Market Regimes:
How to Detect and Trade Them

The same strategy can thrive in one market environment and bleed in the next. This guide covers what market regimes are, why volatility regimes are the easiest to forecast, and three data-driven regime signals, plus 215 ready-made regime signals to download for Build Alpha.

By David Bergstrom··21 min read

What Is a Market Regime?

A market regime is a persistent state of market behavior, defined by characteristics such as trend direction, volatility level, and correlation, that lasts long enough to change which trading strategies work. Regimes shift because the participants and pressures driving price shift.

Some common drivers of market regime: monetary policy, dealer hedging, investor positioning, and risk appetite.

A strategy's backtest is an average across every regime in the data. If the next six months look like the calm half of that history, a trend-following system may struggle while a mean-reversion system outperforms. If the market looks like 2008 or 2020, the opposite happens. Knowing the current regime is context a price-only strategy desperately needs.

Market Regimes: Quick Summary

  • A regime is a persistent market state, usually described by trend and volatility
  • Volatility regimes are far easier to forecast than direction, because volatility clusters
  • Dealer gamma, market breadth, and positioning add regime information beyond volatility
  • Regime signals work best as filters: they decide when a strategy should trade

The Four Common Market Regimes

We can start with two dimensions, trend and volatility, and create four regimes. Real markets are messier, but this is enough of a starting point.

Uptrend · Low Vol
Calm Bull

Steady grind higher, shallow pullbacks, low realized volatility. The most common regime.

Favors: trend following, buy-the-dip, premium selling

Uptrend · High Vol
Volatile Bull

Rising prices with large daily swings, often late in a rally or early in a recovery.

Favors: wider stops, smaller size, momentum with confirmation

Sideways or Down · Low Vol
Quiet Range

Little net progress and contained ranges. Breakouts tend to fail and revert.

Favors: mean reversion, fading extremes, range strategies

Downtrend · High Vol
Crisis

Sharp declines, violent rallies, correlations converge toward one. Rare. Dangerous.

Favors: short-term breakouts, reduced exposure, or standing aside

Most regime-aware traders spend their effort on the volatility axis rather than the trend axis, and for good reason, as the next section shows.

Volatility Regimes Persist: The Best Predictor of Future Volatility

Past volatility is the single best predictor of future volatility. Volatility clusters. Calm periods tend to stay calm and turbulent periods tend to stay turbulent, a property economists have documented since the ARCH and GARCH models of the 1980s.

Scatter plot of S&P 500 realized volatility over the past 21 days versus realized volatility over the next 21 days sampled weekly, showing strong volatility persistence
S&P 500 realized volatility over the past 21 days vs. the next 21 days, sampled weekly

That is why so many regime filters start with a volatility measure: realized volatility percentile, ATR relative to its average, or the VIX.

The chart above looks a month ahead. The three signals below look one day ahead, at the size of tomorrow's S&P 500 range, which is what matters most for stop placement, position sizing, and whether a short-term strategy should trade at all.

Three Regime Signals Beyond Past Volatility

Single-Stock Gamma
Lower gamma → bigger next-day range
Strength 0.62
% Above 200-Day SMA
Weaker breadth → bigger next-day range
Strength 0.45
VIX Specs − Gamma
Crowded specs, low gamma → bigger next-day range
Strength 0.43

Strength is the rank correlation with the next day's S&P 500 range, from 0 (no relationship) to 1 (perfect).

1. Single-Stock Gamma (GEX500)

GEX500 is gamma exposure aggregated across the options on every S&P 500 component stock. When it is high, dealers are long gamma and their hedging leans against price moves, which compresses ranges. When it falls, the brake comes off and dealers push price to extremes.

Scatter plot of S&P 500 single-stock gamma exposure GEX500 versus next day S&P 500 range showing large ranges cluster when gamma is low or negative
The lower single-stock gamma, the bigger tomorrow's range. Nearly every range above 6% happened when gamma was low or negative.

When single-stock gamma is negative, the biggest next-day ranges in the data show up. For the full breakdown of how dealer gamma works and 36 ready-made gamma signals, see the Gamma Exposure (GEX) guide.

2. Market Breadth: Percent of Stocks Above Their 200-Day Average

Breadth measures participation. When most S&P 500 stocks are above their 200-day average, the market is in a healthy, broad-based uptrend. When that percentage collapses, the index is held up by a shrinking group of names, and ranges widen.

Scatter plot of percent of S&P 500 stocks above their 200-day moving average versus next day S&P 500 range showing wider ranges when breadth is weak
The fewer stocks above their 200-day average, the bigger tomorrow's range.

The Market Breadth guide shows how the 20-day version predicts next-day range even after controlling for recent volatility, plus 90 downloadable breadth signals.

