I’d actually like some clarification on this, too. I’m having a similar issue, and don’t quite follow whether this is intentional or not. Initially, I had myself convinced that the parses were re-evaluated at a later time, which resulted in rank decay. That said, I’m seeing for other people that this does not seem to happen – and my last example showed a discrepancy within 10 minutes of the Logs being posted.
I have a thread up about this for myself, too. Still very recent, so I’m not expecting any immediate replies. But, hopefully, we’ll get some answers.
Following up on Kihra’s comment, when viewing data that focuses on % rankings (overall or iLvl) there are generally two things you can change that influence what you’re looking at if the default numbers aren’t working for you:
Today vs Historical - If you’re looking at older data and want to see how you did compared to numbers at that point (especially useful if your class has experienced a change - buff/nerf), use Historical.
Rankings vs Parses - Rankings compares to the best ever parse of other eligible players, Parses compares to the last two weeks of parses for other eligible players. I feel that Rankings is the more “fair” comparison.
The different pages still don’t perfectly match so there are other things going on there but I haven’t been able to figure out those discrepancies yet and they’re usually minor.