Quote Originally Posted by lars573
Having seen a reenactment of a viking shield wall, I can say that it is similar to a phalanx. But different in 1 important aspect, size. A shield wall was between 1 and 3 ranks deep, hence the reason it's called a wall. Where as a phalanx was a square block of a few hundred men who weren't quite as close together as men in a shield wall would be.
I'll make an attempt at a small clarification, elements of the hoplite phalanx and Macedonian phalanx were square (or even thin rectangles) but the elements were united into a formation commonly employed as a lon line, with depth. (Extreme depth for the Macedonian "double depth" units.)

One unfortunate thing about computer limitations is that while we can see ranks in the game at only slightly reduced scale, we cannot really appreciate the width of big armies (deployed with perhaps 2 miles of frontage vs. several hundred yards in RTW.) Consider that the game scale is about 10:1. So we can have the right number of ranks and lines (or nearly so) but the overall deployment is an order of magnitude too narrow. When you get full stack armies you can start to approximate a legion. But consular armies were often 2 Roman legions, plus two allied legions. Bigger ones might be 4 Roman legions, with as many allied legions.

I think we are hitting the wall with respect to AI's ability to cope. Even when computers and graphics cards can handle full size armies, the AI is unlikely to be sophisticated enough to move so many units with any level of competency. The complexity of everything else increases, but the complexity of the AI stays the same...and to top it off, it has a lot more to manage in a coherent fashion. The AI might be able to move single units properly, but it will always be more difficult to get it to wield an army. That requires much higher level decisions, and programmers are unlikely to have the time/money to build such an AI.