Bo here -- author of letsrun's program. There are two main reasons for discrepancies (unless I have a bug in the program):
(1) there is a mistake somewhere in the season results
(2) different interpretations of the common opponent tiebreaking rule
For (1), hopefully the letsrun sleuths have rechecked these enough that they're accurate. However, with A and B teams it's always a lot of work and you have to be really careful.
For (2), it's important to realize that the common-opponents rule is not exactly defined in the rulebook. It just says that the committee should break ties using record against common opponents (but how?). So Flotrack may be using a different interpretation than us. This almost never ends up mattering, but this year it does. This program uses the interpretation "take the group of opponents that both teams have faced, and pick whichever has the most wins against that group". Maybe the committee and/or Flotrack use a different interpretation -- I don't know (and would love to find out).
So the bottom line that I can see is (a) you can't be sure till the committee makes an announcement, and (b) if you want the best guess possible, you have to look through the season results and make a judgement call about the common opponents tiebreaker.