DVN vs XOP forward 10d after up-close / negative-last-30m with Brent +0.5% (N=96)
DVN closed higher, the last half hour of trading was weak, and Brent finished up at least half a percent. That combination showed up 96 times over roughly three years — and in the ten sessions that followed, DVN trailed XOP by an average of 0.67%, with a median gap of -0.62%.
The direction is consistent, but the benchmark is the problem. DVN already lagged XOP by about 0.20% on average across every day in the sample, so the setup adds only around 0.47 percentage points — a difference that doesn't clear statistical significance.
The full panel, test results, and charts are below.
Over the past ~3 years, when DVN closes up on the day but its last-30-minute return is negative while Brent crude closes up at least 0.5%, does DVN lag XOP over the next 10 trading sessions?
How this was measured
Built a daily panel from DVN and XOP RTH minute bars: daily close-to-close return, DVN last-30-minute return (first print at or after 15:30 ET to final RTH print), and Brent daily close-to-close return. Event condition required DVN daily return > 0, DVN last-30-minute return < 0, and Brent daily return >= 0.5%. Forward 10-session returns were computed as close[t+10]/close[t] - 1 for both tickers. The key outcome is diff_fwd10 = DVN_forward - XOP_forward; negative means DVN lagged XOP over the next 10 sessions. Panel covers 2023-10-02 to 2026-08-31. We compared event days against both zero and the unconditional/non-event daily differential using one-sample and Welch two-sample t-tests.
The key numbers
Reading the numbers
On the 96 event days, DVN trailed XOP by about 0.67% on average over the following 10 sessions, and it lagged in 58% of cases. The gap is statistically distinguishable from zero, but comparing event days with all other days shows no clear difference, so treat this as a lead worth watching rather than a proven pattern.
The charts
This histogram spreads out the 96 event days by how much DVN beat or trailed XOP over the next 10 sessions. The mass sits mostly left of zero, which is why the average lands at -0.67%, but the tails are wide, stretching from about -9.8% to +9.2%. That spread matters: the negative tilt is a lean, not a rule, and any single event can easily land on the opposite side.
These three bars compare the average DVN-minus-XOP gap for event days, all days, and non-event days. The event bar is the deepest at -0.67%, roughly triple the all-day figure of -0.20% and clearly below the non-event bar at -0.13%. So the event setup does line up with worse relative performance, though the modest separation is consistent with the event-versus-non-event test not reaching statistical significance.
Event-level 10-session forward differential
