Reading the market regime before prices move

Skip to content
Last close 00:43 UTC
QQQ 743.57 -0.34%
IWM 276.71 -0.43%
RSP 213.30 +0.06%
VIX 15.28 +0.46%
SVIX 30.11 +0.14%
VXX 16.68 -0.06%
SMH 588.12 -0.49%
SOXX 543.90 -0.18%
XLF 54.82 -0.42%
XLE 67.17 +1.76%
GLD 381.67 -0.13%
Brent 125.44 -0.06%
WTI 96.24 +0.11%
Dollar index 121.60 +0.13%
US2Y 4.79% −1 basis point
US5Y 5.00% −2 basis points
US10Y 5.21% −3 basis points
US20Y 5.60% −5 basis points
US30Y 5.56% −4 basis points
Bitcoin 83,266.00 +0.12%
After hours
INDICATOR REFERENCE · Indicator reference

Sixty-four technical and economic indicators in Arabic, with their limits

For each indicator: how it is read and where it goes wrong, a chart computed from its own formula, and an indicator to read with it. The limit is part of the definition: a tool used outside its context produces worse decisions than having none. Completely free, no sign-up.

  • 64indicators
  • 6families
  • 21live indicators in the platform tools
  • 100%Free, no sign-up

How do you read the charts?

  • One path for all the charts: a rally that loses momentum, then a sideways range that narrows, then a decline that starts with a gap. So you can see on the same charts which indicator is faster and which lags.
  • Orange marks where the indicator errs on this path: a leg under the chart or a specific session. Its error is written on the card below.
  • The numbers on the axis are the indicator's own thresholds (70 and 30 on RSI), not prices. The path is a worked example, not a real stock price, and the economic charts illustrate the idea.
Reference path: a worked exampleRisingSidewaysFallingPriceVolume
The reference path: 120 hypothetical sessions, with each session's volume below.

Trend

Trend · 21

It shows which way price is heading and how strongly. Most of it lags by nature, and it errs in sideways markets.

DMA

SMA
Simple moving average: a worked exampleIt errs in the sideways leg of the reference path.DMA-20

Price above it = uptrend, below it = downtrend. Its slope matters more than its position.

Where does it err? Lagging by nature: it confirms, it does not lead. In a sideways market it produces back-to-back opposite signals.
Read it with ADX: the average shows direction, and ADX measures its strength: below 20, ignore its crosses.

EMA

EMA
Exponential moving average: a worked exampleIt errs in the sideways leg of the reference path.EMA-20DMA-20

A fast line crossing above a slow one = rising momentum. The gap between them measures its strength.

Where does it err? Its speed costs noise: more false signals on short timeframes.
Read it with Average True Range: the distance of price from the average is read in units of ATR, not dollars, so tickers can be compared.

ADX

ADX
Average directional index: a worked exampleIt errs in the downtrend leg of the reference path.ADX-14

Above 25 = a real trend · below 20 = sideways. It does not tell the direction.

Where does it err? Reading it as an uptrend is a common mistake · it shows strength, not direction.
Read it with RSI: above 25 means the extreme of RSI can last, so don't read it as a reversal.

Parabolic SAR

Parabolic SAR
Parabolic SAR: a worked exampleIt errs in the sideways leg of the reference path.

The position of the dots sets the direction of the trailing stop.

Where does it err? In a sideways market it flips repeatedly, and commissions eat the account.
Read it with ADX: its flips in a sideways market drop when read with ADX above 20 only.

Aroon

Aroon
Aroon: a worked exampleAroon-UpAroon-Down

Aroon-Up above 70 with Down below 30 = an emerging uptrend.

Where does it err? Slow in fast-reversing markets.
Read it with MACD: Aroon catches the start of a trend, and MACD. confirms its acceleration.
Weighted average: a worked exampleIt errs in the sideways leg of the reference path.WMA-20DMA-20

A middle ground between the simple and exponential averages in speed and smoothness.

Where does it err? It shares the structural lag of all averages.
Read it with On-Balance Volume: the average's slope is more believable when OBV rises with it.

Double EMA

DEMA
Double exponential average: a worked exampleIt errs in the sideways leg of the reference path.DEMA-20EMA-20

Closer to price than the ordinary exponential average at turning points.

Where does it err? Cutting the lag adds wobble around peaks.
Read it with Average True Range: its speed produces many signals, so place the stop at a multiple of ATR, not right under the line.

Triple EMA

TEMA
Triple exponential average: a worked exampleIt errs in the uptrend leg of the reference path.TEMA-20EMA-20

The fastest to react in the moving average family.

Where does it err? The most exposed to noise · not suited to entries on its own.
Read it with ADX: the noisiest of the averages, followed only when ADX is above 25.
Triangular average: a worked exampleTRIMA-20EMA-20

The smoothest line in the family. It shows the broad trend.

