📈 M-Competitions (Makridakis Competitions)

M-Competitions (Makridakis Competitions)#

Influential forecasting competitions benchmarking time-series methods.

Important

✨ AI-generated content. This page was written with the assistance of an AI language model and is provided as a learning aid. Despite careful review, it may still contain mistakes, omissions, or out-of-date information. Whether you are new to the topic, a team lead, or a senior practitioner, treat it as a starting point rather than an authoritative reference: read it critically and independently verify anything you act on (code, commands, figures, and factual claims) against official documentation and primary sources before relying on it.

What they are#

The M-competitions are a series of large-scale forecasting competitions (M1 through M6) organized by Spyros Makridakis and colleagues to gather empirical evidence about which forecasting methods actually work best in practice — not just in theory.

Key findings#

Across the early competitions no single method dominated, and simple methods (naïve, exponential smoothing, ARIMA) proved tough baselines that often matched or beat more complex statistical models; combining forecasts reliably improved accuracy. M4 (2018; 100,000 series, 61 methods) found the best results came from hybrid statistical + ML approaches and combinations, while pure-ML methods fared poorly. M5 (2020; Walmart hierarchical retail data on Kaggle, roughly $100,000 in prizes) was the first in which ML methods dominated the leaderboard, and it put probabilistic / uncertainty forecasting center stage.

Why they matter#

For four decades the M-competitions have shaped forecasting — establishing that combinations and hybrids win, that simple baselines must always be checked, and that probabilistic forecasting is now the standard. They directly inspired modern ML forecasting competitions (such as those on Kaggle).

Theme: Signal Processing & Time Series · All terminology

Hint

Mind map — connected ideas

Forecasting Competitions · Forecasting Benchmarks · Naïve Baseline Forecast · Forecast Error · Probabilistic Forecasts · Time Series Forecasting

Hint

More in Signal Processing & Time Series

ARIMA (AutoRegressive Integrated Moving Average) · Bayesian Time Series · Forecast Error · Forecasting Benchmarks · Forecasting Competitions · Log-Space · Low-pass Filtering · LSTM — Long Short-Term Memory Networks · Naïve Baseline Forecast · Prophet — Time Series Forecasting by Facebook (Meta) · Seasonal Lag · Seasonality · Signal Processing · Simple Baseline Methods