OK, I'm a bit confused, been a bad day at the office and my brain is fried. Bromley loaded RB's earlier saves and found that loading had a detrimental effect on the AI when he worked from the earliest save using the standard passive player test technique. RB says that's what he expected, because the early saves are close enough to the start of the campaign that it should still be wonky after a load. If both of you load up the *latest* save and go forward from there with player passive tests like Bromley did, what happens? I guess I still haven't seen (or been bright enough to recognize) evidence for or against RB's theory that the AI gets better in the midgame, away from the starting extrema. I think it's a good theory, and it would explain both what we've tested so far and what CA said to us, so it's attractive emotionally (we can all be right!), but I want to know for sure one way or another.