I've done some reflecting, and yes, I think 17 starting baddies is certainly not possible, but 15 starting baddies is possible - if whoever the sorc is gets really lucky with scanning. Here's how it would work:
Assuming a 3 pack setup, a typical pack would probably have have the following:
3 wolves
1 cultist
That makes 3 packs of 12. A sorcerer would make 13, and a sorcs apprentice or an unattached cultist would make 14 or even 15 starting baddies. for a 36 player game
Now, If the sorc gets really lucky and picks up an apprentice or even 2 apprentices, that makes 14/15 baddies to start the game... 2 extras picked up over the course of the game would make that 16/17. Suddenly that ridiculously high number makes a bit more sense, setupwise - 14 (14 out of 36 equals 38%) or 15 starting baddies (15 out of 36 equals 42%). Bad luck (wolves only hitting each other for 1 hunt, an apprentice or 2 falling to the sorcerer, the cloaking incident losing the village a turn and two villagers) makes the rest possible.
That's the only way the numbers make sense. Or else wagon is either wrong about the numbers (I suppose that's a possibility, he did get the initial parity count wrong) or lying, but I don't see him having incentive to lie about that.
(I have no idea what a 4 pack setup would look like, but I suppose a 4 pack game is theoretically possible.)