This bias of engagement at strategic points also showed up in early play-testing feedback and prompted the designers to reassess a lot of the decisions around how the single-player campaign would work and how its missions were laid out.īut don't take my word for it-Creative Assembly's Pawel Wojs explains things better than I do and also throws in some great details about how the game's design process worked. Along the way, they would also find areas where their procedural terrain generation hadn't gotten things quite right (like requiring a campaign battle to awkwardly play out on a near-vertical slope).īut the biggest issue was that gameplay mirrored real life in a very interesting way: rather than having battles ranging across the vast and varied landscape as the designers had intended, battles tended to occur near cities or at strategic chokepoints. In the evenings when it was time to go home, Creative Assembly would set as many PCs as they could to playing the game in AI-only mode, iterating through battles and scenarios in order to help see which units needed balancing and which scenarios needed tweaking. In order to properly test a game with thousands of square miles of playable space, the designers employed automated tools running on their office PCs.

Game total war rome 2