Momentum & oscillators

Stochastic Momentum Index · SMI

William Blau's refinement of the stochastic that measures the close relative to the midpoint of the range, not its extremes.

Works best in ranging marketsEngine-computed on a fixed sample series
145120962080Above 80 = overboughtBelow 20 = oversold
SMI 86.68%D 87.66How to read SMI on the chart — the callouts mark what to look for.

The formula

Measure the distance from the close to the center of the recent range, and divide by half the range — then double-smooth both the numerator and the denominator with EMAs. Zero marks the midpoint of the range; the double smoothing makes SMI far less erratic than the raw stochastic.

Midpoint = (Highest High + Lowest Low) ÷ 2 SMI = 100 × EMA(EMA(Close − Midpoint)) ÷ ( EMA(EMA(Highest High − Lowest Low)) ÷ 2 )
Worked example
StepValue
Highest high (n)110
Lowest low (n)90
Midpoint (110 + 90) ÷ 2100
Close − midpoint+6
Half-range (110 − 90) ÷ 210

SMI ≈ 100 × 6 ÷ 10 = +60 before smoothing — the close is in the upper half, overbought above +40.

What the Stochastic Momentum Index is

The Stochastic Momentum Index, created by William Blau, is a refined version of the classic stochastic oscillator that measures where the current close sits relative to the midpoint of the recent trading range rather than relative to its extremes. That single change makes it a cleaner, more centred momentum gauge. Where the ordinary stochastic asks how close price is to the top or bottom of its range, the SMI asks how far above or below the middle of that range price has closed, then smooths the answer heavily. The result oscillates roughly between minus 100 and plus 100 and is centred on zero, so a positive reading means the close is in the upper half of the recent range and a negative reading means it is in the lower half. It answers the question: relative to its own recent range, is price leaning bullish or bearish, and how strongly?

How it is calculated

First the highest high and lowest low over the lookback period are found, and their midpoint is taken as the centre of the range. The SMI then measures the distance from the current close to that midpoint — a positive number if the close is above centre, negative if below. Both that distance and half of the full range are smoothed twice, typically with a pair of exponential moving averages, which is the double smoothing that gives the SMI its calm character. The index is then 100 times the double-smoothed distance divided by the double-smoothed half-range, which scales it into the minus 100 to plus 100 band. Finally a signal line, usually a short exponential average of the SMI itself, is plotted on top for crossovers. Typical settings are a 10-period range with 3-period double smoothing and a 3-period signal.

Reading it, step by step

The zero line is the reference: above zero the close is in the upper half of the recent range and momentum leans bullish, below zero it leans bearish. The extremes matter next — readings above plus 40 are considered overbought and below minus 40 oversold, marking points where price is stretched within its range. Because of the double smoothing, the SMI line is noticeably less jagged than a raw stochastic, so its turns and crossovers are cleaner and easier to trust. The main trigger is the SMI crossing its signal line: a cross up out of oversold territory is a buy cue and a cross down out of overbought a sell cue. A zero-line cross is a secondary, momentum-shift signal. As with all such oscillators, divergence between the SMI and price often warns of a turn before price itself confirms it.

Reading the signals on the chart

145120962080
SMI 86.68%D 87.66The ▲/▼ marks flag the most recent crossings of the 20 and 80 lines — the classic oversold / overbought signals.

Best timeframes

  • Scalping1m – 5msmoothing adds lag
  • Day trading5m – 15m
  • Swing1h – 4h
  • PositionDaily

Typical settings are a 10-bar range with 3-bar double smoothing and a 3-bar signal line; SMI is cleaner than the raw stochastic but turns later.

SMI vs other momentum oscillators

SMIStochasticRSI
Scale−100 to +1000 to 1000 to 100
Centered onZero5050
Close measured vsRange midpointRange high/lowAvg gain/loss
SmoothingDoubleSingleWilder

Common price-action setups

How the signal typically plays out on the chart.

Cross out of oversold

SMI turns up through its signal line from below −40 — buy the momentum shift with a stop under the recent swing low.

Buy the cross
Momentum shift up
Cross out of overbought

SMI rolls down through its signal line from above +40 — sell or exit longs with a stop above the recent swing high.

Sell the cross
Momentum shift down
Bullish divergence

Price makes a lower low but SMI makes a higher low, warning the downside is exhausting — buy the turn once SMI crosses its signal, stop under the low.