| event_date | dvn_daily_ret | dvn_last30_ret | brent_daily_ret | dvn_fwd10 | xop_fwd10 | dvn_minus_xop_fwd10 |
|---|---|---|---|---|---|---|
| 2023-10-09 | 0.0551 | -0.0015 | 0.0399 | 0.0347 | 0.0149 | 0.0198 |
| 2023-10-13 | 0.0385 | -0.0008 | 0.0671 | -0.0348 | -0.0307 | -0.0042 |
| 2023-11-17 | 0.0244 | -0.0038 | 0.0449 | -0.0152 | -0.016 | 0.0008 |
| 2023-11-20 | 0.0062 | -0.0052 | 0.025 | -0.041 | -0.0391 | -0.0019 |
| 2023-12-18 | 0.0094 | -0.0006 | 0.0267 | 0.0336 | 0.0135 | 0.0201 |
| 2024-01-22 | 0.0046 | -0.0013 | 0.0123 | -0.0119 | 0.006 | -0.0179 |
| 2024-02-07 | 0.0056 | -0.0016 | 0.0089 | 0.0591 | 0.0546 | 0.0045 |
| 2024-02-16 | 0.0069 | -0.0036 | 0.0065 | 0.0212 | 0.0198 | 0.0014 |
| 2024-03-28 | 0.0139 | -0.0026 | 0.0145 | 0.0658 | 0.0196 | 0.0462 |
| 2024-04-05 | 0.0171 | -0.0016 | 0.0276 | -0.0269 | -0.0458 | 0.019 |
| 2024-04-26 | 0.0027 | -0.001 | 0.021 | -0.0491 | -0.0404 | -0.0088 |
| 2024-05-02 | 0.001 | -0.0029 | 0.0151 | -0.0224 | 0.0083 | -0.0308 |
| 2024-05-09 | 0.0057 | -0.0004 | 0.0101 | -0.0515 | -0.0324 | -0.0191 |
| 2024-05-17 | 0.0083 | -0.0023 | 0.0072 | -0.0512 | -0.0324 | -0.0188 |
| 2024-06-17 | 0.0031 | -0.0014 | 0.0118 | 0.0453 | 0.0315 | 0.0138 |
| 2024-07-02 | 0.0008 | -0.0024 | 0.0198 | 0.0182 | 0.0124 | 0.0058 |
| 2024-07-31 | 0.0302 | -0.0006 | 0.0269 | -0.054 | -0.0613 | 0.0073 |
| 2024-08-07 | 0.0283 | -0.0099 | 0.0214 | 0.0155 | 0.0333 | -0.0178 |
| 2024-08-12 | 0.0091 | -0.0013 | 0.0245 | -0.0042 | 0.0133 | -0.0175 |
| 2024-08-15 | 0.0214 | -0.0013 | 0.0114 | -0.0095 | -0.0003 | -0.0092 |
| 2024-08-29 | 0.0094 | -0.0068 | 0.0189 | -0.1131 | -0.0808 | -0.0323 |
| 2024-09-12 | 0.002 | -0.0006 | 0.0498 | -0.0432 | 0.0122 | -0.0554 |
| 2024-09-17 | 0.0155 | -0.0044 | 0.008 | -0.0123 | 0.0184 | -0.0307 |
| 2024-09-19 | 0.0149 | -0.0044 | 0.0189 | 0.0198 | 0.0444 | -0.0246 |
| 2024-10-01 | 0.0164 | -0.0035 | 0.0408 | 0.0362 | -0.002 | 0.0382 |
| 2024-10-07 | 0.0059 | -0.0035 | 0.0305 | -0.0594 | -0.0527 | -0.0067 |
| 2024-10-10 | 0.0212 | -0.0014 | 0.031 | -0.0779 | -0.0346 | -0.0434 |
| 2024-10-11 | 0.0028 | -0.0041 | 0.0103 | -0.0912 | -0.0451 | -0.046 |
| 2024-10-22 | 0.0022 | -0.0033 | 0.0314 | 0.0082 | -0.0117 | 0.02 |
| 2024-10-30 | 0.0055 | -0.0022 | 0.0298 | 0.0198 | 0.0732 | -0.0534 |
| 2024-11-04 | 0.0222 | -0.0057 | 0.0171 | -0.0079 | 0.0898 | -0.0978 |
| 2024-11-18 | 0.0091 | -0.0017 | 0.0123 | -0.0327 | 0.0021 | -0.0348 |
| 2024-11-21 | 0.0065 | -0.0015 | 0.0102 | -0.0974 | -0.0755 | -0.0219 |
| 2024-12-30 | 0.0237 | -0.0054 | 0.0064 | 0.2069 | 0.1147 | 0.0922 |
| 2025-01-10 | 0.0239 | -0.0004 | 0.0168 | -0.0183 | -0.0092 | -0.0091 |
| 2025-01-13 | 0.0222 | -0.0004 | 0.0367 | -0.0519 | -0.0323 | -0.0196 |
| 2025-02-10 | 0.0335 | -0.0028 | 0.0208 | 0.0473 | -0.0441 | 0.0914 |
| 2025-02-20 | 0.0253 | -0.0006 | 0.008 | -0.1207 | -0.1374 | 0.0167 |