3. VIX Speculator Positioning Minus Single-Stock Gamma

This one combines two data sets. From the CFTC Commitment of Traders report, take where VIX futures speculators are positioned relative to the past year. Subtract where single-stock gamma sits relative to the past year. A high reading means speculators are leaning one way while dealer gamma is low: crowded positioning with no brake.

Scatter plot of VIX futures speculator positioning percentile minus single-stock gamma percentile versus next day S&P 500 range
The higher the reading (crowded VIX speculators, low gamma), the bigger tomorrow's range. Held up before and after 2015.

Weekly COT positioning data rarely lines up with a next-day target on its own, but paired with a faster series it becomes a usable regime input. Holding gamma constant, VIX positioning still carries its own information about tomorrow's range. This is exactly the kind of multi-source condition Build Alpha's search engine is built to find.

Method Notes

S&P 500 measured with SPY daily high, low, and close, which has no futures roll adjustments. Next-day range is (high − low) divided by the prior close. Gamma and positioning values are read after the close and compared to the following session, so the entire target happens after the data is known. Percentiles use a trailing one-year window.

How to Detect the Current Market Regime

There is no single correct regime model. Each approach has a trade off between speed, stability, and complexity.

MethodHow it worksStrengthWeakness
Trend filterPrice vs. 200-day MASimpleSlow, whipsaws
Volatility percentileVol ranked vs. historyPersistentNo direction
VIX term structureVX1 vs. VX3Forward-lookingEquity only
Dealer gammaOptions GEX, GEX500Next-day rangeNeeds options data
Market breadth% above MA, A/DShows participationStocks only
PositioningCOT, AAII, NAAIMCaptures crowdingWeekly, slow
Statistical modelsHMM, clusteringMany inputsEasy to overfit

In practice, the most robust regime filters combine a volatility measure with one or two inputs from a different family, which is exactly what the three charts above aim to show.

Using Regime Filters in Build Alpha

Every series on this page is a built-in Build Alpha signal source: options gamma, breadth, VIX futures, COT positioning, sentiment, yields, and more. Any of them can be used three ways:

  • As a filter. Set a regime condition as a required signal so a strategy only trades when the environment fits, such as breakouts only when gamma is low or mean reversion only when breadth is healthy.
  • As a sizing input. Trade smaller when regime signals point to wider ranges and larger when they point to calm.
  • As a strategy selector. Build separate strategies for different regimes and let the regime signal decide which one is active.

Select regime signals and let the genetic search decide which ones improve a strategy and which ones are noise. Regime conditions apply to any instrument, an equity regime signal can filter strategies on gold, bonds, crude oil, or currencies just as easily as on the S&P 500.

Build Alpha interface showing custom market regime signals loaded for use as strategy filters
Market regime signals loaded in Build Alpha, ready to use as filters, entries, or exits

215 Custom Market Regime Signals

The download below adds 215 regime signals to Build Alpha's Custom Indicators list, built from every data family on this page: price structure, volatility, the VIX curve, the yield curve, dealer gamma, breadth, sentiment, and COT positioning. Each one works as an entry, an exit, or a filter, and every threshold and lookback can be optimized.

Expand a family to see the custom signals in it. These are meant as ideas and starting points, not absolute truths.

