Quote Originally Posted by Marshall Louis-Nicolas Davout View Post
I don't see why you are telling people to provide even more sources,...
Greg is not asking for "more sources"; he's asking for more substance. The first post in a thread sets the tone for the subsequent discussion. If the original poster does not put much effort into it, it tends not to generate high quality responses. You can see a good example of that by comparing the two recent abortion threads in the backroom.

I've seen on other forum websites from where they ask just one or two simple questions.
Well, without meaning to sound arrogant, the Org aims to be more than another "forum website": our niche is to encourage more substantive discussions; we were recently praised for that by CA. We try to do better than the "yes it is" "oh no it isn't" "oh yes it is" back and forth pantomime that passes for discussion in some other forums.

To answer the question...
My take on the question is that the ETW campaign AI could not cope with the incredible scope and freedom of the inter-continental map (except for Maratha, they rock!). By contrast, STW2 narrows that down dramatically with mountain walls creating choke points so that provinces become almost Risk style ones with limited lines of approach. This gives the AI more of a chance to perform (it is possible to program very good AI in games where movement options are limited, eg chess, but hard to do so for more free form games such as TW battles).

The ETW AI also suffered from losing the tech race (fire by rank and 2nd rates anhilating all in their path), which seems less crucial in STW2.

I am not sure about the battle AI: STW2s AI is certainly competent and seems to be put on a better fight. I found ETWs musket based AI a little passive whereas NTW seemed to go to the other extreme, hurling itself onto your grapeshot firing cannon (or rather marching forward determinedly then suddenly becoming passive when it reached grapeshot range). I suspect TW is better for programming melee based AI than firepower based AI. (The native Americans in ETW could be extremely effective early on, when they just form an enormous battleline and walk over nascent European armies.)

STW2 still suffers from a few AI flaws that carry forward from ETW and NTW: an inability to recognise the importance of trade nodes and lamentable performance in siege defenses (it will mill around the centre getting shot to death or fixate on initial feint moves ignoring subsequent real enemy approaches elsewhere).