| 2025-02-27 | 0.0124 | -0.0001 | 0.0207 | -0.0588 | -0.0535 | -0.0053 |
| 2025-03-07 | 0.0248 | -0.0067 | 0.0198 | 0.0487 | 0.0717 | -0.023 |
| 2025-03-11 | 0.0058 | -0.0007 | 0.006 | 0.0819 | 0.0823 | -0.0005 |
| 2025-03-12 | 0.0099 | -0.0069 | 0.0119 | 0.0858 | 0.0761 | 0.0096 |
| 2025-03-20 | 0.0022 | -0.0009 | 0.0093 | -0.0789 | -0.0992 | 0.0203 |
| 2025-03-24 | 0.0224 | -0.0004 | 0.0182 | -0.2051 | -0.1981 | -0.007 |
| 2025-03-31 | 0.0174 | -0.0016 | 0.034 | -0.2318 | -0.1839 | -0.0479 |
| 2025-04-09 | 0.1583 | -0.0038 | 0.0376 | 0.007 | -0.0085 | 0.0156 |
| 2025-04-14 | 0.0095 | -0.0003 | 0.0052 | 0.0895 | 0.0597 | 0.0297 |
| 2025-04-17 | 0.0345 | -0.0057 | 0.0205 | 0.0373 | 0.027 | 0.0103 |
| 2025-05-08 | 0.0541 | -0.0087 | 0.0317 | -0.0247 | 0.0319 | -0.0566 |
| 2025-05-09 | 0.0188 | -0.0014 | 0.0328 | -0.0421 | 0.0124 | -0.0545 |
| 2025-05-12 | 0.0314 | -0.0012 | 0.0159 | -0.0656 | -0.0042 | -0.0614 |
| 2025-05-13 | 0.0221 | -0.0057 | 0.0253 | -0.0997 | -0.0506 | -0.0491 |
| 2025-06-03 | 0.0279 | -0.0002 | 0.014 | 0.0918 | 0.0905 | 0.0013 |
| 2025-06-06 | 0.0277 | -0.0011 | 0.0131 | 0.024 | 0.0521 | -0.0281 |
| 2025-06-09 | 0.0111 | -0.0056 | 0.0104 | -0.0024 | 0.0363 | -0.0388 |
| 2025-06-13 | 0.0297 | -0.0017 | 0.0728 | -0.0943 | -0.0505 | -0.0438 |
| 2025-06-17 | 0.0026 | -0.0013 | 0.0486 | -0.0369 | -0.0356 | -0.0013 |
| 2025-07-11 | 0.0006 | -0.004 | 0.0239 | -0.0303 | -0.0379 | 0.0076 |
| 2025-08-20 | 0.0068 | -0.0009 | 0.0149 | 0.0524 | 0.0656 | -0.0132 |
| 2025-08-27 | 0.0094 | -0.0028 | 0.0129 | -0.0076 | -0.0002 | -0.0074 |
| 2025-09-23 | 0.0225 | -0.002 | 0.0163 | 0.0148 | 0.0037 | 0.0111 |
| 2025-09-24 | 0.0212 | -0.0024 | 0.0247 | -0.0068 | -0.0101 | 0.0032 |
| 2025-10-03 | 0.0058 | -0.0016 | 0.0105 | -0.0816 | -0.0722 | -0.0094 |
| 2025-10-06 | 0.0101 | -0.0021 | 0.0145 | -0.0854 | -0.0662 | -0.0192 |
| 2025-10-22 | 0.0122 | -0.0028 | 0.021 | 0.0043 | -0.0013 | 0.0056 |
| 2025-10-23 | 0.0318 | -0.0015 | 0.0649 | -0.0302 | -0.0218 | -0.0084 |
| 2025-11-03 | 0.0039 | -0.0061 | 0.0053 | 0.0753 | 0.0282 | 0.0471 |
| 2025-11-11 | 0.0368 | -0.0024 | 0.0135 | 0.0185 | -0.0292 | 0.0476 |
| 2025-11-18 | 0.0228 | -0.0026 | 0.0269 | 0.0535 | 0.0224 | 0.0311 |
| 2025-11-24 | 0.0079 | -0.0014 | 0.0327 | 0.0409 | 0.0284 | 0.0125 |
| 2025-11-26 | 0.0162 | -0.0053 | 0.0128 | 0.0313 | 0.0101 | 0.0212 |
| 2025-12-03 | 0.0142 | -0.0015 | 0.006 | -0.0096 | -0.0629 | 0.0533 |
| 2025-12-19 | 0.0034 | -0.0043 | 0.0109 | 0.0014 | 0.0071 | -0.0057 |
| 2026-01-08 | 0.0461 | -0.003 | 0.037 | 0.0707 | 0.0528 | 0.0179 |
| 2026-01-14 | 0.0249 | -0.0013 | 0.0191 | 0.053 | 0.0535 | -0.0005 |
| 2026-02-03 | 0.0252 | -0.008 | 0.0338 | 0.081 | 0.0574 | 0.0236 |
| 2026-02-06 | 0.0148 | -0.0002 | 0.0083 | -0.0032 | 0.021 | -0.0242 |
| 2026-02-19 | 0.0049 | -0.0058 | 0.0194 | -0.0031 | 0.08 | -0.0831 |