Trend & VolatilityATR ratios, percentiles, MA structure14 signals+
Trend / Price Structure · 6
  • ATRShort_abv_ATRLongShort-term ATR is above long-term ATR, indicating a higher-volatility short-term regime.
  • ATRShort_blw_ATRLongShort-term ATR is below long-term ATR, indicating a lower-volatility short-term regime.
  • MovAvgDistancePrice is more than roughly +3 ATR above its moving average, indicating strong upside extension.
  • MovAvgDistance_below_neg3Price is more than roughly -3 ATR below its moving average, indicating strong downside extension.
  • MA_bull_structureMoving averages are arranged in a bullish trend structure.
  • MA_bear_structureMoving averages are arranged in a bearish trend structure.
Volatility / ATR Regimes · 8
  • ATR_Ratio_abvoveShorter-term ATR relative to long-term ATR is above the selected high-volatility ratio threshold.
  • ATR_Ratio_belowShorter-term ATR relative to long-term ATR is below the selected low-volatility ratio threshold.
  • ATR_ExpandingATR is increasing versus its prior value, identifying expanding volatility.
  • ATR_ContractingATR is decreasing versus its prior value, identifying contracting volatility.
  • ATR_90_percentileATR is in an unusually high historical volatility regime near the 90th percentile.
  • ATR_10_percentileATR is in an unusually low historical volatility regime near the 10th percentile.
  • Range_countA high proportion of recent bars have larger short-term ATR than the prior longer-term ATR.
  • Range_count_belowA low proportion of recent bars have larger short-term ATR than the prior longer-term ATR.
VIX Term StructureContango vs. backwardation6 signals+
  • VX1_above_VX3Front-month VIX futures are sufficiently above VX3, identifying backwardation/stress.
  • VX1_below_VX3Front-month VIX futures are sufficiently below VX3, identifying stronger contango.
  • Vix_Curve_WideningThe VX1–VX3 curve spread is widening versus the prior period.
  • Vix_Curve_NarrowingThe VX1–VX3 curve spread is narrowing versus the prior period.
  • VX1_VX3_Spread_Above_SMA20The VX1 minus VX3 spread is above its own 20-period moving average.
  • VX1_VX3_Spread_Below_SMA20The VX1 minus VX3 spread is below its own 20-period moving average.
Yield Curve10Y minus 2Y spread4 signals+
  • TenYr_minus_TwoYrThe 10-year minus 2-year Treasury yield spread is above the selected positive threshold.
  • TenYr_minus_TwoYr_belowThe 10-year minus 2-year Treasury yield spread is below the selected negative threshold.
  • YieldCurve_10Y2Y_Above_SMA20The 10Y–2Y yield spread is above its own 20-period moving average.
  • YieldCurve_10Y2Y_Below_SMA20The 10Y–2Y yield spread is below its own 20-period moving average.
Dealer GammaGEX, GEX500, spreads, ratios, extremes38 signals+
Core GEX / GEX500 Regimes · 8
  • GEX_90_percentileGEX is in an unusually high historical regime near its 90th percentile.
  • GEX_10_percentileGEX is in an unusually low historical regime near its 10th percentile.
  • GEX_GEX500_spread_wideningThe spread between SPX GEX and component GEX is widening.
  • GEX_GEX500_spread_narrowingThe spread between SPX GEX and component GEX is narrowing.
  • GEX_GEX500_spread_ratioThe GEX/GEX500 relationship is above its selected ratio threshold.
  • GEX_GEX500_spread_ratio_belowThe GEX/GEX500 relationship is below its selected ratio threshold.
  • GEX_GEX500_Spread_Above_SMA20GEX minus GEX500 is above its own 20-period moving average.
  • GEX_GEX500_Spread_Below_SMA20GEX minus GEX500 is below its own 20-period moving average.
Additional GEX Spread / Ratio Regimes · 16
  • AbsSpread_Above_2500MThe absolute difference between GEX and GEX500 exceeds 2.5 billion.
  • AbsSpread_Below_2500MThe absolute difference between GEX and GEX500 is below 2.5 billion.
  • Ratio_Above_030The GEX/GEX500 ratio is above the specified 0.30 threshold.
  • Ratio_Below_030The GEX/GEX500 ratio is below the specified 0.30 threshold.
  • Spread_Above_SMA20The GEX minus GEX500 spread is above its 20-period moving average.
  • Spread_Below_SMA20The GEX minus GEX500 spread is below its 20-period moving average.
  • SpreadPct_Above_90thThe GEX/GEX500 spread is in its upper historical percentile regime.
  • SpreadPct_Below_10thThe GEX/GEX500 spread is in its lower historical percentile regime.
  • GEXpct_Above_GEX500pctGEX ranks higher historically than GEX500 on a percentile basis.
  • GEXpct_Below_GEX500pctGEX ranks lower historically than GEX500 on a percentile basis.
  • Zscore_spread_aboveThe standardized GEX–GEX500 spread is above the selected z-score threshold.
  • Zscore_spread_belowThe standardized GEX–GEX500 spread is below the selected z-score threshold.
  • Ratio_Above_SMA20The GEX/GEX500 ratio is above its own 20-period moving average.
  • Ratio_Below_SMA20The GEX/GEX500 ratio is below its own 20-period moving average.
  • RatioPct_Above_90thThe GEX/GEX500 ratio is in its upper historical percentile regime.
  • RatioPct_Below_10thThe GEX/GEX500 ratio is in its lower historical percentile regime.
GEX Momentum / Extremes · 14
  • GEX_Above_Max_Prior2Current GEX is above both of its previous two readings.
  • GEX_Below_Min_Prior2Current GEX is below both of its previous two readings.
