# Predicting Succession Outside Dynasty

I am playing as a one province count in India. I have murdered the heirs of Rashtrakuta until the grandchildren of the king are the only males left in the dynasty tree other than the king himself. Now that all the traditional heirs are dead, the line of succession is-

The duke currently second in the realm tree, with 19% of the king's forces, Steward, 8 provinces The duke fifth in the realm tree with 15%, Spymaster, 10 provinces The duke fourth in the tree with 16%, 11 year old, 9 provinces

And so on like this. Is there a preference for councilors? What drives succession after the direct heirs are dead?

• The realm tree doesn't show the line of succession. If the succession law is primogeniture or gavelkind then the eldest son of the eldest son of the king is the first in line to inherit. If it's ulltimogeniture then its more complicated, one of the king's collateral relatives should inherit. To see the line of succession hover over the kingdom's shield.
– user86571
Commented Aug 18, 2017 at 3:10
• Thanks Ross, I understand that, the line of succession tooltip is the order in which I've murdered all 23 male heirs of his dynasty. The succession is agnatic gavelkind. It appears that there are no subordinate titles that cause the realm to split significantly. My problem here is that his only collateral relatives outside his dynasty are my grandchildren. So is there any pattern outside of the standard succession that it is using as a backup? Commented Aug 19, 2017 at 3:59
• If there are no valid gavelkind successors the game falls back primogeniture succession which should mean one of the king's grandchildren inherits. If there's no primogeniture heirs then the king's liege inherits. If the king is independent then king's highest ranking vassal inherits (ties broken by order of the vassal's title in game's data files) and failing that one of the king's courtiers inherits (ordered by the sum of age and prestige), and failing that a randomly generated character inherits.
– user86571
Commented Aug 19, 2017 at 4:21

First of all, there are some restrictions about who can inherit at all. Anyone excluded by the following points can be ignored:

• "Agnatic" means that this title can only be inherited by males. So any females can be ignored for the succession of this title (agnatic-cognatic means females only matter when she has no brothers, cognatic means gender equality. There is theoretically also enatic and enatic-cognatic succession which reverse these gender roles, but these are only available through modding).
• When the kingdom got a crown authority of "High" or "Absolute", anyone outside of the realm is not eligible either and can also be ignored (unless we are talking about the title of the realm itself).
• Some religions do not allow theocratic leaders to inherit (those whose capital is a temple holding). Note that Court Chaplains do not count as theocratic leaders, even though their portraits look like they do.

Now that we determined who is eligible at all, we need to perform the following algorithm to find the heir:

1. Gravelkind means that the inheritance is divided among all direct children, with the eldest child getting the largest share.
2. When there are no living direct children, then Gravelkind succession behaves like Primogeniture. That means the title performs a depth-first search among all descendants, preferring the eldest, until it finds someone alive.
3. When there are no living descendants at all, then the title travels up the family tree to the father.
4. When the father is dead, the downward search repeats from him: It looks for the next living descendant of the father. When it finds nobody, it goes to the grandfather and so on.
5. When the whole family is deceased, the liege inherits.
6. When the title-holder is independent, one of their courtiers inherits
7. When there is no eligible courtier, a random character is generated to gain the title
• Good answer, but I think the asker is most curious about predicting which courtier will inherit if you get to step 6. Commented Oct 28, 2017 at 21:34