The Energy Playbook's Dirty Secret: Tiny Samples

There is a certain comfort in an energy thesis: oil goes up, names follow, insiders know best, earnings beat and the stock gaps. The only problem is the data. This week's batch of quant checks on FANG, XOM, OXY, BKR, EQT and BP all point the same direction — the edges are real-looking, but built on so few observations that they are closer to anecdotes than signals.

Earnings day and the crude-split illusion

Take FANG. Across 34 positive EPS surprises, the crude-split pattern — Brent up versus Brent down — seems to tell a clean story. The five Brent-down days opened the next session up 0.47%, then handed all of it back, closing -0.47% open-to-close. The 29 Brent-up days opened +0.94% and kept grinding, closing +1.13%. That looks like a cue: don't chase gap-ups without a Brent tailwind. But the research note stops at the obvious caveat: five down days. The evidence is far too thin to call real. One bad headline, one supply shock, and that average flips.

Small samples everywhere

Same story in the insider lane. OXY insider purchases near Brent's 20-day range extremes: bottom-quartile purchases averaged 3.18% forward 20-day returns, versus 0.43% for top-quartile — a 2.75 percentage-point edge. Sounds like something. Then you see the counts: 11 bottom-quartile purchase days and 8 top-quartile days. The t-test says the gap could easily be noise. Meanwhile XOM's supposedly bearish setup — rising Brent with contracting free-cash-flow margin — only qualified once in the study window. That one quarter produced a next-60-day return of +2.35%, which beats the median baseline (+1.89%) but trails the mean (+3.82%). One event, sitting at the 51.9th percentile. That's not a signal; that's a coin flip.

The benchmark doesn't care about your thesis

Then there's the RSI setup. Buying EQT when its 14-day RSI drops below 35 and selling on a still-unstated exit condition returned +26.72% on $100k across 13 trades. SPY buy-and-hold over the same window returned +68.30%. That's not just underperforming; it's the kind of drawdown that makes a strategy feel smart until you look at the benchmark. And BKR's revenue-growth acceleration while Brent sat below its 50-day SMA? Only two qualifying quarters — June 2024 and June 2025 — averaging +5.5%, just under the +6.1% baseline. Both positive, but the study's own verdict: this did not produce above-baseline returns.

The pattern here is less about crude and more about the research process. The platform is publishing negative or null results: signals that look like edges but fail the sample-size test. That's the opposite of the usual promotional quant content. The data leans toward a simple truth: energy markets are messy, and most neat correlations in this sector are built on a handful of days. If you're going to use a crude-oil overlay on your equity thesis, the burden of proof is on the pattern — not on the stories we tell about it.