  • GEX500_Above_Max_Prior2GEX500 is above both of its previous two readings.
  • GEX500_Below_Min_Prior2GEX500 is below both of its previous two readings.
  • GEX500_Above_90thPercentileGEX500 is in its upper historical percentile regime.
  • GEX500_Below_10thPercentileGEX500 is in its lower historical percentile regime.
  • Zscore_GEX_aboveGEX is unusually high relative to its recent mean and standard deviation.
  • Zscore_GEX_belowGEX is unusually low relative to its recent mean and standard deviation.
  • Zscore_GEX500_aboveGEX500 is unusually high relative to its recent distribution.
  • Zscore_GEX500_belowGEX500 is unusually low relative to its recent distribution.
  • GEX_NewHigh_20GEX is at a new 20-period high.
  • GEX_NewLow_20GEX is at a new 20-period low.
  • GEX500_NewHigh_20GEX500 is at a new 20-period high.
  • GEX500_NewLow_20GEX500 is at a new 20-period low.
Market BreadthTRIN, TICK, A/D, highs/lows, % above SMA85 signals+
Core Breadth Regimes · 8
  • Breadth_trendMarket breadth is improving versus its previous reading.
  • Breadth_trend_downMarket breadth is deteriorating versus its previous reading.
  • Breadth_spread_wideningThe selected breadth spread is widening.
  • Breadth_spread_narrowingThe selected breadth spread is narrowing.
  • Breadth_ratio_abvThe selected breadth ratio is above its regime threshold.
  • Breadth_ratio_blwThe selected breadth ratio is below its regime threshold.
  • Risers_2xAdvancing stocks are more than twice declining stocks.
  • Fallers_2xDeclining stocks are more than twice advancing stocks.
TRIN · 3
  • TRIN_Above_1_5TRIN is above 1.5, indicating unusually heavy downside pressure.
  • TRIN_Below_0_7TRIN is below 0.7, indicating unusually strong upside pressure.
  • TRIN_Below_PriorTRIN is lower than its prior reading, indicating easing downside pressure.
TICK · 4
  • TICK_Above_1000TICK exceeds +1000, indicating unusually broad simultaneous buying.
  • TICK_Below_Neg1000TICK is below -1000, indicating unusually broad simultaneous selling.
  • TICK_Above_MaxPrior2TICK is above both of its previous two readings.
  • TICK_Below_MinPrior2TICK is below both of its previous two readings.
Advance / Decline Breadth · 2
  • ADD_Above_0NYSE Advance-Decline Difference is positive.
  • ADD_Below_0NYSE Advance-Decline Difference is negative.
52-Week High / Low Breadth · 16
  • High52_Above_SMA20The number of 52-week highs is above its 20-period average.
  • High52_Below_SMA20The number of 52-week highs is below its 20-period average.
  • Low52_Above_SMA20The number of 52-week lows is above its 20-period average.
  • Low52_Below_SMA20The number of 52-week lows is below its 20-period average.
  • High52_Above_PriorMore stocks are making 52-week highs than in the prior period.
  • High52_Below_PriorFewer stocks are making 52-week highs than in the prior period.
  • Low52_Above_PriorMore stocks are making 52-week lows than in the prior period.
  • Low52_Below_PriorFewer stocks are making 52-week lows than in the prior period.
  • High52_Above_Low52More stocks are making 52-week highs than 52-week lows.
  • Low52_Above_High52More stocks are making 52-week lows than 52-week highs.
  • HighLowSpread_WideningThe difference between 52-week highs and lows is increasing.
  • HighLowSpread_NarrowingThe difference between 52-week highs and lows is decreasing.
  • High52_Low52_Ratio_Above_SMA20The highs/lows ratio is above its own 20-period moving average.
  • High52_Low52_Ratio_Below_SMA20The highs/lows ratio is below its own 20-period moving average.
  • High52_Low52_Spread_Above_SMA20Highs minus lows is above its own 20-period moving average.
  • High52_Low52_Spread_Below_SMA20Highs minus lows is below its own 20-period moving average.
Percentage of S&P 500 Stocks Up / Down · 16
  • PctUp_Above_70More than 70% of S&P 500 stocks are up.
  • PctUp_Below_30Fewer than 30% of S&P 500 stocks are up.
  • PctDown_Above_70More than 70% of S&P 500 stocks are down.
  • PctDown_Below_30Fewer than 30% of S&P 500 stocks are down.
  • PctUp_Above_PriorPercentage of stocks up is greater than its prior reading.
  • PctUp_Below_PriorPercentage of stocks up is lower than its prior reading.
  • PctDown_Above_PriorPercentage of stocks down is greater than its prior reading.
  • PctDown_Below_PriorPercentage of stocks down is lower than its prior reading.
  • PctUp_Above_PctDownThe percentage of stocks up exceeds the percentage down.
  • PctDown_Above_PctUpThe percentage of stocks down exceeds the percentage up.
  • PctUpDownSpread_WideningThe spread between % stocks up and % stocks down is increasing.
  • PctUpDownSpread_NarrowingThe spread between % stocks up and % stocks down is decreasing.
  • PctUp_PctDown_Ratio_Above_SMA20The % up/% down ratio is above its own 20-period average.
  • PctUp_PctDown_Ratio_Below_SMA20The % up/% down ratio is below its own 20-period average.
  • PctUp_PctDown_Spread_Above_SMA20% up minus % down is above its own 20-period average.
  • PctUp_PctDown_Spread_Below_SMA20% up minus % down is below its own 20-period average.
% of S&P 500 Stocks Above SMA20 · 6
  • ABV20_Above_80More than 80% of S&P 500 stocks are above their 20-day moving average.
