Concept
Outlier Detection
Outlier Detection is a Statistics concept. The Library holds 3 implementations, each one a working definition you can pull into Quant.
Top Outlier Detection indicators
3 total
What is Outlier Detection?
Outlier detection flags observations that sit far outside the typical run of a series: a return, a volume print, or an indicator value that does not belong to the recent distribution. The standard tools are threshold rules. A z-score beyond a chosen multiple of standard deviations is the textbook version; robust variants use the median and MAD (median absolute deviation) so the outliers being hunted have little pull on the yardstick; interquartile fences and percentile cutoffs avoid normality assumptions; and sequential methods such as CUSUM (see change-point detection) catch shifts that accumulate rather than spike.
Markets complicate the textbook because returns are heavy-tailed: extreme prints occur far more often than a normal distribution implies, so a fixed three-sigma rule over-fires in volatile regimes and under-fires in quiet ones. Purpose also splits in two. One use is hygiene, catching bad ticks before they poison indicators; the other treats the outlier as the signal, since an abnormal volume spike or return jump is often exactly the event a strategy wants to react to.
How traders use it
- As data hygiene: filtering or winsorizing bad prints before computing averages and volatility estimates, since a single spike can distort every downstream indicator for its full lookback.
- As event detection: returns or volume far outside the recent distribution serve as triggers for breakout, reversal, or news-driven logic.
- As robust engineering: median-based filters and clipped inputs let indicators tolerate outliers instead of requiring their removal.
Related concepts · Normalization & distribution
Concept family
Statistics
45 concepts mapped · 37 in the Library
Outlier Detection FAQ
What threshold counts as an outlier in market data?
There is no universal cutoff. Two or three standard deviations is the convention under normality, but returns are heavy-tailed, so fixed sigma rules over-fire in volatile regimes and under-fire in calm ones. Robust thresholds built from the median and MAD, or plain percentile cutoffs, adapt better. The right choice depends on purpose: data cleaning tolerates false positives, signal generation does not.
Should outliers be removed from price data?
Only when they are errors: bad ticks, misprints, exchange glitches. Genuine extreme moves are part of the market's real distribution, and deleting them biases volatility and risk estimates toward calm conditions that will not hold. A common compromise is winsorizing, capping values at a percentile for model inputs, while keeping the raw series for risk measurement.
Build Outlier Detection your way.
Quant writes, tests, and refines it with you — then it runs on LuxAlgo charting or ports to TradingView.