Energy Headlines Shine While Quant Signals Stay Coin-Flips
If you read the energy headlines this week, you'd think the sector had finally broken free from the noise. Imperial Oil posted a record Q2 net income of $2,190 million, Exxon is touting a 9% production CAGR through 2030, and even a tragic safety incident at Cheniere's Sabine Pass hasn't dented the bullish narrative. But the latest batch of quant research from trades.run paints a different picture: many of the relationships traders lean on in oil are, statistically speaking, coin flips.
The CPI-Surprise Story Doesn't Hold Up
Take the notion that big oil names should outrun the market when inflation surprises. The platform tested XOM's monthly excess return in months when CPI MoM came in above its trailing 12-month average. Across 21 overlapping months, 13 were surprise months, and XOM beat SPY in only 7 of them — a 54% hit rate. Mean excess return was +0.38% in surprise months versus -1.69% in non-surprise months, but the t-test p-value of 0.54 says that gap could easily be luck. Headline writers may love the inflation-hedge angle, but the data leans toward no consistent edge.
Relative Strength and Volatility Correlations Fail the Test
Two backtests also underperformed badly. Buying VLO when its five-day return outpaced XLE by a certain margin returned +22.86% on $100,000 over 62 trades, with a 44% win rate. SPY buy-and-hold over the same window returned +68.30%. That's a 45.44-point shortfall. The EQT strategy, triggered when the stock closed 5% below its 20-day moving average while Brent crude did something undisclosed, fared even worse relative to the benchmark: +13.15% versus +68.30%, a 55.15-point trail. These are small samples — 62 and 11 trades — but the point isn't the edge; it's the absence of one.
The cross-asset relationships don't help either. The XOP-to-XLE weekly spread has essentially no predictive power for next-week Brent moves: r = 0.069, p = 0.40, and it called the direction correctly only 48.4% of the time — slightly worse than a coin flip. And OXY's correlation with Brent doesn't get a meaningful boost when SPY volatility is in its top decile: 0.526 on high-vol days versus 0.513 on other days, with a one-sided Fisher p of 0.444. Even the BKR-Brent beta, which you'd expect to rise in high oil-vol regimes, stays flat: 0.263 in high-vol months versus 0.255 in normal months, a difference that's economically trivial and statistically indistinguishable from noise (t-stat 0.46, p=0.646).
What the Headlines Miss
The news flow is doing what it always does: projecting confidence. Imperial's record earnings are real — $2,190 million in Q2 2026, up sharply from Q2 2025 — but they're a function of higher commodity prices, not a sustainable signal. Exxon's Permian-and-technology story is upbeat, but the 9% CAGR is a target, not a result. And the Cheniere tragedy at Sabine Pass is a reminder that operational risk is always one incident away. Morgan Stanley's Apple comment — that investors are 'barking up the wrong tree' on supply issues — is a good broader reminder that consensus fears and consensus confidence often overshoot.
The platform's research doesn't say energy is a bad trade. It says the obvious signals — the ones that feel like they should work — are mostly noise. The data leans toward humility: if you're trading oil stocks based on CPI surprises, relative strength versus XLE, or Brent correlations, you're betting on a coin flip. That doesn't make for a thrilling blog post, but it might keep your account from becoming a backtest's worst trade.