  • ABV20_Below_20Fewer than 20% of S&P 500 stocks are above their 20-day moving average.
  • ABV20_Above_PriorThe percentage above SMA20 is increasing.
  • ABV20_Below_PriorThe percentage above SMA20 is decreasing.
  • ABV20_Above_50More than half of S&P 500 stocks are above SMA20.
  • ABV20_Below_50Fewer than half of S&P 500 stocks are above SMA20.
% of S&P 500 Stocks Above SMA50 · 6
  • ABV50_Above_80More than 80% of S&P 500 stocks are above their 50-day moving average.
  • ABV50_Below_20Fewer than 20% of S&P 500 stocks are above their 50-day moving average.
  • ABV50_Above_PriorThe percentage above SMA50 is increasing.
  • ABV50_Below_PriorThe percentage above SMA50 is decreasing.
  • ABV50_Above_50More than half of S&P 500 stocks are above SMA50.
  • ABV50_Below_50Fewer than half of S&P 500 stocks are above SMA50.
% of S&P 500 Stocks Above SMA200 · 6
  • ABV200_Above_80More than 80% of S&P 500 stocks are above their 200-day moving average.
  • ABV200_Below_20Fewer than 20% of S&P 500 stocks are above their 200-day moving average.
  • ABV200_Above_PriorThe percentage above SMA200 is increasing.
  • ABV200_Below_PriorThe percentage above SMA200 is decreasing.
  • ABV200_Above_50More than half of S&P 500 stocks are above SMA200.
  • ABV200_Below_50Fewer than half of S&P 500 stocks are above SMA200.
SMA20 vs SMA50 Breadth Relationship · 6
  • ABV20_ABV50_Ratio_Above_SMA20The ABV20/ABV50 ratio is above its own 20-period average.
  • ABV20_ABV50_Ratio_Below_SMA20The ABV20/ABV50 ratio is below its own 20-period average.
  • ABV20_ABV50_Spread_WideningABV20 minus ABV50 is increasing.
  • ABV20_ABV50_Spread_NarrowingABV20 minus ABV50 is decreasing.
  • ABV20_ABV50_Spread_Above_SMA20ABV20 minus ABV50 is above its own 20-period average.
  • ABV20_ABV50_Spread_Below_SMA20ABV20 minus ABV50 is below its own 20-period average.
SMA20 vs SMA200 Breadth Relationship · 6
  • ABV20_ABV200_Ratio_Above_SMA20The ABV20/ABV200 ratio is above its own 20-period average.
  • ABV20_ABV200_Ratio_Below_SMA20The ABV20/ABV200 ratio is below its own 20-period average.
  • ABV20_ABV200_Spread_WideningABV20 minus ABV200 is increasing.
  • ABV20_ABV200_Spread_NarrowingABV20 minus ABV200 is decreasing.
  • ABV20_ABV200_Spread_Above_SMA20ABV20 minus ABV200 is above its own 20-period average.
  • ABV20_ABV200_Spread_Below_SMA20ABV20 minus ABV200 is below its own 20-period average.
SMA50 vs SMA200 Breadth Relationship · 6
  • ABV50_ABV200_Ratio_Above_SMA20The ABV50/ABV200 ratio is above its own 20-period average.
  • ABV50_ABV200_Ratio_Below_SMA20The ABV50/ABV200 ratio is below its own 20-period average.
  • ABV50_ABV200_Spread_WideningABV50 minus ABV200 is increasing.
  • ABV50_ABV200_Spread_NarrowingABV50 minus ABV200 is decreasing.
  • ABV50_ABV200_Spread_Above_SMA20ABV50 minus ABV200 is above its own 20-period average.
  • ABV50_ABV200_Spread_Below_SMA20ABV50 minus ABV200 is below its own 20-period average.
SentimentAAII bulls, bears, neutral, NAAIM32 signals+
AAII Bulls / Bears · 10
  • Bulls_Above_50AAII bullish sentiment is above 50%.
  • Bulls_Below_50AAII bullish sentiment is below 50%.
  • Bulls_Above_PriorAAII bullish sentiment is increasing.
  • Bulls_Below_PriorAAII bullish sentiment is decreasing.
  • Bears_Above_50AAII bearish sentiment is above 50%.
  • Bears_Below_50AAII bearish sentiment is below 50%.
  • Bears_Above_PriorAAII bearish sentiment is increasing.
  • Bears_Below_PriorAAII bearish sentiment is decreasing.
  • Bulls_Above_BearsAAII bullish sentiment exceeds bearish sentiment.
  • Bears_Above_BullsAAII bearish sentiment exceeds bullish sentiment.
AAII Bull / Bear Relative Regimes · 6
  • BullBearSpread_WideningBulls minus Bears is increasing.
  • BullBearSpread_NarrowingBulls minus Bears is decreasing.
  • Bulls_Bears_Ratio_Above_SMA20The Bulls/Bears ratio is above its own 20-period average.
  • Bulls_Bears_Ratio_Below_SMA20The Bulls/Bears ratio is below its own 20-period average.
  • Bulls_Bears_Spread_Above_SMA20Bulls minus Bears is above its own 20-period average.
  • Bulls_Bears_Spread_Below_SMA20Bulls minus Bears is below its own 20-period average.
AAII Bulls vs Neutral · 6
  • Bulls_Neutral_Ratio_Above_SMA20The Bulls/Neutral ratio is above its own 20-period average.
  • Bulls_Neutral_Ratio_Below_SMA20The Bulls/Neutral ratio is below its own 20-period average.
  • Bulls_Neutral_Spread_WideningBulls minus Neutral is increasing.
  • Bulls_Neutral_Spread_NarrowingBulls minus Neutral is decreasing.
  • Bulls_Neutral_Spread_Above_SMA20Bulls minus Neutral is above its own 20-period average.
  • Bulls_Neutral_Spread_Below_SMA20Bulls minus Neutral is below its own 20-period average.
AAII Bears vs Neutral · 6
  • Bears_Neutral_Ratio_Above_SMA20The Bears/Neutral ratio is above its own 20-period average.
  • Bears_Neutral_Ratio_Below_SMA20The Bears/Neutral ratio is below its own 20-period average.
  • Bears_Neutral_Spread_WideningBears minus Neutral is increasing.
  • Bears_Neutral_Spread_NarrowingBears minus Neutral is decreasing.
  • Bears_Neutral_Spread_Above_SMA20Bears minus Neutral is above its own 20-period average.
  • Bears_Neutral_Spread_Below_SMA20Bears minus Neutral is below its own 20-period average.