Where does it err? It is the slowest of them all: not suited to timing.
Read it with RSI: it draws the broad trend, and RSI. measures how stretched the move is inside it.
Adaptive average: a worked exampleKAMA-10EMA-10

It whipsaws less in sideways markets, where fixed averages fail.

Where does it err? Hard to calibrate · a wrong setup turns it into an ordinary average.
Read it with ADX: it slows in a sideways market, and ADX below 20 confirms the market really is sideways.
Mean adaptive average: a worked exampleMAMAFAMA

A MAMA/FAMA crossover signals a turn in the cycle.

Where does it err? It assumes a cyclicality that may not exist in a news-driven market.
Read it with Hilbert mode: the Hilbert mode tells you whether the cycle dominates, and only then does the cross of MAMA with FAMA. count.
T3 average: a worked exampleT3-5TRIMA-20

High smoothness with less lag than the triangular average.

Where does it err? The volume factor changes its behavior radically · it needs testing.
Read it with Average True Range: smooth with less lag, so set the stop under it at a multiple of ATR.
Aroon oscillator: a worked exampleIt errs in the sideways leg of the reference path.Aroon Osc

Above zero the uptrend leads, below it the downtrend leads · and distance from zero shows the strength of the trend, not its direction.

Where does it err? Near zero it swings and produces conflicting signals · read it with ADX, not alone.
Read it with ADX: near zero it swings, so read it when ADX confirms the trend has strength.
Extended Parabolic SAR: a worked exampleIt errs in the sideways leg of the reference path.

Finer control over how fast the stop accelerates.

Where does it err? More parameters = more chances of overfitting.
Read it with Average True Range: compare its acceleration setting with the ATR distance so the stop doesn't sit in the noise.
Minus directional movement: a worked example−DM

Raw material, not a signal: -DI and then ADX are built from it. When it rises, selling pressure is widening.

Where does it err? Never use it alone · an absolute number not normalized by the true range cannot be compared between assets.
Read it with Average True Range: a raw number divided by the true range gives −DI, and ADX. is built from it.
Plus directional movement: a worked example+DM

Counterpart of the negative line: +DI is built from it. A rise means buying pressure is widening.

Where does it err? Like its counterpart: an intermediate input, not a signal. Read alone, it gives confidence with nothing behind it.
Read it with Average True Range: divided by the true range it gives +DI, and alone it can't compare two tickers.
Directional movement index: a worked example2520DX-14ADX-14

The raw ADX before smoothing.

Where does it err? Very jumpy · ADX is better for reading.
Read it with MACD: when it rises with MACD above zero, the uptrend is strong; below zero, it is a downtrend.

ADX rated

ADXR
Rated ADX: a worked exampleIt errs in the downtrend leg of the reference path.ADXR-14ADX-14

Smoother than ADX · for classifying regimes, not for timing.

Where does it err? Double lag.
Read it with DMA: above 25 the average's slope reads as a trend; below it, ignore its crosses.

Momentum

Momentum · 17

It measures the speed and stretch of a move. It works in a range and errs when it reads a strong trend as a reversal.

RSI

RSI
Relative strength index: a worked exampleIt errs in the uptrend leg of the reference path.RSI-14

Above 70 overbought · below 30 oversold.

Where does it err? ⚠ In a strong trend it stays extreme for weeks · an extreme is not a reversal signal by itself.
Read it with ADX: when ADX is above 25, an extreme can last for weeks, so don't read it as a reversal.

MACD

MACD
Moving average convergence divergence: a worked exampleIt errs in the sideways leg of the reference path.0MACDSignal line

A cross above zero confirms a trend, and the bars measure how fast it is picking up.

Where does it err? It lags like its components · its divergence sometimes precedes the reversal and sometimes lies.
Read it with ADX: its crosses in a market with no trend are noise, and ADX shows when to count them.
Stochastic: a worked exampleIt errs in the uptrend leg of the reference path.

A cross in the extreme zone is stronger than merely reaching it.

Where does it err? It produces many signals · needs a trend filter.
Read it with DMA: its crosses are more reliable in the direction of the DMA-50 slope than against it.
Commodity channel index: a worked exampleIt errs in the uptrend leg of the reference path.+100−1000CCI-20

Above +100 unusual strength · below −100 unusual weakness.

Where does it fail? No upper bound: a reading of +300 is possible and does not mean a reversal.
Read it with Average True Range: it has no upper bound, and ATR tells you whether a large reading is volatility expanding rather than a reversal.

Williams %R

Williams %R
Williams %R: a worked exampleIt errs in the uptrend leg of the reference path.−20−80%R-14

Above −20 overbought · below −80 oversold.

