Finding the important factors in battle outcomes: a statistical exploration of data from major battles

• Main Author: Yigit, Faruk
• Other Authors: Lucas, Thomas W.
• Language: en_US
• Published: Monterey, California. Naval Postgraduate School 2012
• Online Access: http://handle.dtic.mil/100.2/ADA386719; http://hdl.handle.net/10945/9233

This study explores important factors in battle outcomes through a statistical analysis of data from major historical battles. The data set of CDB90FT has been made available and documented by the Center for Army Analysis (CAA). The quality of the historical data is good. There are 660 battles listed in the data set containing over 140 numerical features for each battle. The earliest battle in the data Set is the Netherlands' War of Independence in 1600, while the last one is from the Israel- Lebanon War in 1982. The data set contains many interesting facts on the battles including the initial strengths, the total strengths, the number of casualties, the lengths of the front lines, terrain features, command capability of leaders, weather conditions, etc. The approach is to use the data set as the basis for an objective and scientific comprehensive analysis, seeking patterns, trends, and relationships in combat. After making campaign-wise grouping and analysis, it is found that the Force Ratio is a valid estimator of the battle outcome. In addition, the Casualty Rate has declined steadily over the past four centuries while Dispersion has increased.