| 2026-02-26 | 0.0035 | -0.003 | 0.0137 | 0.0789 | 0.1125 | -0.0336 |
| 2026-03-11 | 0.0318 | -0.0019 | 0.0127 | 0.1238 | 0.1086 | 0.0152 |
| 2026-03-12 | 0.0217 | -0.0014 | 0.1253 | 0.1204 | 0.1165 | 0.0039 |
| 2026-03-17 | 0.0153 | -0.0004 | 0.0727 | 0.0623 | 0.0706 | -0.0083 |
| 2026-03-18 | 0.0112 | -0.0031 | 0.0895 | 0.0144 | 0.0121 | 0.0023 |
| 2026-03-25 | 0.0022 | -0.0012 | 0.0066 | -0.05 | -0.0762 | 0.0262 |
| 2026-03-26 | 0.0186 | -0.0044 | 0.0389 | -0.0693 | -0.0909 | 0.0216 |
| 2026-03-27 | 0.0127 | -0.0001 | 0.0713 | -0.0912 | -0.1036 | 0.0124 |
| 2026-04-22 | 0.0317 | -0.0014 | 0.0688 | -0.0051 | 0.011 | -0.0161 |
| 2026-04-27 | 0.0075 | -0.0025 | 0.0181 | -0.0305 | -0.0048 | -0.0257 |
| 2026-04-29 | 0.0323 | -0.0021 | 0.0556 | -0.084 | -0.0538 | -0.0301 |
| 2026-05-11 | 0.0223 | -0.002 | 0.0254 | -0.0332 | -0.0132 | -0.02 |
| 2026-05-14 | 0.0096 | -0.0003 | 0.0057 | -0.0591 | -0.0307 | -0.0284 |
| 2026-05-18 | 0.004 | -0.0005 | 0.0243 | -0.0766 | -0.0464 | -0.0302 |
| 2026-05-22 | 0.0025 | -0.0004 | 0.01 | -0.0523 | -0.0255 | -0.0268 |
| 2026-06-10 | 0.0574 | -0.003 | 0.0168 | -0.0791 | -0.0744 | -0.0047 |
| 2026-07-07 | 0.0545 | -0.0034 | 0.0319 | 0.0362 | 0.1037 | -0.0675 |
| 2026-08-11 | 0.0006 | -0.0027 | 0.0056 | 0.0337 | 0.0232 | 0.0106 |
The takeaway
Yes, DVN does tend to lag XOP after this setup — but the edge is mild and mostly looks like a generic DVN habit rather than something special about the pattern. Across 96 qualifying days, DVN trailed XOP by an average of 0.67% over the next 10 sessions, with a median gap of -0.62% and DVN the laggard in 58% of cases. The problem is the benchmark: DVN already lagged XOP by 0.20% on average across all 721 days in the sample, and the extra 0.47 percentage points on event days is not statistically distinguishable from that baseline (p = 0.12). Against zero the event gap does clear the bar, at roughly a 4-in-100 chance of being luck (p = 0.043), but that test treats overlapping 10-day windows as independent observations, which flatters the result. Bottom line: this is a weak lean, not a signal you'd trade — the direction is consistent, the magnitude is small, and the evidence that the setup itself causes the underperformance is thin.
The fine print
- 96 events, and the 10-session windows overlap whenever days cluster — the effective independent sample is smaller, so p = 0.043 is softer than it reads.
- Event days barely separate from ordinary days (p = 0.12); much of the -0.67% is the same mild lag DVN shows all the time (-0.20% baseline).
- Only ~3 years of data in a single oil-price regime; DVN-vs-XOP relative behavior is sensitive to crude macro conditions.
- Gross price returns only — no transaction costs, borrow costs, or liquidity assumptions for a DVN-vs-XOP relative position.