TTE vs XLE forward 20-day excess: joint EPS-revision/Brent condition is untestable; Brent-tape-only comparison shown
The joint condition the question asks for — analyst estimates revised up while crude rises — has no history to test. Over the past three years, zero sessions qualified, because the estimate data holds one current snapshot per quarter rather than the 30-day revision path needed to measure it.
What survived is the Brent leg alone, and it came back cleaner than expected. Across 306 sessions when crude's prior 20-day return was positive, TTE beat XLE by an average 0.86% over the next 20 sessions; when crude wasn't rising, TTE lagged by 0.62%. TTE won 61% of the positive-crude windows.
That's a real tilt, but it's a crude-tape effect, not evidence for slower repricing in a European integrated. The revision thesis stays open, untested. The full breakdown, including the Welch tests, is below.
Over the past ~3 years, when TTE's next-quarter consensus EPS estimate is revised upward by at least 1% over the prior 30 days while Brent crude's 20-day return is positive, does TTE outperform XLE over the next 20 trading sessions? I expect positive analyst revisions into a rising crude tape to be repriced more slowly in a European integrated than in U.
How this was measured
Resampled TTE and XLE minute bars to daily closes, aligned inner on trading dates, and computed close-to-close forward 20-session excess return as TTE minus XLE. Brent crude daily values were reindexed to the trading calendar and the 20-session return was lagged one day for T+1 publication. Days were split by whether the lagged Brent 20-day return was positive, and one-sample plus Welch tests were used on the forward excess. The requested TTE 30-day EPS revision condition was not applied historically because TTE_estimates lacks within-quarter revision history; only the current as-of-today revision snapshot is reported.
The key numbers
Reading the numbers
Strictly as asked, zero days qualified because there is no historical 30-day EPS revision series to combine with Brent. On the Brent leg, TTE beat XLE by 0.86% on average over the next 20 sessions when Brent's 20-day return was positive, versus losing 0.62% when it wasn't; the gap is clear.
The charts
This box plot splits the next-20-session TTE-minus-XLE return into three groups: Brent up, Brent flat-or-down, and all days. The key detail is the average: +0.86% when Brent's 20-day return was positive versus -0.62% when it wasn't, with all days near flat at +0.04%. The distributions still overlap heavily and the Brent-up group's outcomes range from about -16.6% to +15.5%, so the tape tilts the odds rather than determining the result. This supports the Brent leg of your question, but not the analyst-revision leg, which had zero testable history.
The scatter plots each day's lagged Brent 20-day return against the following 20-session TTE-minus-XLE excess, with 690 observations. Brent readings span a very wide range, from -32.6% to +70.3%, while the forward excess stays between -20.2% and +15.5% and averages just +0.04%. The cloud is widely dispersed rather than a tight line, so a single crude reading does not map one-for-one to the next month's relative performance. That is why the cleaner read is regime averages, and why the full joint condition remains untested with zero eligible days.
Forward 20-day TTE minus XLE return by Brent tape
| Group | N | Mean | Median | Std | Positive share |
|---|---|---|---|---|---|
| Brent 20d > 0 | 306 | 0.0086 | 0.01 | 0.0462 | 0.6111 |
| Brent 20d <= 0 | 384 | -0.0062 | -0.0001 | 0.0429 | 0.5 |
| All days | 690 | 0.0004 | 0.0055 | 0.045 | 0.5493 |
The takeaway
The question as asked can't be answered: there are zero historical days in the past three years where TTE's 30-day EPS revision could even be measured, because the estimate data only holds a single current snapshot per quarter rather than a revision history — so no joint revision-and-Brent test was run. What could be tested is the Brent leg alone, and it came back unusually clean. On the 306 sessions when Brent's prior 20-day return was positive, TTE beat XLE over the following 20 sessions by an average of 0.86%, versus a 0.62% lag on the 384 days when Brent wasn't rising — a gap that's very unlikely to be noise (Welch p ≈ 0.00002). The Brent-positive group's outperformance on its own also clears zero (p = 0.001), and TTE won 61% of those windows versus a literal 50-50 coin flip when Brent was flat or down. That's a real signal, not a lean — but it is a Brent-tape effect, not the analyst-revision effect you were asking about, so it doesn't confirm your thesis about slower repricing of European integrateds. Practically: rising crude does seem to tilt the next month in TTE's favor against XLE, and it's worth testing properly once a revision history exists — but today the revision half of the setup is simply unmeasured, and the current snapshot (revision at -1.98%, Brent 20-day at +18.77%) is a starting condition, not evidence.
The fine print
- The EPS-revision half of the trigger has no historical data — the 0.86% result is purely a Brent-tape effect and says nothing about your revision thesis.
- Forward 20-day windows overlap heavily day to day, so the t-stat and p-values overstate how many truly independent observations back them.
- Returns are price-only close-to-close; TTE ADR and XLE pay different dividends, which can nudge relative performance.
- Window runs 2023-10-31 to 2026-08-03, and the current revision (-1.98%) and Brent (+18.77%) figures are snapshots, not backtest inputs.