I do not remember the reasoning behind every stat for every creature at this point in time, but I think a lot of that is due to my being too much of an expert player and thus having some very OP powerbuilds in mind. The player is starting to approach epic levels at that point and a powerbuilt Weapon Master or Arcane Archer (for example) can get ridiculously high ABs at high levels. Another factor is that at the beginning of Ch. 3 you do have a warrior type (Quord) in the party whose AB can be made quite decent through things like Zarala's Bard Song and Bless/Aid/Bull's Strength potions, even if the PCs AB happens to be on the low side. Finally, at the time I was doing a lot of the design work for Ch. 3, Swordflight was even less well known or popular than it is now, and it appeared as if it would be a very small (rather than just small) niche thing only of interest to really dedicated power-gaming combatfags, so I was kind of assuming that anyone not in that category would have dropped the series long before reaching Ch. 3.
The Marilith was something of a special case precisely because she can potentially briefly join the party, and I wanted her to be an actual asset if she did so, not die in a few seconds. The latter is what she would do against the kind of enemies one runs into in Swordflight unless I gave her some seriously scary stats. If she were not unusually formidable as an enemy she would have been basically useless as an ally. Theoretically, I suppose I could have given the "henchman" version different stats but I did not want to do something so gamey and unrealistic.
