I don't think anyone at all is debating whether or not D&D has been the historical best seller RPG over the last decade. The only thing being debated is whether Pathfinder has recently (in the last few quarters) caught up to and perhaps surpassed D&D in sales. For that we need to look at current data and not rely on historical data.
Look, I'll give you a real world example from my work. I work with oceanographic and meteorologic data to compute drifts of objects at sea (it's a search and rescue application). We consider both long term background climatic data (like your decade of facts on D&D's sales) but more useful than that is real time data. Knowing that on average over the past 10 years current flowed two knots north in this particular area is a good fallback, but if we have ship observations that say otherwise, or we have model data from an oceanographic model we will use that instead because it is closer what is really going on at this particular moment in time. The best data we can get is if we send an aircraft or vessel out to deploy a buoy that transmits position data (from which can derive surface current data). If we were to fall back to the climatic data if we had closer to real time data we would be doing a disservice to the public and be putting lives at risk. In some cases we are forced to fall back to climatic data if we are missing any other data, but even then we usually modify that data based on modeled or observed winds for which we have really good coverage (the wind drag effect on water will produce a noticeable surface current).
So what's the point? The point is that by rejecting this near real time data (Amazon, ICv2, etc) you are simply falling back to the historical norm which does not necessarily represent what is really going on on the ground at this particular time. You're falling back to an even worse quality of data than the data that you're criticizing.
Okay, this month we have both Pathfinder and 4E releases (Ultimate Combat and the Neverwinter campaign setting. If you look at the Amazon.com rankings, Pathfinder has an overall book rank somewhere close to 500 (it's 665 today - it was under 500 a few days ago). If you look at the Neverwinter book, it has an overall book rating of 2000+. A few days ago I believe it was somewhere between 1700 and 1800.
The rankings have been pretty consistent over a period of days. So we can clearly say that as far as the Amazon channel goes, Pathfinder has been outselling the most recent 4E book over the past week or so. That is real data. Can we extrapolate that to mean that Pathfinder is more popular than D&D? Probably not. However we can probably expect that sales in other book channels are probably similar especially if you work in other data like the recent ICv2 data over the past few quarters.
It seems to me that what you're doing is simply rejecting data that doesn't match your view of how things should be because it doesn't match the long term trend line. You are favoring longer term data over more recent short term data which is a bad way to model what is going on right now.
, and I used the position buoys to locate a vessel, I can be reasonably certain a simple search in that area would yield a vessel. If it did not I could conclude something bad happened.
