Crude-Based Strategies Look Tempting but Keep Trailing the Benchmark

There is something about oil charts that makes a trader want to believe. A headline drop in Brent, a stock closing near its high, a clean cross of the 20-day moving average — each one feels like a signal with a story attached. But the latest run of research from the platform keeps telling the same story: the oil tape is full of compelling narratives, and almost none of them survive contact with the data.

A Sentiment Edge That Fails the Smell Test

Take the study on Phillips 66, published this morning. On days when crude fell more than 1% and PSX news sentiment sat in the top quintile, the stock averaged +2.21% over the next five days. On rising-crude days with the same sentiment level, the average was +1.63%. That 58-basis-point gap looks like a tradable idea, and the median tells the same story: +2.15% versus +1.06%. So the direction is right. But here is where the platform's numbers get uncomfortable: only 18 down-crude days and 36 rising-crude days underpin that whole comparison. With that sample size, the stats say the edge could easily be luck. In other words, the pattern is exactly what a trader would invent in hindsight — but the evidence is too thin to trust.

The Moving-Average Experiments: A Graveyard of Good Ideas

Then there are the backtests that lean on Brent for timing. The results are not just disappointing; they are a systematic lesson in how far a narrative can drift from reality. Buying XLE when Brent sits above its 100-day moving average generated +7.29% on $100,000 across 28 trades, with a 46% win rate. That sounds like a modest positive. But the same window handed SPY buy-and-hold +68.30%, a gap of 61 points. The relative-strength play between Exxon and Chevron did a bit better — +25.08% across 31 trades, with a 61% win rate — yet it still trailed SPY by 43 points. A rule that buys Valero when Brent closes below its 20-day moving average earned +9.26% but only a 37% win rate, once again miles behind the benchmark. And the SLB version, buying when Brent closes above the 20-day, produced an outright loss of -16.69%, the worst of the bunch.

These are not edge cases. Four different, plausible crude-driven rules all underperformed a boring buy-and-hold by enormous margins. The best win rate among them — 61% — still could not overcome the drag of a lagging asset class. The message is not that energy is a bad sector forever; it is that the simple versions of crude timing are not doing the work people hope they are.

Close Location Turns Out to Be a Dead End

One more piece of the puzzle comes from the XLE study on high-volume Brent-led selling days. When crude is getting hammered, does it matter whether the energy ETF closes near its high or near its low? Intuition says it should — a strong close on a red day suggests buyers are stepping in, and a weak close suggests the selling is unfinished. The data, however, says no. Across 67 high-volume down days, top-half closes led to an average 5-day forward return of 1.04%, versus 0.91% for bottom-half closes. That 0.13-point gap is statistically indistinguishable from noise, with a Welch p-value of 0.87. The medians pointed the same way.

Put all three studies together and a clear picture emerges. The oil market is constantly generating patterns that look meaningful until you measure them properly. The PSX sentiment signal has the right shape but not enough evidence. The moving-average rules have enough trades to matter and they all lag the index by a mile. The close-location effect simply does not exist. If there is a viable edge in energy timing, it is not hiding in these obvious corners. The data leans toward humility — and maybe toward accepting that the benchmark is a tougher opponent than any of these rules have beaten.