Buy divergence
Bullish reversal

Best timeframes and settings

The SMI's smoothing makes it well suited to swing trading on hourly, 4-hour, and daily charts, where its clean crossovers cut through the noise that plagues the raw stochastic. The standard configuration is a 10-period range with 3 and 3 double smoothing and a 3-period signal line, which balances responsiveness against calm. Shortening the range length or the smoothing periods makes the SMI quicker and more reactive but reintroduces some of the choppiness it was designed to remove. Lengthening them produces an even smoother line that lags more at turns but generates fewer false signals. Because the double smoothing already adds lag, scalpers on very fast charts sometimes find the SMI too slow and prefer a lighter oscillator, while position traders appreciate the steadiness of longer settings on daily and weekly data.

When and where to use it

Like the stochastic it descends from, the SMI is fundamentally a range and swing tool, most effective in markets that oscillate or trend gently rather than run in a straight line. It excels at timing entries within a range and at spotting momentum shifts through zero-line and signal-line crosses. Its overbought and oversold zones are most reliable in rangebound conditions, where reversion to the mean is the norm. In a strong, sustained trend the SMI, like every stochastic, will pin near its extreme and its reversal signals will fail repeatedly, so it should not be used to fade a powerful move. It suits liquid instruments across equities, futures, and FX. The best practice is to consult it for timing inside a trend or range you have already defined with a separate trend tool, not to call major turns on its own.

Strategies that use it

The primary strategy trades signal-line crossovers out of the extremes: buy when the SMI crosses above its signal line from below minus 40, and sell when it crosses below its signal line from above plus 40, placing stops beyond the recent swing. A second strategy uses the zero line as a momentum switch, going long when the SMI crosses above zero and short when it crosses below, which captures shifts in the balance of the range earlier but with more whipsaw. A third, higher-probability approach trades divergence: when price makes a new low but the SMI makes a higher low, it flags fading downside momentum and sets up a long once the signal-line cross confirms. Pairing any of these with a higher-timeframe trend filter, taking only signals in the trend's direction, markedly improves the hit rate.

Combining it with other indicators

The SMI's chief blind spot is trend context, so a trend filter is the natural partner. ADX tells you whether to trust the SMI's reversal signals — low ADX means a range where they work, high ADX means a trend where they should be ignored or only taken in the trend's direction. A longer moving average defines the bias so you only act on SMI crosses that align with it. Because the SMI already measures momentum, pairing it with a volume tool such as On-Balance Volume adds a participation check that momentum alone lacks. Support and resistance levels give the SMI's overbought and oversold signals a price context, so an oversold SMI cross right at a support shelf is a stronger buy than the cross in open space. The SMI supplies timing; trend and level tools supply the setting.

Where it fails

The double smoothing that makes the SMI so clean is also its weakness: it adds lag, so the SMI turns later than the raw stochastic at abrupt reversals, and a fast V-bottom can be half over before the signal fires. Like all stochastics it saturates in a strong trend, pinning above plus 40 or below minus 40 for long stretches while its overbought and oversold reversal signals fail again and again. Traders who fade those extremes in a trend get repeatedly run over. On very fast timeframes the lag can make it useless for precise entries. And because it is a momentum oscillator, it can diverge from price for a long time before price finally turns, so divergence is a warning, not a timer. The remedy is to respect the regime, wait for the signal-line confirmation, and never treat an extreme reading alone as a trade.

A worked example

Suppose a stock is chopping in a range on the daily chart, and over the last 10 bars the highest high is 52 and the lowest low is 48, putting the range midpoint at 50. Price closes at 51, one point above the midpoint, so before smoothing the SMI is positive and, after the double smoothing scales it, the line reads about plus 45 — into overbought territory above plus 40. The next day the SMI ticks down and crosses below its signal line while still above plus 40; that is your sell trigger. You short near 51 with a stop above the recent swing high at 52.20 and target the lower half of the range near the 48 area. Price rotates down over the following sessions to 48.50, where the SMI drops below minus 40, and you cover as it prepares to cross its signal line back up — a full swing captured from one extreme to the other.

Common mistakes

  • Trading its extremes in a strong trend, where SMI pins near +100 or −100 and reversals fail.
  • Forgetting the double smoothing adds lag, so it turns later than the raw stochastic.
  • Acting on the raw SMI alone and ignoring the signal-line cross.
  • Treating +40/−40 as automatic sell/buy without price confirmation.
  • Using the same thresholds for every market instead of tuning them.
  • Skipping the higher-timeframe trend that decides whether a cross follows through.