NAAIM Exposure · 4
  • NAAIM_Above_80NAAIM exposure is above 80, indicating relatively high active-manager equity exposure.
  • NAAIM_Below_20NAAIM exposure is below 20, indicating relatively low active-manager equity exposure.
  • NAAIM_Above_PriorNAAIM exposure is increasing versus the prior reading.
  • NAAIM_Below_PriorNAAIM exposure is decreasing versus the prior reading.
COT PositioningS&P 500 spreads and acceleration36 signals+
COT: Commercials vs Non-Commercials · 10
  • Spread_Comm_NonComm_SP500_Above50The Commercial minus Non-Commercial COT spread is above +50.
  • Spread_Comm_NonComm_SP500_Below50The Commercial minus Non-Commercial COT spread is below the selected lower threshold.
  • AbsSpread_Comm_NonComm_SP500_Above50The absolute Commercial/Non-Commercial disagreement exceeds 50.
  • AbsSpread_Comm_NonComm_SP500_Below50The absolute Commercial/Non-Commercial disagreement is below 50.
  • Widening_Comm_NonComm_SP500_AboveThe Commercial/Non-Commercial positioning spread is widening in the positive direction.
  • Widening_Comm_NonComm_SP500_BelowThe Commercial/Non-Commercial positioning spread is widening in the negative direction.
  • Narrowing_Comm_NonComm_SP500_AboveA previously positive Commercial/Non-Commercial spread is narrowing.
  • Narrowing_Comm_NonComm_SP500_BelowA previously negative Commercial/Non-Commercial spread is narrowing.
  • Comm_NonComm_SP500_Spread_Above_SMA20Commercial minus Non-Commercial positioning is above its own 20-period average.
  • Comm_NonComm_SP500_Spread_Below_SMA20Commercial minus Non-Commercial positioning is below its own 20-period average.
COT: Commercials vs Nonreportables · 10
  • Spread_Comm_NonReport_SP500_Above50The Commercial minus Nonreportable COT spread is above +50.
  • Spread_Comm_NonReport_SP500_Below50The Commercial minus Nonreportable COT spread is below the selected lower threshold.
  • AbsSpread_Comm_NonReport_SP500_Above50Absolute disagreement between Commercials and Nonreportables exceeds 50.
  • AbsSpread_Comm_NonReport_SP500_Below50Absolute disagreement between Commercials and Nonreportables is below 50.
  • Widening_Comm_NonReport_SP500_AboveThe Commercial/Nonreportable spread is widening in the positive direction.
  • Widening_Comm_NonReport_SP500_BelowThe Commercial/Nonreportable spread is widening in the negative direction.
  • Narrowing_Comm_NonReport_SP500_AboveA positive Commercial/Nonreportable spread is narrowing.
  • Narrowing_Comm_NonReport_SP500_BelowA negative Commercial/Nonreportable spread is narrowing.
  • Comm_NonReport_SP500_Spread_Above_SMA20Commercial minus Nonreportable positioning is above its own 20-period average.
  • Comm_NonReport_SP500_Spread_Below_SMA20Commercial minus Nonreportable positioning is below its own 20-period average.
COT: Non-Commercials vs Nonreportables · 10
  • Spread_NonComm_NonReport_SP500_Above50The Non-Commercial minus Nonreportable COT spread is above +50.
  • Spread_NonComm_NonReport_SP500_Below50The Non-Commercial minus Nonreportable COT spread is below the selected lower threshold.
  • AbsSpread_NonComm_NonReport_SP500_Above50Absolute disagreement between Non-Commercials and Nonreportables exceeds 50.
  • AbsSpread_NonComm_NonReport_SP500_Below50Absolute disagreement between Non-Commercials and Nonreportables is below 50.
  • Widening_NonComm_NonReport_SP500_AboveThe Non-Commercial/Nonreportable spread is widening in the positive direction.
  • Widening_NonComm_NonReport_SP500_BelowThe Non-Commercial/Nonreportable spread is widening in the negative direction.
  • Narrowing_NonComm_NonReport_SP500_AboveA positive Non-Commercial/Nonreportable spread is narrowing.
  • Narrowing_NonComm_NonReport_SP500_BelowA negative Non-Commercial/Nonreportable spread is narrowing.
  • NonComm_NonReport_SP500_Spread_Above_SMA20Non-Commercial minus Nonreportable positioning is above its own 20-period average.
  • NonComm_NonReport_SP500_Spread_Below_SMA20Non-Commercial minus Nonreportable positioning is below its own 20-period average.
COT Positioning Acceleration · 6
  • Accel_Comm_SP500_Above10Commercial positioning has accelerated upward by more than the selected 10-point threshold.
  • Accel_Comm_SP500_Below10Commercial positioning has accelerated downward beyond the selected threshold.
  • Accel_NonComm_SP500_Above10Non-Commercial positioning has accelerated upward by more than 10.
  • Accel_NonComm_SP500_Below10Non-Commercial positioning has accelerated downward beyond the selected threshold.
  • Accel_NonReport_SP500_Above10Nonreportable positioning has accelerated upward by more than 10.
  • Accel_NonReport_SP500_Below10Nonreportable positioning has accelerated downward beyond the selected threshold.
Download all 215 Custom Market Regime Signals
Place the file in your BuildAlpha folder. Rename your existing CustomIndicators.xml first so you don't overwrite your own signals!
Download ↓
Want to keep your existing custom signals too? Upload both XML files (these signals and your current CustomIndicators.xml) to an AI assistant and ask it to merge them. Merging two XML files is a simple task for AI. Then place the merged file, named exactly CustomIndicators.xml, back into your BuildAlpha folder.

