APA vs XOP after positive surprises during negative Brent 20-day tape
The setup is intuitive: a high-beta E&P like APA beating estimates while Brent's trailing 20-day tape is negative should signal company-specific execution and hedges that the peer basket underreprices, letting APA lead XOP. We tested that condition on the past three years of earnings and forward returns.
Only seven positive surprises occurred while Brent was falling. The average 20-session APA-minus-XOP gap was a wide 4.82%, but the median was just 0.06%, and APA outran XOP in only 4 of 7 cases. Two 2025 events supplied essentially all the edge; the earliest events saw APA lag.
The one-sample t-test puts p at 0.214—roughly a 1-in-5 chance of luck—and the falling-crude filter actually made the average gap 2.47% smaller than the unconditional set. That is not a repeatable signal. The full event-by-event evidence is below.
Over the past ~3 years, when APA posts a positive quarterly earnings surprise while Brent crude's trailing 20-day return is negative, does APA outperform XOP over the next 20 sessions? I expect a high-beta E&P beating estimates into a falling crude tape to signal company-specific execution and hedges that the peer basket underreprices, so APA leads the producer basket until the crude tape stabilizes.
How this was measured
Daily closes were built from minute bars for APA and XOP and joined on common trading days. APA quarterly earnings reports from the past 3 years were filtered to positive surprises, defined as reported EPS above consensus estimate or a positive surprise field where available. For each event, the trailing 20-session Brent return was computed from the daily Brent series with a one-business-day publication lag, then aligned to the first common APA/XOP trading day on or after the report date. The conditional sample is the subset of positive surprises where that Brent trailing 20-session return was negative. For each event, 20-session forward returns were measured for APA and XOP from the anchor day close, and the main outcome is APA − XOP over those 20 sessions. A one-sample t-test on the APA−XOP gap tests whether conditional outperformance is distinguishable from zero.
The key numbers
Reading the numbers
Across 7 qualifying events, APA beat XOP by an average of +4.82% over the next 20 sessions, but the middle event's gap was only about +0.06%, so a couple of big wins are pulling the average up. With p=0.214, that gap is well within what random noise could produce — treat this as a pattern worth watching, not a proven edge.
The charts
Each pair of bars is one earnings event where APA beat estimates while crude's prior 20-day return was negative, showing what APA and the XOP basket did over the following 20 sessions. The eye should go to the twin towers in the middle: 2025-08-06, where APA returned +26.3% against XOP's +6.6%, and 2025-05-07, at +16.4% versus +6.4%. But the leftmost three events are all negative and APA's losses run deeper than XOP's (for example -10.5% versus -6.6% on 2024-07-31), which is exactly the high-beta behavior you'd expect if the crude tape keeps falling. So the answer to your question is 'sometimes, and mostly in the two events where crude turned' rather than a consistent lead.
This box plot compares the spread of the seven forward returns for APA and XOP side by side. APA's average is higher (+5.97% versus +1.15%) and its best outcome reaches +26.3% against XOP's +9.3%, but its worst case is also deeper at -10.5% versus -6.6%, meaning the whole distribution sits lower and wider. In plain terms, owning APA into these setups bought you more upside and more downside — a higher-beta bet on the same crude tape, not a steadier one. The fact that the typical (middle) outcome for both is close to flat is why the average gap flatters the strategy.
This scatter plots the eight events with Brent history, putting the crude drawdown on the horizontal axis and APA's outperformance versus XOP on the vertical. Note that every point sits at or below a Brent trailing return near zero — the deepest reading is about -14% — yet APA's relative result still ranges from roughly -3.9% to +24.6%, averaging +7.3%. That vertical scatter at similar crude readings is the key message: how much crude had fallen before the print did not reliably dictate whether APA led the basket. Your 'hedges plus execution' thesis shows up in the outliers, but the chart gives no clean slope to lean a rule on.
Qualifying event details
| report_date | anchor_date | surprise_pct | brent_20d_ret | APA_20d_ret | XOP_20d_ret | outperform |
|---|---|---|---|---|---|---|
| 2023-11-01 | 2023-11-01 | 25.47 | -0.0809 | -0.0876 | -0.0509 | -0.0367 |
| 2024-07-31 | 2024-07-31 | 25.81 | -0.1022 | -0.1051 | -0.066 | -0.0391 |
| 2024-11-07 | 2024-11-07 | 1.01 | -0.007 | -0.0245 | -0.0251 | 0.0006 |
| 2025-05-07 | 2025-05-07 | 29.17 | -0.1402 | 0.1636 | 0.0641 | 0.0995 |
| 2025-08-06 | 2025-08-06 | 81.25 | -0.0463 | 0.2628 | 0.0656 | 0.1972 |
| 2025-11-05 | 2025-11-05 | 25.68 | -0.0307 | 0.2168 | 0.0933 | 0.1236 |
| 2026-05-06 | 2026-05-06 | 157.99 | -0.1027 | -0.0078 | -0.0005 | -0.0073 |
The takeaway
The thesis doesn't hold up: when Brent's tape was falling, APA's positive earnings surprises did not reliably lead XOP. On the seven qualifying events, APA beat XOP by an average of 4.82% over the next 20 sessions (APA +5.98% vs XOP +1.15%) — but that average is doing all the heavy lifting from just two events. The median gap was a rounding-error 0.06%, and APA only outran the basket in 4 of 7 cases, so the typical outcome was a dead heat. In fact the two earliest events (late 2023 and mid-2024) saw APA lag XOP by roughly 3.7–3.9%, while the entire edge comes from two 2025 events where APA ran +10% and +20% ahead. The stats agree with that read: p = 0.214 is about a 1-in-5 chance this is luck, nowhere near conclusive. Tellingly, the falling-crude condition didn't help at all — the average gap was 2.47% smaller than the unconditional set of positive surprises, the opposite of what you'd expect if it were adding signal. Net: this looks like a coin flip dressed up by a couple of outsized winners, not a repeatable setup worth leaning on.
The fine print
- Just 7 qualifying events over three years — far too few to call this a stable edge; the t-test is descriptive, not decisive.
- The mean rests on two 2025 events (+10% and +20% relative); strip them and the median gap collapses to essentially zero.
- Anchor is the first common trading day on or after the report, so whether the surprise hit pre-open or after-close isn't modeled.
- Three years may span only one or two crude regimes, and APA−XOP is a relative comparison, not a clean tradeable hedged spread.