Some general thoughts on long ancestral lines. One is the old saying that "a chain is no stronger than its weakest link" - meaning that one undocumented relationship will break the entire line.
Secondly, errors are cumulative, in the sense that the longer a line gets, the greater the possibility of eg. one unfaithful wife. In twenty generations' time, that probability may approach 1.
Third: even in a perfectly documented line, how many genes do you share with an ancestor 20 generations back? Answer: about 1 millionth.
And last but not least: You can't always take medieval sources at face value. The author may have had his own axe to grind, eg. forging a relationship in order to get hold of some land. This is really an area for experts.
Personally, I feel that deep genealogy before year 1600 mostly is futile. I find it far more rewarding to make sense of original documents from the 17th and 18th centuries. Going back 400 years or 12 generations, you've got potentially 4,096 ancestors. Good luck in trying to find them all.