Example: A Mean Reversion Strategy vs. 215 Regime Filters

In this video, I take a simple mean-reversion strategy in Build Alpha and test it against all 215 Market Regime filters above. The winners are obvious fits that describe calmer markets, which mean reversion should benefit from.

Using Market Regimes to Size Positions

A regime filter is an on/off switch. It assumes the edge drops to zero when the filter is off. That is rarely the case. More often the edge shrinks, gets noisier, or both. If it is still positive, you still want exposure, just not as much.

That is what Build Alpha's Position Sizing Scaler does. Instead of a hard cutoff like "do not trade when the filter is above X," it scales position size continuously with the regime reading. As the scaler rises, size rises. As it falls, size shrinks. And if the average edge holds but the variance rises, that is a sizing problem by definition, not an on/off one.

Why Hard Filters Often Make a Strategy Worse

1. The threshold gets overfit. Take "skip the trade if VIX is above 20." Where did 20 come from? Usually from trying a few numbers and keeping the prettiest equity curve. The strategy is now fully on at 19.9 and completely flat at 20.1. No market changes that abruptly.

2. Filters throw away trades, and statistical confidence scales with the square root of the trade count. Here is how much a filter must improve the edge per trade just to keep the strategy's t-statistic where it was:

Trades keptEdge increase neededA t-stat of 2.50 becomes
80%+12%2.24
60%+29%1.94
40%+58%1.58
20%+124%1.12

Last column assumes no improvement in edge per trade.

Filter down to 40% of the original trades and the remaining trades need to be 58% better just to break even on confidence. Most filters are nowhere near clearing that bar.

What a Position Sizing Scaler Is

A scaler is simply a number you multiply your position size by, recomputed every bar. Anything you can calculate works, price-based or not. Some ideas:

  • (Close − SMA) / ATR
  • ATR252 / ATR20
  • Volume / Avg(Volume)
  • AAII Bulls / AAII Bears
  • COT Commercials − COT Nonreportables

Alternative data is especially useful here. The series behind the regime signals on this page, such as dealer gamma, breadth, and the VIX curve, all work as scalers too.

On / Off
Regime Filter
  • One lever, and it only moves one way: off
  • Can take trades away, but never lets you press
  • Hard threshold to overfit, with a cliff at the cutoff
  • Smaller sample, lower statistical confidence
Continuous
Position Sizing Scaler
  • Moves both directions: bet more when conditions favor you, less when they don't
  • Size follows the edge available
  • No threshold to fit and no cliff
  • Keeps every trade, so the sample never shrinks
A filter decides which trades you are allowed to take. A scaler decides how much you believe each one.

