations and more actionable signals.
KEY FINDINGS AND PRACTICAL FRAMEWORK
Based on 60 CPI releases and 847 diamond signals, we offer the following framework for CPI day trading:
This analysis is retrospective and based on historical data. CPI sensitivity has varied over time and may continue to evolve as the inflation cycle matures. All options trading involves risk of loss. The strategies discussed are for educational and research purposes and do not constitute investment advice.