The limitations of EB

[johnhughthom]

Allegedly the AI is much more passive (but also smarter about building armies) with Alexander's executable, maybe I need to try that?

It may be slightly better, but nowhere near enough to stop the game feeling like a chore after a while.

[vollorix]

Imho, the army compositions of the AI depends on the balancing of the unit size/cost factor, and on the total amount of money it gets.
For example: i´ve halfed the size and the upkeep of Triarii, but kept the recruitment costs, and see - there are no more then 2 of this units per large Roman army most of the time. Or cut the size of Pedites EXtr. to 30 men per unit ( 60 on large ), kept the price for recruitment + upkeep, and now you see far more Samnites Heavies on the field, which i didn´t touched. But then, once Roman AI becomes quite rich, it starts to crank out Pedites again, so if one would reduce their recruitment pool to Roma only, and perhaps tweak the recruitment costs a little bit more, AI won´t get so obsessed in hiring them. Farther: once i´ve done those changes, the AI started mass recruiting Rorari, so i lowered the costs for Hastati by 100 and Princeps by 50, iirc, and now i can see very nice historical compositions for the Romans: up to 6 Princeps, 4 Hastati, 2 Triari and some Pedites + couple of Samnites, with all the other support troops. But this i could only achieve by taking the abilities of the Romans to recruit any Celtic or Greek troops at all.

Therefore my conclusion would be: if one balances the costs of the units to force specific recruitment behavior rather than simply following kind of formula for the unit stats, one can get quite satisfying results.

Edit: It should be Rorari instead of Leves.

[Titus Marcellus Scato]

@vollorix: Very interesting post - I think I'd like to make your modifications to my own EB installation, it sounds fantastic!

Would you care to post your edu file?