Testing Scalers in Build Alpha

Build Alpha's Position Sizing Scaler test runs any selected strategy through all of your saved scalers at once, so you can compare how each one changes the equity curve and risk-adjusted performance side by side. Find it under Robustness Testing → Pos Sz Scaler.

Build Alpha Position Sizing Scaler test running a trading strategy through every saved position sizing scaler
Robustness Testing → Pos Sz Scaler: one strategy, every saved scaler, compared side by side

Validate Before You Trade

Every regime filter adds parameters and ways to fit the past. A filter that only works at one exact threshold is noise. Run regime-filtered strategies through Build Alpha's robustness tests, especially out-of-sample, walk-forward, and parameter permutation. Regime filtering is not an excuse to skip validation.

Takeaways

✓ Key Takeaways

  1. A market regime is a persistent state, usually described by trend and volatility, that changes which strategies work
  2. Past volatility is the best single predictor of future volatility. A good place to start
  3. Single-stock gamma, breadth, and positioning add regime information about tomorrow's range beyond past volatility

💡 Need to Know

  1. Regime signals work best as filters, sizing inputs, and strategy selectors
  2. Combining a volatility measure with a second data family is more robust than either alone
  3. Build Alpha ships every series on this page built in, plus the 215 custom regime signals above, for any market

Market Regimes FAQ

A market regime is a persistent state of market behavior, defined by characteristics such as trend direction, volatility level, and correlation, that lasts long enough to change which trading strategies work. Common examples are calm bull markets, volatile bull markets, quiet ranges, and high-volatility declines.
The most common framework crosses trend and volatility into four regimes: uptrend with low volatility, uptrend with high volatility, sideways or down with low volatility, and downtrend with high volatility. Each favors different strategy types, such as trend following in calm uptrends and mean reversion in quiet ranges.
Popular methods include a trend filter such as the 200-day moving average, realized volatility percentile, the VIX and its futures term structure, dealer gamma exposure, market breadth, futures positioning, and statistical models such as hidden Markov models. Robust regime filters usually combine a volatility measure with one or two inputs from a different data family.
Yes. Volatility clusters, so calm periods tend to stay calm and turbulent periods tend to stay turbulent. Recent realized volatility is the single best predictor of near-term volatility, which is why other regime signals should be judged on whether they add information beyond it.
Often because the market regime changed. A backtest averages performance across every regime in the data, so a strategy built mostly on calm, trending years can struggle when volatility rises or the trend breaks. Regime filters let a strategy trade only in the environments where it has historically worked.
Yes. Equity regime signals such as dealer gamma, breadth, and VIX positioning describe market-wide risk appetite, which carries into gold, bonds, crude oil, and currencies. Build Alpha lets you apply any regime signal as a filter on any instrument or timeframe.
Add regime signals to your signal pool, or set one as a required signal so a strategy only trades when the condition is true. Build Alpha's genetic search then tests which regime conditions improve a strategy, and its robustness tests check whether the improvement holds out of sample.
Build Alpha algorithmic trading software desktop for building and validating regime-filtered strategies
Build Alpha: discover, validate, and export regime-filtered strategies with no coding
David Bergstrom Build Alpha founder and quantitative strategy developer

David Bergstrom

Founder, Build Alpha

A decade-plus in professional trading as a market maker and quantitative strategy developer at a high-frequency trading firm with a CME seat, consulting for CTAs and RIAs. Self-taught C++, C#, and Python programmer specializing in data science, machine learning, and trading strategy development.

What Traders Say About Build Alpha

Trader testimonial about Build Alpha regime filters and alternative data
User review of Build Alpha algorithmic trading platform
Positive feedback on Build Alpha alternative data signals
Build Alpha review on regime-filtered trading strategy generation
Trader recommending Build Alpha futures strategy tools
Review of Build Alpha genetic algorithm strategy generator
Algorithmic trader praising Build Alpha strategy automation
Build Alpha review for systematic trading strategy development
Positive testimonial about Build Alpha portfolio construction
Trader feedback on Build Alpha automated strategy generation
Build Alpha user review highlighting robustness testing
Quantitative trader endorsing Build Alpha trading software
Build Alpha testimonial on Monte Carlo simulation and validation
Trader praising Build Alpha walk forward optimization tools
Build Alpha customer review on automatic trading system generation
User feedback on Build Alpha no-code strategy builder
Professional trader testimonial about Build Alpha platform
Latest Build Alpha user review on regime analysis workflow

Trade with Regime Context

Gamma, breadth, VIX, and positioning data built in. Point-and-click regime filters for any market. No coding required.

No coding needed

Stop Wasting Time
Coding, Testing, Coding, Testing…

Build Alpha’s Genetic Algorithm uses your inputs to create 1000s of trading strategies in seconds

View Demo
Recent Posts
Get updates & ideas
Learn more about BA & get exclusive insights