Regime Feed

Every trader talks about regimes. Almost every regime indicator on the market cheats: it relabels the past once the future arrives, so the historical chart looks clairvoyant and the live edge does not exist. The Regime Feed is built around one design constraint that forbids the cheat.

The constraint: never repaint, by construction

Once the feed prints a regime label for a bar, that label is final — prefix-stable by construction. The label you see live is the label the history will show forever. This costs accuracy at regime boundaries, and we say so on every surface: honest uncertainty at the edge is the price of a history you can actually trust. Any regime product whose historical labels are cleaner than its live ones is showing you a backtest of hindsight.

What it publishes

• Regime states. State-space identification of volatility and liquidity regimes per instrument — trending, mean-reverting, transition, stress — each with its model definition pre-registered before the feed goes live.
• Uncertainty, visibly. Each label ships with its confidence, and boundary bars are marked as boundary bars instead of being silently sharpened later.
• Transition receipts. When the model changes state, the inputs that drove the transition are recorded and recomputable — a regime call you can audit, not an oracle you must trust.

What it does not do

It does not predict the next regime. No regime model honestly can. It identifies the one you are in, with stated lag and stated uncertainty. That is less exciting than clairvoyance and considerably more real.

Who it is for

Systematic traders who condition sizing or strategy selection on market state and need the conditioning variable to be as honest live as it looks in the backtest.