SVI 203 · Tier 200 — Intermediate — Inference & Modeling · 2 credits

Technical Indicators: A Statistical Audit

RSI, MACD and Bollinger Bands on trial: what they actually compute, why their predictive claims fail honest statistical testing, and the statistical tools that do each job properly.

Prerequisites

SVI 102 Probability & the Return Model

Learning Objectives

State precisely what RSI computes and show it is a bounded transformation of past returns, containing no information the return series does not already hold.
Show that MACD is a difference of two exponential moving averages — a lagging band-pass filter of price with no predictive content of its own.
Explain why Bollinger Bands' apparent probability statement (±2σ ≈ 95%) is invalid for non-stationary, fat-tailed, volatility-clustered price data.

Module Schedule

01Module 1: RSI and Oscillators: The Claim Versus the Evidence18 min 02Module 2: MACD: A Lagging Filter Dressed as a Signal17 min 03Module 3: Bands and Ranges: What Survives Statistical Scrutiny18 min