Regimes

Regime detection, and its limits

Every strategy has conditions in which it works. The tempting conclusion is to hold whichever strategy suits the current conditions. The uncomfortable evidence is that regimes are only reliably identifiable after they end.

In short

A market regime is a persistent behavioural state — trending, range-bound, high or low volatility — that favours some strategies and punishes others. Regimes are recognisable in hindsight and ambiguous in real time, so holding several archetypes at once is more dependable than switching between them.

What a regime actually is

A regime is not a forecast; it is a description of how price has been behaving. Trend regimes reward continuation, range regimes reward fading, and volatility regimes change the cost of being wrong in both.

Regimes also differ by market. Hyperliquid majors and long-tail perps can occupy opposite regimes simultaneously, which is one reason a basket spanning several markets behaves differently from one concentrated in a single ticker.

Trending: sustained directional drift; trend and breakout archetypes earn.
Range-bound: repeated rejection at boundaries; mean reversion earns.
High volatility: wider stops, higher slippage, larger outcomes both ways.
Compressed volatility: funding and spread strategies dominate.

Why switching underperforms holding

Detecting a regime change requires enough data to distinguish it from noise, and by the time that data exists a meaningful part of the move has happened. Acting early produces false switches; acting late produces the worst of both regimes.

Holding uncorrelated archetypes concurrently sidesteps the problem entirely. You give up the upside of being perfectly positioned in exchange for never being completely wrong-footed.

How the system responds instead

Rather than predicting regimes, scores are recomputed continuously so allocation drifts toward whatever is currently producing durable results, and probation removes leaders whose records degrade. That is a reactive mechanism, not a predictive one, and it is deliberately described as such.

Past performance is not indicative of future results. Perpetual futures are leveraged instruments and carry a substantial risk of loss, including the loss of your entire position.

Questions

Frequently asked

Does the system predict regimes?

No. It reacts to changes in trader performance rather than forecasting market conditions.

Can regimes be detected at all?

They can be described with confidence in hindsight and estimated with considerable uncertainty in real time. The uncertainty is the point.

Do all markets share one regime?

No. Different perp markets can be in different regimes at the same time, which is why market spread matters alongside strategy spread.

How fast does allocation react?

Scores update continuously against fresh data; weight changes follow score changes rather than a fixed schedule.

Is a bad month evidence a strategy is broken?

Usually not. It is more often evidence that its regime is absent. Persistent degradation across regimes is the real signal.

Keep reading

Strategy driftWhen a leader quietly changes archetype.Combining strategiesHolding several regimes at once instead of picking one.Trend followingThe archetype most sensitive to regime absence.Live model performance7D and 30D score-weighted ROI estimates from public trader history.

Start Autopilot View live performance

Diversified copy trading. On autopilot.

Score-weighted allocation across up to 10 elite Hyperliquid traders, each isolated in its own sub-account. Your funds never leave your account.