Calculation of the area of the fin and rudder and its effectiveness as a stabilser is notoriously difficult, even with todays' advanced numeric methods. The difficult arises primarily because of the difficulty of predicting boundary layer behaviour. The boundary layer is already thick by the time the airflow reaches the tail area, and its thickness varies according to angle of attack. The wake flow over the fin area is therefore relatively sluggish, and is dependent to varying degrees on the downwash from the wing (in general low wings result in more efficient fins than high winged aeroplanes). Another variable that is difficult to predict in the case of prop driven aircraft is the helical wash from the prop.
These influences are extremely difficult to calculate when trying to predict flight behaviour across a wide flight regime and even more so if the aircraft is required to be spin-certified. The real behaviour of the aircraft can only be determined in flight testing, when deficiencies become apparent.
Designers always tend to err on the small side when designing fin area as generally a smaller fin will be lighter and cause less drag. Most of the time, the intelligent guesswork is good enough, but as you note, sometimes it is not, resulting in the various retrofitted strakes that you see on some designs.