Mean Reversion to Climatology (Regression)

Overview

This pillar identifies extreme weather anomalies and predicts their regression to the long-term climatological average. It leverages the statistical principle that unusual weather patterns are temporary and tend to balance out over time.

What It Does

It analyzes current weather data against a 30-year historical baseline to quantify how unusual the present conditions are. The pillar then calculates the historical probability of such an anomaly persisting versus reverting to the mean. This provides a clear signal on whether to bet for or against the continuation of extreme weather.

Why It Matters

Public perception and market odds often overreact to recent extreme weather, creating a recency bias. This pillar offers a data-driven counter-signal, identifying valuable opportunities to bet against the crowd's assumption that the trend will continue indefinitely.

How It Works

First, the pillar establishes the climatological norm for a specific location and time period using 30-year average data. It then calculates the current deviation from this norm, often expressed in standard deviations (Z-score). Finally, it analyzes historical records to determine how long similar deviations have lasted in the past, generating a probability of reversion.

Methodology

The core calculation is the Z-score of a 14-day rolling average weather metric (e.g., temperature) against the 1991-2020 climatological mean for that same period. When the Z-score exceeds +/- 1.5, the pillar queries a historical database for all past events with a similar score, calculating the median duration before reversion to within +/- 0.5 standard deviations.

Edge & Advantage

This pillar provides a systematic edge by capitalizing on the market's recency bias during and after significant weather events.

Key Indicators

• Climatological Z-Score

  high

  Measures how many standard deviations the current weather is from the 30-year average, indicating the anomaly's severity.

• Anomaly Duration

  high

  The number of consecutive days the current anomaly has persisted, providing context on its maturity.

• Historical Reversion Time

  medium

  The average time it took for similar past anomalies to return to the mean, setting a predictive timeframe.

Data Sources

• NOAA NCEI

  Provides long-term U.S. and global historical climate data (normals, extremes) from the National Oceanic and Atmospheric Administration.

• Copernicus Climate Change Service (C3S)

  European Union's Earth Observation program, offering comprehensive global climate reanalysis datasets.

Example Questions This Pillar Answers

• → Will the average temperature in Phoenix for August 2024 be below 95°F, following a record-hot July?
• → Will total rainfall in London for the next 30 days be above the historical average, after a 6-week dry spell?
• → Will the number of days below freezing in Chicago this December be greater than 10?

Tags