Where does it err? Almost a mirror of the Stochastic · don't combine them.
Read it with DMA: it mirrors the Stochastic, so they don't combine. Read it in the direction of the DMA-50. slope.
Extended MACD: a worked exampleMACD (DMA)Signal line

Read it exactly like MACD. Its value is that changing the average type slows the signal or speeds it up, depending on the asset.

Where does it err? Free tuning tempts you to fit the past (overfitting): settings that shine on one stretch of history and fail after it.
Read it with ADX: like the crosses of MACD: read them with ADX above 20, not in a sideways market.
Fast stochastic: a worked example

Reacts fastest. Suited to short timeframes.

Where does it err? High noise · rarely used alone.
Read it with EMA: its noise eases when read only in the direction of EMA-21.

Stochastic RSI

Stochastic RSI
Stochastic RSI: a worked example

The more sensitive of the two. It catches extremes inside extremes.

Where does it err? It hits its limits often · a signal without context is worthless.
Read it with Bollinger Bands: its extreme deserves attention when price touches the Bollinger Band with it.

Momentum

Momentum
Momentum: a worked example0MOM-10

Its signal tells you the direction of acceleration.

Where does it err? Raw and uncalibrated · hard to compare across tickers.
Read it with Average True Range: its value is in price units, so divide by ATR to compare tickers.
Rate of change ratio: a worked example1ROCR-10

Above 1 up, below it down.

Where does it err? The same limits as the rate of change.
Read it with On-Balance Volume: a ratio above 1 is more believable when OBV is rising.

TRIX

TRIX
TRIX: a worked example0TRIX-15

A zero cross is a trend signal, and a divergence is an early warning.

Where does it err? Triple smoothing = triple lag.
Read it with On-Balance Volume: its zero cross comes late, and OBV may catch a shift in participation sooner.
Ultimate oscillator: a worked exampleUO

It gives fewer false divergences than a single-window oscillator.

Where does it err? Its complexity makes the signal hard to interpret.
Read it with ADX: fewer false divergences, and it stays an oscillator, read when ADX is below 25.
Balance of power: a worked exampleBOPDMA-14

It measures who won the candle: the buyer or the seller.

Where does it err? Very noisy on a single candle · needs smoothing.
Read it with Accumulation/Distribution line: the candle's balance builds up in the Accumulation/Distribution line, which is its smoother reading.
Chande momentum oscillator: a worked exampleIt errs in the uptrend leg of the reference path.+50−50CMO-14

±50 usual extreme limits.

Where does it err? It resembles RSI and shares its flaw in a strong trend.
Read it with ADX: it shares the flaw of RSI, so read its extreme when ADX is below 25.
Percentage price oscillator: a worked example0PPOSignal line

Comparable across symbols with different prices · unlike MACD.

Where does it err? It inherits the full lag of MACD.
Read it with Normalized ATR: both are relative, so you can compare a ticker's momentum and volatility with others.
Absolute price oscillator: a worked exampleAPO

It measures how wide the gap between the averages is.

Where does it err? Not comparable across tickers.
Read it with Average True Range: its gap is in price units, so read it as a ratio to ATR, not by its raw value.

Volatility

Volatility · 4

It measures the size of the move, not its direction. Stop distance and position size are built on it.

Bollinger Bands: a worked exampleIt errs in the uptrend leg of the reference path.Bollinger BandsDMA-20

A shrinking range comes before expansion · a tight range warns of a coming move.

Where does it err? Touching the band is not a sell or buy signal · in a trend it is touched repeatedly.
Read it with MACD: a squeeze doesn't give direction, and MACD does when price leaves the range.
Normalized average true range: a worked exampleNATR-14

Lets you compare the volatility of a $20 stock with one at $800.

Where does it err? It shares ATR's silence on direction.
Read it with Percentage Price Oscillator: both are relative, so you can compare a ticker's volatility and momentum with others.

True Range

True Range
True range: a worked exampleTRATR-14

The building block of ATR · shows instantaneous volatility.

Where does it err? One candle does not make a reading.
Read it with Bollinger Bands: one session doesn't make a reading, and widening Bollinger Bands confirm that volatility has shifted.

Volume

Volume · 4

It asks whether volume joins the move. Its divergence from price is its best-known warning.

On-Balance Volume: a worked exampleOBV

Its divergence from price is the best-known early warning: price rises while OBV falls = a rally without participation.

Where does it err? Its absolute value means nothing · only the direction matters.
Read it with RSI: its divergence at the same time as RSI diverges is a stronger warning than either alone.
Accumulation/distribution line: a worked exampleIt errs on a gap-down session, in the first leg down: it rises while the close is lower than the prior one.A/D

Rising with price confirms it · a divergence between them is a warning.

