Home › Documentation › Validation & robustness
Analysis 5

Market regime

Available in the Regime tab of the Performance window. It answers two questions: which market environment does this strategy live in? and is the market on its ground RIGHT NOW?

Each trade is classified on its entry bar along four axes. Each axis is measured as a percentile against the previous 500 bars (the past only: what was known at entry), so it is a recent, relative regime: "high" means high compared with the recent history at that moment. Each axis is split into three levels.

Trend · Low / Medium / High

Is price moving efficiently or zigzagging? Measured with the Efficiency Ratio(20): net progress divided by the path travelled.

Momentum · Against / Neutral / In favour

How far had price already run in the trade's favour? The 20-bar move measured in ATR and oriented to the trade's direction. In mixed strategies it is seen from the longs.

Volatility

Was the market calm or nervous? ATR(20) as a percentage of price.

Compression / expansion

Was volatility shrinking or expanding? ATR(5) ÷ ATR(50).

THE MARKET NOW and its verdict

At the top, a header with the key data (trades, period, symbol · TF and the Profit Factor in and out of sample). Below it, THE MARKET NOW: the four axes on the last bar, each with its percentile bar and an amber dot where it stands today, and a cautious verdict:

Favourable ground

Only if on at least 2 axes the current level is clearly better than overall (PF ≥ ×1.15 with 30 trades or more), it holds out of sample and no axis is against it.

Unfavourable ground

The same the other way round: the market is where this strategy performs worst, out of sample too. Caution.

Neutral ground

The current regime does not match any level clearly better or worse for this strategy.

If the strategy has no out-of-sample stretch, the verdict is marked as indicative only. There is no "78/100"-style score: there would be no statistical basis for one.

📋 Diagnosis

One table per axis: trades and %, net, average, PF, win rate and PF in / out of sample for each level, with the median value of the axis across those trades and the NOW tag on the current level.

  • ·Green / red stripe: better / worse than overall.
  • ·Green background: better and confirmed out of sample.
  • ·Alarm red: out-of-sample PF below 1.
  • ·Dimmed (⚠): levels with fewer than 30 trades; their figure is not reliable.
🗺️ Map and charts
  • ·The real histogram of each axis over the last 500 bars, with today's dot.
  • ·Trend × Volatility map: the Profit Factor of each combination, with the scale centred on the strategy's PF.
  • ·Bubbles: one per trade (green won, red lost, size = result), with an axis selector and the NOW dot.
  • ·A sample and validation panel, with its legend.
🧭
A diagnosis, not a rule

A level that only shines in sample is usually chance. If you decide to add a regime filter to the strategy, validate it out of sample and with Walk-Forward like any other condition: picking the best level after looking at this table is overfitting. The price series is needed, with about 160 bars of history before the trades; the first ones may remain unclassified. On narrow screens, the tables switch to a single column.

Try it yourself

AniQuant can be tried free for 14 days, with every module and no card. If you have the book, 30.

← Previous
Monte Carlo simulation
Next →
Survival test
More in Validation & robustness