Klinger Volume Oscillator

A fast-minus-slow EMA of signed volume force — an indicator whose published formulas genuinely disagree across platforms.

What it is

The Klinger Volume Oscillator (KVO), attributed to Stephen Klinger, attempts to measure long-term volume-based money flow while staying responsive to shorter-term shifts. It does this by classifying each bar's volume as buying or selling pressure based on trend direction, then taking the difference between a 34-period and a 55-period EMA of that signed 'volume force', usually with a 13-period EMA signal line. An honest note before anything else: KVO is one of the most inconsistently implemented indicators in technical analysis. Klinger's original 'volume force' formula involves a trend-run accumulation term (cumulative and current daily ranges) and an absolute-value ratio, and published sources disagree on its exact form — as a result, MetaTrader, TradingView community scripts, and various books compute visibly different KVO values on identical data. The version documented here is the widely used simplified form (signed volume, where the sign comes from the direction of high+low+close), which is deterministic and reproducible; our Pine example implements exactly that, and we state so rather than pretending all versions match. As always: this is descriptive, educational material, not a prediction machine, and no indicator eliminates the risk of loss.

How it works

In the simplified form, each bar is first classified by trend: if today's high + low + close is greater than yesterday's, the bar is treated as accumulation and its volume takes a positive sign; otherwise it is distribution and volume takes a negative sign. This signed series is the volume force. The oscillator is then KVO = EMA(volume force, 34) − EMA(volume force, 55): when recent bars have been dominated by positively signed volume, the fast EMA sits above the slow one and KVO is positive; when negatively signed volume dominates, KVO is negative. A 13-period EMA of KVO serves as the signal line, and some platforms also plot the KVO-minus-signal difference as a histogram. Klinger's original formulation replaces the simple ±volume term with volume × |2 × (dm/cm) − 1| × trend × 100, where dm is the bar's high−low range and cm is a running accumulation of ranges that resets when the trend flips — this weights volume by range behaviour within a trend run. The two constructions can diverge meaningfully, which is exactly why you should verify which formula your platform uses before comparing readings across tools. Both versions are unbounded and centered on zero.

How traders read it

• Zero-line side: KVO above zero means recent signed volume force has favoured accumulation-classified bars; below zero, distribution-classified bars. It describes how volume has been classified, not where price is going.
• Signal-line crossings: KVO crossing its 13-period EMA is watched by some traders as a shift in volume-force momentum; in ranging markets these crossings occur frequently and reverse often.
• Divergence: price making a new low while KVO makes a higher low is read by some as selling volume force fading — a context clue that can persist for many bars without price responding.
• Cross-platform sanity check: before reading KVO at all, confirm which formula your platform implements — a signal on one platform's KVO may simply not exist on another's.
• Magnitude is instrument-relative: like all volume-scaled oscillators, KVO's amplitude depends on typical volume, so levels are only comparable within one chart's own history.

Common settings

The near-universal defaults are 34 and 55 periods for the fast and slow EMAs of volume force, with a 13-period EMA signal line — Fibonacci-flavored numbers that are convention, not optimization. Some implementations expose the choice between the simplified signed-volume force and the original trend-run formula; most do not, which is the root of cross-platform mismatches. Shortening the EMAs makes the oscillator more reactive and noisier; the settings interact with timeframe, so any comparison should state timeframe, lengths, and which volume-force formula was used.

Strengths

• Combines trend classification with volume in one series, rather than treating volume as directionless.
• The dual-EMA structure (34/55) explicitly separates a faster and slower view of volume force, analogous to MACD's treatment of price.
• Zero-centered and signal-lined, giving consistent reference points on any instrument.
• The simplified form is fully deterministic and easy to verify bar-by-bar against its stated formula.
• Its very inconsistency is instructive: comparing implementations is a practical lesson in why formula-level verification matters before trusting any indicator.

Pitfalls to watch

• Implementation ambiguity is the headline risk: published KVO formulas genuinely differ, so two platforms can show different oscillators under the same name — always verify the formula before comparing or automating anything.
• Binary trend classification is crude: a bar where hlc3 rises by a fraction of a pip contributes its entire volume as 'accumulation', identical to a strong trend bar.
• Heavy smoothing means heavy lag: 34/55-period EMAs respond slowly, so KVO confirms volume-force shifts well after they begin.
• Whipsaw on the signal line: the 13-period signal crossing produces frequent flip-flops in sideways markets despite the long base EMAs.
• Volume-data caveats apply in full: on spot forex, 'volume' is tick volume from one feed, so KVO there measures feed activity classified by trend, not actual money flow.
• No predictive power: KVO describes past signed volume. It forecasts nothing, and no formula variant removes the risk of loss.

Pine v6 example

//@version=6
indicator("Klinger Volume Oscillator (simplified signed-volume form)", overlay = false)

fastLen = input.int(34, "Fast EMA Length", minval = 1)
slowLen = input.int(55, "Slow EMA Length", minval = 1)
sigLen  = input.int(13, "Signal EMA Length", minval = 1)

// Trend classification: direction of high+low+close vs prior bar
hlcSum = high + low + close
trendDir = hlcSum > nz(hlcSum[1], hlcSum) ? 1.0 : -1.0

// Simplified volume force: signed volume (NOT Klinger's original trend-run formula)
volForce = trendDir * volume

kvo = ta.ema(volForce, fastLen) - ta.ema(volForce, slowLen)
sig = ta.ema(kvo, sigLen)

crossUp = ta.crossover(kvo, sig) and barstate.isconfirmed
plotshape(crossUp, "KVO Cross Up", style = shape.triangleup, location = location.bottom, color = color.teal, size = size.tiny)

plot(kvo, "KVO", color = color.blue)
plot(sig, "Signal", color = color.orange)
hline(0, "Zero", color = color.gray)

Pro tip: Treat KVO as a case study in verification: before relying on it anywhere, plot your platform's built-in version against a formula you implemented yourself (like the simplified form above) and check whether they match. If they do not, you have learned which formula your platform uses — and that any KVO commentary you read online may be describing a different indicator entirely. That habit of validating the formula, not the name, transfers to every tool you use. Educational information only, not financial advice; no indicator predicts price, and trading involves risk of loss.

Educational only — not financial advice, not a recommendation to trade. No indicator is predictive; trading involves substantial risk of loss.