The statement that tiny and small are unchanged is a specific exception to the optional general rule that large start at BS + 1
OoC:
The alternative method in Draft 2.25 (in which the bloodline score of a regent is stability-relevant) seems to indicate, conceptually, that, when a domain's province/holding levels are greater than the bloodline score of its regent, the domain is unstable (i.e., large or huge, depending on the extent); furthermore, that, below bloodline-score-plus-one province/holding levels, a domain is tiny at 0-20 levels, small at 21-40 levels, and medium at 41+ levels.
Not quite true.
The method described in draft 2.25 state that no domains with 40 or less holdings get a penalty regardless of its regents BS.
So, conceptually, it seems to indicate that a domain
over a certain size(40 levels) is unstable if it has more holding/province levels than its regents BS.
But I question the concept of linking BS to stability. It in effect make BS a stronger stat.
What, from a balancing point of view, does it bring to the game to make one stat stronger?
Answer: It makes it easier to game the system since you then have a more limited field of effort, or a new opportunity for improving your realms overall power level.
The only reason that the alternative rules in 2.25 could be acceptable, from a balancing point of view, is in fact that no realms with 40 or less holdings get affected by it. Thus, hopefully, making stability gain from investment in BS a long term plan that might not dominate the game.
The best solution from a balancing point of view would still be to ignore the option and go with static domain sizes.