CVX next-day intraday range after extreme Brent volatility days (proxy: |daily return| > 90th pctile)
An extreme move in Brent crude doesn’t just end at the close. For CVX, the question is whether that volatility bleeds into the next session’s trading range — and the numbers say it often does. Using a daily absolute return proxy for Brent’s intraday swing, days past the 90th percentile were followed by an above-median CVX range 67.6% of the time, versus 48.0% after calmer Brent days. That’s an odds ratio of 2.26, with a one-sided Fisher p-value of 0.001.
This isn’t a forecast of direction or a call on the stock. It’s a measurement of how uncertainty in the crude market maps onto trading behavior in the supermajor a day later. The evidence below covers the full methodology, the distribution of next-day ranges, and the sensitivity of the finding — including the caveat that about one in three extreme Brent days still yields a quiet CVX session. The detail is in the analysis that follows.
For CVX, does a day when Brent crude's intraday high-low range (as % of open) exceeds its 90th percentile predict an above-median intraday range for CVX on the following trading day? Extreme oil swings create uncertainty, causing traders to adjust hedging and position sizes in the supermajor a day later.
How this was measured
Brent crude intraday data is unavailable in this runtime; daily absolute return (|close-to-close pct_change|) is used as a proxy for intraday high-low range. The 90th percentile of Brent's historical |return| is computed over the full available window. A day is classified as 'extreme' when Brent's |return| exceeds that threshold. The next trading day's CVX intraday range — defined as (day-high − day-low) / day-open — is extracted from CVX minute bars. The overall median of CVX next-day range is used to split days into above/below median. Fisher's exact test (one-sided) assesses whether the odds of an above-median CVX range are higher after an extreme Brent day. A Welch t-test on the continuous range values is also reported.
The key numbers
Reading the numbers
After a day when Brent's move is in the most extreme 10%, CVX's next-day range is above its median about 68% of the time, versus 48% after ordinary days. The p-value of 0.0010 says that gap is very unlikely to be just luck.
The charts
This histogram shows the 74 days that followed extreme Brent moves. The average next-day CVX range is 6.19% of the open, and the widest day reached 21.4%. Compare that average with the non-extreme group's 4.79%: on the whole, CVX does move more after oil shocks.
The 648 non-extreme Brent days produce a next-day CVX range averaging 4.79% of the open, clearly below the 6.19% after extreme days. There is one very large outlier at 114% of the open, but the average stays lower than the extreme-day group. So normal oil days generally do not push CVX around as much.
The left bar is the headline comparison: 67.6% of extreme Brent days are followed by an above-median CVX range, versus 48.0% after non-extreme days. A Fisher exact test p-value of 0.0010 means a gap like this would occur by chance only about once in a thousand trials, and the odds ratio of 2.26 says extreme days make an above-median CVX range about 2.3 times as likely.
Contingency table (extreme vs above-median next-day range)
| Brent day type | Above median | Below median |
|---|---|---|
| extreme | 50 | 24 |
| non-extreme | 311 | 337 |
The takeaway
Yes — at least for the daily-return proxy used, extreme Brent volatility does predict a wider-than-usual CVX range the next trading day. After the 74 extreme Brent days, CVX's next-day range was above its median 67.6% of the time (50 of 74), versus 48.0% after non-extreme days (311 of 648). The odds ratio is 2.26, and the one-sided Fisher p-value is 0.001, so this is a clear signal rather than a coin flip. The continuous range comparison also lands at p = 0.0315 — conventionally significant, though not overwhelming. Practical takeaway: an extreme Brent swing roughly doubles the odds of an above-median CVX range the next session, but about one in three extreme days still produces a quiet CVX day.
The fine print
- Brent intraday high-low range wasn't available; daily close-to-close absolute return is the proxy, which misses intraday reversals and doesn't exactly match true intraday range.
- The 90th percentile threshold was computed in-sample over the full window; rolling or out-of-sample thresholds could change which days count as extreme.
- Only 74 extreme Brent days — the effect is statistically significant, but the odds ratio of 2.26 has meaningful uncertainty.
- Next-day alignment assumes overlapping trading days; holidays can insert extra calendar days between a Brent event and CVX's response.