Z-Score

TL;DR. (value − mean of the last n values) / their standard deviation. 0 is the window's average; 2 means two standard deviations above it, an unusual stretch for this window; −2 the mirror. One scale for every instrument and price level.

#What it measures

• Stretch from a recent average. How far the value has moved from its own mean, in units of its own recent variability.
• A unit-free reading. Dividing by the standard deviation removes the price scale, so a Z-Score of 2 means the same thing on a stock at 20 and an index at 5,000.
• The idea behind the Bollinger Bands. Price at the upper Bollinger Bands band is a Z-Score of about 2 at the same period. The Z-Score node gives you the number itself.

#Formula

#Inputs in AlgoLift

#Recommended periods

• 10 to 20 (fast): Short stretches inside a session.
• 30 (default): A balanced window.
• 100 to 200 (slow): How unusual today is against a longer regime.

#Outputs in AlgoLift

#How to read it

• Above 2: the value is unusually high for this window. A mean-reversion rule reads it as stretched; a momentum rule as strong.
• Below −2: the mirror, unusually low.
• Crossing back inside ±2: the stretch is easing, the moment many fade rules enter.
• Persistently on one side of 0: a trend. The mean is chasing the value and the score stays positive (or negative) for a long time.

#Best in, worst in

Best in ranging markets and mean-reversion rules, where an extreme score tends to unwind toward the average.

Worst in strong trends: the score stays high, and a rule that fades every reading above 2 keeps selling into strength.

#Three setups

#1. Fade the stretch

Enter short when the Z-Score crosses below 2 from above (a Crossover node reading the line and a constant), long when it crosses above −2 from below. Exit at 0, through a Comparison wired into Set Exit Condition.

#2. Stretch with a trend filter

Only fade toward the trend: buy a Z-Score below −2 only while price is above its 200-bar average. The filter keeps the rule out of the falls that keep falling.

#3. Breakout strength

Read a Z-Score above 2 as strength instead: enter long when it crosses above 2 and the session is trending, with the stop placed through a Bracket Order.

#Common mistakes

• Treating ±2 as a probability. It is a distance in the window's own deviation, not a percentage chance.
• Fading in a trend. An extreme score in a trend is the trend. Add a filter or a confirmation.
• Tuning the period and threshold together to the decimal. A rule that works at period 27 and 2.1 and fails nearby has found noise (overfitting in trading).