Where does it err? It ignores gaps between sessions.
Read it with MACD: its divergence from price is more believable when MACD loses momentum with it.
Money flow index: a worked exampleMFI-14

Above 80 / below 20 is an extreme backed by volume · more reliable than RSI alone.

Where does it err? It needs clean volume data · it goes astray in thin markets.
Read it with ADX: its extremes in a strong trend can last, and ADX shows when.
Chaikin oscillator: a worked example0ADOSC

A zero cross shows when the accumulation balance changed.

Where does it err? Sensitive to the choice of windows.
Read it with DMA: its zero cross is more believable when the slope of DMA-20. agrees.

Cycles

Cycles · 8

Looks for a cycle in price and measures its length and phase. It fails when there is no cycle at all.

Hilbert sine wave

HT Sine Wave
Hilbert sine wave: a worked exampleSineLead Sine

The two waves crossing suggests a cycle top or bottom.

Where does it err? It fails completely when trend dominates the cycle.
Read it with ADX: it fails when the trend dominates, and ADX above 25 says it does.
Hilbert mode: a worked exampleIt errs in the sideways leg of the reference path.TrendCycleTrend Mode

A filter that says which indicator family fits right now.

Where does it err? Frequent switches near the turn.
Read it with ADX: it shows which family fits, and ADX confirms it with an independent calculation.
Hilbert cycle: a worked exampleIt errs in the uptrend leg of the reference path.14Dominant CycleActual cycle

It calibrates the windows of the other indicators automatically.

Where does it err? It assumes a cycle exists at all.
Read it with RSI: the length of RSI is tuned to it: half a cycle instead of a fixed 14.

Hilbert trendline

HT Trendline
Hilbert trendline: a worked exampleHT TrendlineEMA-20

It separates the trend component from the cycle.

Where does it err? Heavy math · its meaning is not intuitive.
Read it with ADX: it separates the trend from the cycle, and ADX measures the strength of that trend.

Midpoint

MIDPOINT
Midpoint: a worked exampleMIDPOINT-14

A simple reference for the center of the range.

Where does it err? It ignores what lies between the two ends.
Read it with ADX: it works as a reference for the middle of the range when ADX below 20 confirms the market is sideways.

Mid price

MIDPRICE
Midprice: a worked exampleMIDPRICE-14

The center of the actual range, not of the closes.

Where does it err? Sensitive to outlier candles.
Read it with Average True Range: outlier candles move it, and ATR reveals them when it jumps.
Hilbert phase: a worked example0°180°Phase

It tells you where we are in the cycle, not where we are going.

Where does it err? Sensitive to noise.
Read it with Stochastic: it shows where the cycle stands, and the Stochastic confirms when it is near a peak or trough.
Hilbert phasor: a worked example0InPhaseQuadrature

Raw material for building custom filters.

Where does it err? It is not read directly.
Read it with Hilbert cycle: the cycle length is derived from it, so read it through that rather than directly.

Macro

Economic · 10

US economic data that moves rates and markets. It arrives late and gets revised, so read it as a trend, not a single number.

Nonfarm payrolls: an illustrative chart0First releaseAfter revision

The release that moves the market most · published the first Friday of the month.

Where does it err? It is revised twice, by large amounts · the first number is not final.
Read it with unemployment rate: the same report's unemployment from another survey, so when they agree it is stronger than either alone.

Real GDP

Real GDP
Real GDP: an illustrative chart0Two negative quartersReal growth

Two negative quarters in a row = the conventional definition of a recession.

Where does it err? Published two months late and revised twice · not suited to timing.
Read it with nonfarm payrolls: it comes out quarterly and late, and monthly payrolls are a faster pulse.
GDP per capita: an illustrative chartConsecutive yearsAverage per personMedian

A more precise gauge of living standards than the total.

Where does it err? It hides inequality in the distribution.
Read it with unemployment rate: living standards really improve when falling unemployment goes with them.

Retail sales

Retail Sales
Retail sales: an illustrative chartNominalAfter inflation

The fastest pulse of consumption, which is two-thirds of the US economy.

Where does it err? It is nominal, not real: inflation lifts it with no actual growth.
Read it with annual inflation: it is nominal, so subtract inflation to read it as real growth.
Durable goods orders: an illustrative chartHeadlineCore, excluding transport

A leading indicator of business and investment confidence.

Where does it err? Aircraft and defense cause misleading jumps · watch the core figure.
Read it with nonfarm payrolls: company confidence is more believable when hiring follows.

Indicators are measuring tools, not orders to execute. None of them alone makes a decision. The value is in independent evidence lining up with the macro setting, which is why each indicator here comes with another one to read beside it, not a lookalike that repeats its reading.