Market Regime
Also known as: Market State, Market Condition, Trading Regime, Regime Filter
What is it?
A market regime is the prevailing behaviour of a market over a stretch of time - trending, ranging, or volatile - and it decides which strategies have an edge and which quietly bleed. Regimes are measured, not felt. One common test compares how far price actually travelled against how far it ended up: if EUR/USD covers 340 pips of movement in a week but closes only 25 pips from where it opened, price is oscillating and the regime is ranging.
| Regime | What price does | What tends to work | What tends to fail |
|---|---|---|---|
| Trending | Closes far from where it started; successive higher highs or lower lows | Trend following, breakout entries, trailing stops | Fading extremes, fixed take-profits, range scalping |
| Ranging | Travels a long way but ends near the open, inside a defined band | Mean reversion, selling the band edges, tighter targets | Moving-average crossovers, breakout entries |
| Volatile | Wide bars, gaps and direction changes inside a single session | Smaller size, wider stops, or standing down entirely | Tight stops, scalping, position-doubling after a loss |
Cover the same 340 pips and close 280 from the open, and the same rules now describe a trend. Readings like ADX, ATR or an efficiency ratio put a number on that distinction so a bot can act on it. The practical point is that a strategy is not good or bad in the abstract - it is good or bad for a regime.
A moving-average crossover system is built to capture sustained direction, so it earns its keep in a trending regime and gives that back in a ranging one, without a single rule having changed.
Why it matters: Most strategies only have an edge in one market regime, so knowing which regime you are in tells you whether to trade your system or stand aside.
Efficiency ratio = |close(n) - close(0)| / sum of |close(i) - close(i-1)|
Running a trend-following bot through a ranging regime is one of the most common reasons a strategy that backtested well starts losing in live trading.
Real-world example
EUR/USD spent much of 2023 inside 1.0500-1.1100. A crossover bot tuned on the trending 2022 data kept firing inside that band and handed back a large share of its earlier gains without a single broken rule.
How SignalBots handles it
SignalBots tags each signal with the market conditions its model detected, so you can pause a trend-following bot when conditions turn choppy instead of finding out three losing trades later. Your capital is at risk.
See /risk-warning.
Pro tip
Read the regime on a timeframe at least four times higher than the one you trade - a 4H chop is invisible on the 15m chart where your bot is placing orders.
Common pitfalls
Re-optimising the strategy after every losing week, when the rules were never the problem - only the regime changed, and it will change back.
Frequently asked questions
How many market regimes are there?
Most frameworks use three - trending, ranging and volatile - because each rewards a different edge: continuation, reversion, or staying flat. Some models split trending into up and down, giving four or five states.
Can a bot detect the regime automatically?
Yes. Bots commonly read ADX, ATR or an efficiency ratio on a higher timeframe and either switch rule sets or stand down when the reading leaves the band the strategy was built for.
How long does a regime last?
There is no fixed length. FX ranges can hold for weeks or months, while a volatility regime around a rate decision may last hours. Treat duration as unknowable and react to the measurement rather than a calendar.
Does the regime change the win rate or the reward-to-risk?
Usually both, but the reward-to-risk moves further. A trend system in a range keeps a similar win rate while losing its large winners, so expectancy can turn negative even though the rules are untouched.
Should I trade in every regime?
No. Sitting out the regime your strategy was not built for is itself a decision. Most retail systems have an edge in one or two of the three, and forcing trades in the third is where backtested results get given back.
Trading involves substantial risk of loss. Historical and backtested results do not guarantee future performance. Read the full risk warning.
