Trend & directionVariable Index Dynamic Average · VIDYA
Tushar Chande's volatility-scaled EMA that reacts faster when the market is moving strongly.
Works in most conditionsEngine-computed on a fixed sample series
What it is
The Variable Index Dynamic Average (VIDYA) is an adaptive moving average created by Tushar Chande that changes its own speed based on how volatile the market is. A normal exponential moving average always reacts at the same pace, which forces a compromise: fast enough to catch trends means jumpy in quiet markets, while smooth enough to filter noise means slow to turn. VIDYA breaks that compromise by speeding up when the market is moving strongly and slowing down when it goes quiet, so it hugs price during real trends and flattens out to filter chop during consolidations. For a beginner, picture a moving average with an automatic transmission that shifts gears with market conditions, rather than being stuck in one gear like an ordinary EMA.
How it is calculated
VIDYA starts from the standard exponential moving average formula, where a smoothing constant alpha, equal to 2 divided by the period plus one, controls how much weight the newest price gets. The innovation is to multiply that alpha by a volatility ratio, usually called k, so the effective smoothing rises and falls with market activity. In Chande's later and most common version, k is the absolute value of the Chande Momentum Oscillator divided by 100, a number between 0 and 1 that is high when momentum is strong; the original version instead used the ratio of a short-period standard deviation to a longer-period standard deviation. Each new VIDYA value equals alpha times k times the current price, plus one minus alpha times k times the previous VIDYA. When volatility or momentum is high, alpha times k is large and the average tracks price closely; when the market is calm, alpha times k shrinks toward zero and the line barely moves, effectively filtering noise.
Reading it, step by step
Read VIDYA much as you would any moving average, but pay special attention to its changing character. A sloping, responsive VIDYA that is tracking price closely signals an active, trending market worth participating in. A flat, sluggish VIDYA signals a quiet, low-conviction market where the line is deliberately ignoring minor wiggles, and those flat phases are your cue that there is no trend to trade. Price crossing above a rising VIDYA suggests an uptrend is engaging, while price crossing below a falling one suggests a downtrend. Because the line adapts, a crossover during a high-volatility phase is more meaningful than one during a flat phase, where the line is barely responding. In essence, the slope tells you the direction and the responsiveness tells you whether a trend is genuinely present.
Best timeframes and settings
VIDYA typically uses a base period around 14 for the EMA component and a similar or longer window for the volatility or momentum measure that drives k, and it can be applied from intraday charts up to daily and weekly. The right settings depend heavily on the instrument's typical behavior, because the whole point is to tune the volatility input to what normal versus active looks like for that market. A shorter base period and a more sensitive volatility gauge make VIDYA quicker and better for shorter swings, at the cost of more false engagements; a longer base and smoother volatility input make it more deliberate and better for position trading. The core trade-off remains responsiveness versus noise, but VIDYA lets you offload part of that trade-off to the market itself, which is its main appeal over a fixed EMA.
When and where to use it
VIDYA is most valuable in markets that alternate between trending and ranging, because its adaptivity is designed precisely to lean in during trends and sit out during chop. It suits trend-following approaches that suffer from whipsaws when a fixed moving average keeps triggering during quiet periods. It works across asset classes as long as the volatility input is tuned sensibly for that instrument. It is less useful in a market that is either perpetually trending, where a simple EMA would do, or perpetually choppy, where no moving average helps much. Avoid deploying it with default settings blindly on an unfamiliar instrument, since a poorly matched volatility measure can leave it either too jumpy or too sluggish, undermining the very adaptivity that justifies choosing it.
Strategies that use it
A first strategy is a slope-and-cross trend follow: go long when price crosses above VIDYA while the line is rising and responsive, exit when price closes back below it, and stand aside entirely during flat phases where the line signals no trend. A second is an adaptive stop and trend filter: use VIDYA as a trailing reference, staying in a long as long as price holds above the line, letting the average tighten naturally when volatility rises and loosen when it falls. A third pairs VIDYA with a fixed EMA of the same length: when VIDYA pulls decisively away from the EMA, the market is trending and worth following, whereas when they overlap and both go flat, the market is ranging and trend trades should be avoided. Each approach exploits VIDYA's ability to distinguish active markets from dead ones.
Combining it with other indicators
VIDYA sits naturally beside other adaptive averages such as Kaufman's Adaptive Moving Average and the Fractal Adaptive Moving Average, and comparing them can confirm whether the market is genuinely trending. Because its speed is driven by a volatility gauge, an explicit volatility indicator like Average True Range or a standard-deviation band helps you understand why VIDYA is speeding up or slowing down. A momentum oscillator such as the Chande Momentum Oscillator, which underlies one version of k, or the Relative Strength Index adds confirmation of the strength behind a VIDYA engagement. Volume tools can validate that a trend VIDYA is tracking has real participation behind it. Avoid stacking it with several fixed-length moving averages that all say the same thing, since VIDYA's value is its adaptivity, not another redundant trend line.
Where it fails
VIDYA's greatest strength is also its greatest liability: its behavior depends entirely on the volatility measure and lengths you choose, so a badly tuned VIDYA can be worse than a plain EMA, either flapping around too eagerly or lagging too far behind. The extra volatility input adds complexity and one more thing to get wrong, and beginners often accept defaults that do not fit their instrument. In markets that transition abruptly, the volatility gauge can lag the shift, leaving VIDYA momentarily in the wrong gear. It still fundamentally lags price, as all moving averages do, just less than a fixed EMA during active phases. Guard against these problems by tuning the volatility input to the instrument, by backtesting settings rather than trusting defaults, and by treating VIDYA's flat phases as a genuine instruction to stay out rather than a glitch.
A worked example
Take a VIDYA with a base period of 14, so alpha is 2 divided by 15, about 0.133, and let k be driven by the absolute Chande Momentum Oscillator. During a quiet, range-bound stretch the momentum reading is weak, say an absolute CMO of 40, giving k equal to 0.40; the effective smoothing is 0.133 times 0.40, about 0.053, which is very slow, so VIDYA flattens and ignores the sideways noise. Then the stock breaks into a strong trend and momentum surges to an absolute CMO of 80, so k rises to 0.80 and the effective smoothing jumps to 0.133 times 0.80, about 0.107, roughly doubling its responsiveness. Now VIDYA tightens against price and tracks the advance closely, having automatically shifted into a faster gear exactly when a real trend appeared, which is the adaptive behavior a fixed moving average simply cannot deliver.