
"The great battles of Alexander, Hannibal and Caesar arouse endless interest, but our understanding of them is marred by the weakness of the surviving source material from so long ago." "Lost Battles employs simple simulation techniques to make traditional battle diagrams 'come to life'. Readers will actually be able to refight engagements for themselves, gaining greater insights into whether proposed tactics and deployments make military sense, and experimenting with alternative possibilities such as the presence or absence of the Persian cavalry at Marathon." "This book is the culmination of ideas and models developed and refined by Sabin in the course of over 15 years. It is supplemented by free downloads as shown on the colour plates and described in Appendix 2, allowing readers to refight the battles using anything from paper and pencil to dynamic portrayals on their computer screen." "Lost Battles is a toolkit for further study, offering endless thought-provoking insights into the enigmas of ancient warfare."--BOOK JACKET.
https://www.amazon.com/Lost-Battles-Reconstructing-Clashes-A...
The author developed and then iterated a simple simulation (structured like war game rules) for modeling ancient battles. Its been a while since I read the book but IIRC his goal was to create a model that was able to reproduce what actually happened in relatively well-documented battles, and then use that model to explore less well documented battles.
It's not a good cover-to-cover read, but once you understand his overall scheme, its fun to pick up and read his model for a battle or two.