Where outperformance actually lives

Read enough strategy marketing and a pattern shows up. The pitch is always an average: the whole-period return, one smooth equity curve, a single win rate. What it almost never tells you is when the strategy earned that average. For nearly every systematic approach that has ever worked, the answer is: not evenly. Outperformance concentrates. It arrives in specific market regimes and goes quiet in the rest.

This post is about that concentration: why vendors rarely mention it, and the one question that separates builders who tested their system from builders who tested their copywriting.

Edge is lumpy, not smooth

Strategies that beat a benchmark almost never beat it a little every week. The more common shape is long stretches of roughly matching the market, punctuated by bursts where the strategy pulls ahead. Trend following built its reputation in a handful of extended trend and crisis years and looked ordinary in between. Momentum does well in stable trends and gives a chunk back when they snap.

None of that is a defect. It is what an edge is. A systematic strategy is a bet on a specific, recurring market behavior, and that behavior is not always present. When it is not, a well-built strategy degrades gracefully toward the market's own return, and a badly built one bleeds. A system that outperformed in every environment would need a genuine edge in every environment, and nothing with a public track record has ever shown that.

What a regime actually is

A regime is a stretch of time in which the market behaves in a recognizably consistent way. Trending or chopping. Breadth wide or narrow. Volatility calm or violent. Correlations loose, so individual names can separate from the pack, or compressed, so everything moves together. You do not need a formal definition to use the idea, just one question about any strategy in front of you: what does the market have to be doing for this thing to make money?

Every strategy has an answer, written down or not. A breakout system needs follow-through after breakouts. A mean-reversion system needs prices that snap back instead of trending away. An options-selling system needs realized volatility to stay below what it was paid for. That answer defines the strategy's home regime, and the home regime is where the edge lives.

A worked example: leadership rotation

Take leadership rotation, the family of strategies that buys the market's strongest names and rotates as strength moves. When does it have an edge? When leadership is strong and rotating: a minority of names clearly outrunning the index, with the baton passing between them often enough that following relative strength beats sitting still. When does it not? When leadership compresses. In risk-off phases and in broad, undifferentiated rallies, correlations rise, relative strength stops discriminating, and a portfolio of leaders is mostly a sample of the index. In those stretches a leadership strategy should be expected to roughly match its universe, and its job shifts from outperforming to not doing damage: cutting losers by rule and holding cash when nothing qualifies.

Notice what that description gives you that a headline number never could: a falsifiable expectation. If leadership is strong and rotating and the strategy still is not separating from the index, something is wrong. If leadership is compressed and the strategy is merely matching the market, nothing is wrong at all. Without the regime statement you cannot tell those two apart, so you abandon good systems in quiet regimes and trust broken ones in loud ones.

Why vendors leave this out

Mostly incentives, not conspiracy. A whole-period average is a better sales tool than a conditional claim. Beat the market over a decade fits in a headline. Beat it while leadership was strong and roughly match it otherwise takes a paragraph, and invites the follow-up nobody selling wants: what regime are we in right now? So the marketing compresses. The long equity curve gets shown at a scale where the flat years disappear into the slope. The backtest window quietly starts where the friendly regime begins.

To be fair, most of this is compression rather than lying. But this particular compression removes exactly the fact a buyer needs most, because regime concentration is the difference between a system that prints steadily and a system that waits, then works. Those are very different products to live with, even when their long-run numbers look the same. If you want the full checklist for taking a marketed backtest apart, we wrote one in how to read a backtest, and covered its sibling trap in survivorship bias.

The question to ask of any system

All of this collapses into one question you can put to any vendor, any bot marketplace listing, any strategy PDF: when does this not work?

A builder who has genuinely tested a system can answer immediately, because the flat and losing stretches are burned into their memory. A good answer has three properties:

• It names the home regime in observable terms, things you could point to on a chart, not vibes.
• It says what the strategy does outside that regime: match the market, sit in cash, take small rule-based losses.
• It comes with the unflattering evidence: the full curve, quiet years visible, not a cropped highlight.

The bad answers are just as recognizable. It adapts to all market conditions is the most common, a red flag precisely because it is unfalsifiable. Deflecting to a bigger chart is another.

Stating it out loud

This is the posture we try to hold with Coil (coil.trade). Coil is a long-only system built on leadership rotation, so its regime statement is the one from the worked example above: the edge is expected when leadership is strong and rotating, and outside those regimes the honest expectation is roughly market-like behavior with the risk rules doing the real work. That is not a hedge. That is the claim. The research behind the ranking is laid out at /how-it-works, quiet stretches included, because a conditional claim you can check is worth more than an unconditional one you have to take on faith.

None of this is investment advice, and no regime statement changes the basic fact that markets can lose money in any regime. It changes whether you knew what you were buying.

FAQ

What is regime concentration in trading?

It is the observation that a strategy's outperformance is not spread evenly across time. Most systems earn their edge inside specific market regimes and roughly match, or trail, the market elsewhere. A whole-period average hides this, which is why the same headline number can describe two very different experiences.

Does regime concentration mean a strategy is broken?

No. A real edge is a bet on a specific market behavior, so a real edge is always concentrated somewhere. The problem is not concentration, it is undisclosed concentration. A strategy that matches the market while waiting for its home regime can be perfectly sound, as long as you knew that was the deal.

How do I find out when a system does not work?

Ask the builder directly, and look at the full equity curve for flat or losing stretches instead of the summary number. A specific, mechanical answer that names observable market conditions is a good sign. A claim that it works in all market conditions is the red flag, because it is unfalsifiable.