Market regimes: why one strategy cannot fit all markets

The market is always in one of a few states

Markets do not behave one way. They cycle through states, and each state pays a different kind of trader.

Trending. Sustained direction. Higher highs or lower lows, with pullbacks that fail to reverse. Momentum gets paid. Mean reversion fights the move and gets carried out.

Ranging. Price oscillates in a band. Every breakout attempt fails and returns. Mean reversion gets paid. Momentum buys every false breakout and dies of a thousand stops.

High volatility. Huge bars both ways. Breakout traders and short-volatility traders live opposite lives here. One feasts. The other is the meal.

Low volatility. Compressed, quiet. Grids and market-making collect small coins steadily. Momentum starves waiting for a move that does not come.

Real markets are blends, and the boundaries are only clean in hindsight. The categories are still the right mental model, because they answer the only question that matters: what kind of edge is this market paying for right now?

Why a backtest needs regime eyes

A strategy backtests at plus 30 percent over three years. One number. Now decompose it by regime and look again. Perhaps all of it came from one six-month trend, and the other thirty months bled slowly.

Both stories end at the same equity point. They describe different strategies. One has broad edge. The other has one good period and a long leak. Deploy the second one, and unless its favorite regime returns soon, you get the leak without the windfall. The blended number could not warn you. The decomposition could. Always ask a backtest where the money came from, not just how much came.

How regimes get detected

Nothing exotic. Rolling volatility with thresholds separates loud from quiet. Trend filters, ADX or moving-average slope, separate directional from sideways. Where options exist, the gap between realized and implied volatility says what the market is bracing for. And crypto has its own tell: funding. Persistent positive funding means crowded longs. Persistent negative means crowded shorts. Funding shifts often lead to price-regime shifts, which makes funding one of the more useful native signals this market offers.

No detector is clean at the boundary. It does not need to be. Even rough labels turn “what happened” into “what happens when,” and that second question is the one worth paying for.

Drift against change

Two different events are constantly confused.

Drift is movement inside a regime. Volatility creeping up while the trend continues. Good strategies absorb drift. Sizing rules exist for it.

Change is a phase transition. Trend snaps to chop. Quiet snaps to spikes. Strategies built for the old state stop working, often abruptly, and no parameter inside the old assumptions fixes it. Catching the change early is among the highest-leverage skills in this business, because running the wrong strategy in the wrong regime compounds its cost daily, and the cost does not announce itself. It just accumulates.

What regime fit looks like

A trend strategy on BTC 4-hour bars paid handsomely in the long directional runs and bled through the extended chop between them. A funding-carry strategy collects while the funding regime persists and breaks when funding flips weekly and the carry never accumulates. A grid on a sideways pair collects in the range and gets liquidated when the real trend finally arrives.

Here is the test of whether you understand your own strategy. Can you name the market in which it should not be running? If you cannot, you do not yet know what you built. The strategy knows. It will tell you, expensively.

The honest conclusion

Edges are regime-specific. Regimes change. Hold both facts at once and only two coherent plans exist. Run one strategy and consciously eat the losses when its regime is absent. Or run several strategies with complementary regime fits, so something in the book is suited to whatever the market is doing. Serious operations end up at the second. See portfolio construction.

How Edgecraft handles this

Every Edgecraft backtest ships with a regime decomposition, so you see where the money was made and where it leaked instead of judging a blend. Strategies are easily explorable for the target regime and searchable directly.