XOM top‑decile news sentiment and forward 5‑day returns — does extreme bullish hype mark short‑term tops?
The short-term fade trade did not show up in Exxon's numbers. Over the roughly three years tested, the 49 trading days with top-decile news sentiment posted a mean forward 5-day return of +0.68%, versus a +0.65% baseline across all 490 days. That gap is just three basis points — well inside the noise band for a stock with daily swings of roughly three percentage points.
The theory being tested is that extreme bullish hype in energy names marks local tops, as institutional sellers quietly unload into retail enthusiasm. For XOM in this window, that pattern is absent; the p-value of 0.94 says a gap this size would materialize by pure chance nearly every time. If anything, the direction leans slightly positive, not negative, which undercuts the fade hypothesis rather than supporting it.
Below is the full methodology, the charts, and the complete breakdown of the sentiment-return relationship.
For XOM over the past ~3 years, do days with top-decile positive news sentiment lead to below-baseline forward 5-day returns? Extreme bullish news hype marks short-term tops in energy stocks as institutional sellers fade retail excitement.
How this was measured
XOM minute bars resampled to daily closes; forward 5‑day return defined as (close[t+5] / close[t]) − 1. Daily mean sentiment obtained from XOM_news headlines (ticker_sentiment_score primary, overall_sentiment_score fallback). A trading day is flagged as 'top‑decile' when its daily mean sentiment equals or exceeds the 90th percentile of all sentiment scores in the available window (in‑sample ranking). The mean forward 5‑day return of these top‑decile days is compared with the unconditional mean of all days (baseline) via Welch's two‑sample t‑test (unequal variance). A gap significantly below zero would be consistent with the hypothesis that extreme bullish sentiment fades into short‑term under‑performance.
The key numbers
Reading the numbers
Top-decile sentiment days averaged a 0.68% forward 5-day return, basically the same as the 0.65% average across all 490 days (gap 0.03 pts; p=0.94). So extreme positive news hype did not predict below-baseline returns in this sample.
The charts
The two boxes overlap heavily: top-decile sentiment days have a mean forward 5-day return of 0.68%, while all other days average 0.65%. The top-decile group ranges from -4.8% to +6.6%, and all other days range wider, from -10.5% to +10.2%. The thing to notice is that the extreme-news days are not clustered at negative forward returns — their average sits right on top of the baseline. For the question, this means there is no visible tendency for top-decile hype to mark a short-term top.
This scatter plots every trading day's sentiment score against its forward 5-day return, with sentiment ranging from -0.45 to 0.75 and returns from -10.5% to +10.2%. The points form a wide cloud at every sentiment level, so low, medium, and high sentiment days all show both large positive and large negative forward returns. The top-decile sentiment days are highlighted, but their mean forward return of 0.68% mirrors the overall average of 0.65% rather than dropping below it. In plain terms, the chart does not support the idea that extreme bullish news is followed by below-baseline performance.
Top‑decile sentiment days (first 20, chronological)
| date | daily_sentiment | fwd5d_return |
|---|---|---|
| 2023-08-04 | 0.4154 | 0.039 |
| 2023-08-23 | 0.3226 | 0.0353 |
| 2023-08-24 | 0.2983 | 0.0475 |
| 2023-09-19 | 0.4449 | -0.0056 |
| 2023-09-20 | 0.4946 | 0.0369 |
| 2023-10-09 | 0.2882 | -0.0076 |
| 2023-10-10 | 0.2897 | 0.0182 |
| 2023-10-11 | 0.4303 | 0.0656 |
| 2023-10-25 | 0.3479 | -0.0257 |
| 2023-11-17 | 0.3524 | -0.0122 |
| 2023-12-05 | 0.3093 | -0.0235 |
| 2023-12-12 | 0.2894 | 0.0518 |
| 2023-12-14 | 0.4334 | 0.0027 |
| 2023-12-22 | 0.4235 | 0.0041 |
| 2023-12-28 | 0.2881 | 0.0224 |
| 2024-01-24 | 0.4237 | 0.0342 |
| 2024-03-01 | 0.3976 | 0.025 |
| 2024-03-08 | 0.4504 | 0.0263 |
| 2024-03-15 | 0.315 | 0.0193 |
| 2024-03-19 | 0.3182 | 0.007 |
The takeaway
No — over the roughly three years tested, top-decile news sentiment days in XOM did not lead to below-baseline forward 5-day returns. The 49 highest-sentiment days averaged a +0.68% forward return, essentially the same as the +0.65% average across all 490 trading days, a gap of just +0.03 percentage points. The stats say this is basically a coin flip: a p-value of 0.94 means a gap this size would appear by pure luck about 94 times out of 100, so there is no meaningful signal that bullish hype marks a short-term top here. If anything, the direction is slightly positive, not negative, but the edge is tiny relative to daily swings of roughly 3 percentage points. Takeaway: the "extreme sentiment fades" idea is not supported for XOM in this window, so high daily sentiment scores are not a reliable short-term fade signal.
The fine print
- The 90th-percentile threshold is set in-sample using the same data, so the top-decile split is descriptive and carries some look-ahead bias.
- Forward 5-day returns overlap day to day, so the t-test is not based on fully independent observations.
- Only days with news headlines are included; the baseline excludes non-news days, and the window covers a single ~3-year energy regime.