Austerlitz Analytics

Napoleonic Infantry Behaviour — 1792 to 1815

This is Austerlitz Analytics — a site that predicts Napoleonic infantry formations based on the collection of hundreds of recorded interactions from the period 1792–1815. The data was hand-gathered and contains multiple vectors to best describe the situations troops faced, explaining their reactions to different variables.

      How it works: The site uses ML techniques — specifically
      k-Nearest Neighbours (k-NN) classification — to predict the action variable
      in the database. Given a set of battlefield conditions, the model finds the most
      historically similar engagements and votes on the most likely outcome.
    

Each engagement in the dataset is described by the following variables:

Training Level Officer Trait Formation Fatigue Terrain Time of Day Cavalry Charge Infantry Charge Officer Status Volley Fire Outflanked Outnumbered Infantry Tactic Action (target)

Any use or expansion of the database is encouraged. The full dataset can be downloaded below — add your own engagements, correct errors, or extend the period coverage.

Open the Predictor Download Dataset (CSV)

— Napoleon Bonaparte