Because we created the industry's first daily mortgage rate index based on actual lender rate sheets without any subjective distortions, and because the longest-standing mortgage rate index in the U.S. is a once-a-week survey with plenty subjective distortions and some quirky methodology, we often find ourselves pointing out what's "real" on many Thursday afternoons (the weekly survey comes out on Thursdays).  In virtually every case we can remember, there have been quantifiable reasons for periodic discrepancies.  Today may be the first (and certainly the most striking) example of Freddie Mac's weekly survey data simply not making any sense. Reason being: Freddie logged a DECREASE in rates this week.  Before proceeding, we should be clear what that means in the scope of Freddie's methodology.  A "week," in this case, refers to the 5 days starting each Thursday and ending each Wednesday.  As such, if today's index is lower than last Thursday's, it means that the average rate between September 19th and 25th was lower than the average rate between September 12th through 18th. Therein lies the problem.  Rates were quantifiably, clearly, and incontrovertibly higher--even if not significantly so.  Normally, when we apply Freddie's same methodology to our own daily rate tracking, we can at least reconcile any directional discrepancies.  We're not so worried about outright levels matching up because outright levels are not that important for mortgage rate indices (the CHANGE is important).