Radar Perene / concept
Logarithmic Regression Cone
The Logarithmic Regression Cone frames the current price of an asset or index against its long-term historical trajectory, on a logarithmic scale.
What it is
A logarithmic regression fits the long-term price trend (the "central line") over a declared historical window; the cone is the band showing typical deviations around that line, expressed by bands calibrated by the historical standard deviation of the residual.
Price above the upper band: asset stretched relative to its own long-term trajectory. Price inside the cone: within the typical band. Price below the cone: asset contracted.
Why it exists
Because the question is the asset expensive or cheap? is hard when the asset has lived through decades of different regimes. The Cone normalizes that question: is it expensive or cheap relative to its own long-term trajectory, with trend discounted?
The logarithmic scale matters because returns compound multiplicatively. On a linear scale, IBOV at 30,000 looks "far" from IBOV at 130,000 — on log, percentage distance is what counts.
How to read it
| Position | Reading |
|---|---|
| Above the upper band | Asset stretched against long-term trajectory |
| Between upper and central band | Above trend, within the typical band |
| Near the central line | On the long-term trend axis |
| Between central and lower band | Below trend, within the typical band |
| Below the lower band | Asset contracted against long-term trajectory |
What it does not mean
It does not mean price will revert to the mean. Markets can remain in deviation for years. The cone describes current asymmetry; mean reversion is a statistical hypothesis that may take long or never occur within the reader's horizon.
It does not mean a price forecast. The cone is descriptive — here is where price is relative to its history. The Radar never forecasts the destination point.
Limits
The chosen historical window affects the slope of the regression and the width of the cone. The Radar declares window and parameters used in each chart. Assets with structural change (split, deep regulatory shift, recent listing) may have an unstable cone.
Example
IBOV — Logarithmic Regression Cone, window 2000-present. Current price 14% above the central line, within the upper typical band. No extreme deviation; trajectory within the expected historical range.
Related concepts: ERP_BR · Historical Analogs · see the full method