Logistic Regression

Question 1: What is the Logistic Regression indicator?

A 'learn-as-it-goes' model that predicts whether the next move is more likely up or down and labels it long or short (long = betting up, short = betting down). Despite the name, it's a yes/no classifier.

Question 2: How does it learn?

Every new bar it re-studies recent history, adjusts its internal weights to best separate past 'up' bars from 'down' bars, and outputs a probability that becomes the signal.

Question 3: Does it repaint?

Yes - its signals can change after the fact, so only act once the bar has fully closed, and back it up with a tool that doesn't repaint.

Question 4: What are the key settings?

How much history it normalises over (lower for fast markets), how many times it studies the data, its learning speed, and whether it uses raw price in the signal.

Question 5: How should I trade it?

Treat it as a directional lean, only take signals that match the bigger trend and only after the candle closes, and tune the settings for the market you're on.