I know, I was just trying to connect the dots between both of your experiences and figure out why villagers would be going to a castle in the first place, as was happening in the case of OP.
In any case, mystery solved I think. When a villager party takes too long to reach it's destination, the game will sometimes create a new party under certain circumstances. Once the old villager party finally reaches its destination it gets the signal that it's been replaced and obeys the command to disband. It does this by travelling to a nearby allied fief and merging with the garrison, similar to a player's companion party, while the new villager party heads back to its associated village after delivering goods.
The slow walking speed of 200+ villager parties probably causes them to get "lost" and replaced more often than their speedier brethren. It's also way more noticeable when it happens with a large party than a small one. Here is an example of what's happening:
I stood in front of the gates of enemy town Epicrotea, blocking the path of various hostile parties trying to enter until I noticed two villager parties from the same village show up. I then moved and let them enter. Upon exiting from Epicrotea, one party headed to Mecalovea Castle to disband and one headed back home.
Heading Home |
Disbanding |
And here are the garrisons of Mecalovea Castle before and after the villager party enters (Ferth is of Sturgian culture):
Garrison Before |
Garrison After |
Maybe the game should check to see if the garrison is at capacity before merging the peasants with it so desertion doesn't become a problem. As you pointed out though, the hearth growth rate has been lowered, so there will probably be fewer gigantic villager parties roaming the map in any new campaigns from this point on, making this event somewhat more rare.
As a side note, I checked what happens to a bound villager party mid-journey when their fief changes hands. They switch sides and pick a new destination without missing a beat.