Energy Signals Win Most Trades and Still Lose to Buy-and-Hold

Six backtests landed on the platform this week, all of them on energy names, all of them built on triggers that sound reasonable in a research note. Not one of them beat simply owning the index. That is the whole story, and it is worth sitting with for a minute.

The win-rate illusion

Start with the most seductive result. Buying OXY when its 10-day average true range sits in the top quintile returned +28.30% on $100,000 across 20 closed trades, with a 65% win rate, a best trade of +9.01% and a worst of -5.96%. That is a real edge by any trade-by-trade measure. Over the same window, SPY buy-and-hold returned +76.34%, so the strategy finished 48.04 points behind the benchmark.

The pattern repeats. The RSI-below-30 setup on OXY won 75% of its trades and still only compounded to +4.14% across 8 closed trades, trailing by 72.20 points. A high hit rate with small wins and occasional losses is not a compounding machine. It is a metronome. You can be right three times out of four and still watch the index pull away, because the wins are capped and the losses are not.

Where the signals broke

The XOM beta-to-Brent filter is the cautionary tale. It lost 8.10% across 36 closed trades with a 20% win rate, trailing the benchmark by 84.44 points. The revealing detail is that the worst single trade was only -3.45%. Nothing blew up. The strategy was simply ground down, trade after trade, by a filter that kept saying yes when the tape said no.

LNG shows the opposite failure mode. A momentum trigger — 20-day total return beating SPY by at least 5 percentage points — produced +14.61% across 33 trades with a 58% win rate. But the worst trade was -15.87%. One tail event of that size does most of the damage to a 58% hit rate, and the strategy still ended 61.73 points behind. Win rate tells you nothing about the shape of the distribution underneath it.

Sentiment was the better input

The strongest return in the set came from the least technical signal: buying APA when its daily news-sentiment score ranked in the top quintile. That returned +37.23% across 45 closed trades with a 69% win rate, a best trade of +6.04% and a worst of -7.35%. It is the best of the cohort and it still trailed by 39.10 points.

The XLE/SPY ratio at a 20-day low, paired with a Brent condition, returned +3.60% across 13 trades with a 67% win rate — trailing by 72.73 points, with a worst trade of -6.78% against a best of just +2.77%.

What the cohort is actually telling us

Look at the six together and the shape is unmistakable: high win rates, shallow best trades, and a persistent inability to participate in the drift of the index. Every one of these strategies is a timing overlay on a single sector, and timing overlays charge a fee in the form of missed upside. That the benchmark delivered +76.34% over the window is not a footnote — it is the entire context. In a tape that strong, a mean-reversion trigger has to be exceptional just to keep pace, and none of these were.

The data leans somewhere uncomfortable for signal-hunting: a trigger with a 69% win rate can be genuinely predictive at the trade level and still be a poor way to hold risk over time. The interesting question coming out of this week is not which energy trigger has the highest hit rate. It is whether any of these setups can be expressed in a way that captures trend rather than harvesting the wiggle. As currently specified, the honest read is that the signals work and the strategies don't — and those are two very different claims.