Media Create problems or inaccuracies For the beginning of this data I use a text file I saved from GAF at some point, though I didn't save who or where it originally came from. The text file mostly focuses on GCN, Xbox, and PS2. It has GBA, too, but not as much. GCN, Xbox, and PS2 numbers start at the week of February 18, 2002. GBA numbers start on the week of January 26, 2003, shortly before the GBASP launch. All numbers end the week of August 25, 2003. Text file says 8/19/2002 PS2 number is 65,300. GameFront begins on the week after, but their "previous week" number is 63,500. Since I don't have any YTD values before that, I can't tell which is correct. If the first three weeks of the text's 2003 chart are A B C, GameFront lists them as B C A. I'm going with GameFront here. However, from here on out on some the GameFront version is more rounded than the text's version; in those cases I'll use the text's numbers. GameFront's in general has a lot of problems... missing weeks, totals that don't change from the previous week, sometimes totals sitting in the wrong row. Where easy to tell that something has been switched I've made the switch back, or in cases where there's a missing week I can look at the "previous week" column from the next week, but there are still a lot of gaps. I use GameFront as a source from the week of August 19, 2002 through the week of September 8, 2003, after which I can start using GameScience instead. Occasionally GameScience has a gap that I fill in with GameFront data. In July 2004 GameScience has a sizeable gap. I filled it in with GameFront numbers. However, I see that when GameScience picks up on the week of August 9, 2004, their YTD numbers no longer match. Compared to GameFront, GameScience's GCN is +3, GBA is -20, GBASP is -338, PS1 is the same, PS2 is +65, Xbox is -435, and WonderSwan Crystal is the same. As the year goes on they become even more different, but I haven't gone through to see where all the discrepancies come from. No formulas are used on the early part of this chart. Due to the rounded nature of the numbers, adding up the rounded weekly values wouldn't always match things up with the rounded YTD totals, so I left them as-is for anyone else to do with as they pleased. In 2004, for a while the weekly numbers go back and forth from rounded to unrounded. If an unrounded week appears after a rounded week, though, the new YTD number is the previous week's rounded YTD added to the new unrounded weekly number. So don't take any 2005 YTDs as exactly accurate. From the week of October 11, 2004 to the week of October 18, 2004, the PS2 YTD actually decreased (GameScience). If it's an error, it wasn't off by an even amount. It was used as the basis for further YTDs in later weeks, so perhaps it was a Media Create correction? Also from the week of October 11, 2004 to the week of October 18, 2004, the PS1 YTD doesn't change at all (GameScience). The YTD then goes on as usual, so I assume for the rest of the year they're slightly too low. For my purposes here, though, I'm leaving them as-is.