DKII said: I wish all the people who are always saying "I don't really trust VGChartz sales I'm waiting for " would come back an apologize every time they're proven wrong. |
When is that? They've been comparing to NPD for what, 3 months now, and two of those months they've been terribly wrong (on hardware).
And lets not even talk about software.
Last month I guess, they were a bit better, but when you're off by 20% and stuff, it makes the numbers kind of useless. Especially since anybody who follows this closely could make a simple guess at NPD's with similar accuracy or better (the last few months things have been pretty stable). I mean I could guess 200k 360 and 100k PS3 most of the time and be about as accurate as VGchartz.
I consider this NPD will be a good test of VGchart as we will get a good deal of flux due to PS3 cut. That is why I'd like to see if the July NPD's back vgchartz up on ps3 sales increases before proceeding.
I defend vgchartz in general, and their numbers aren't worthless they're a good secondary data point, about as good say as amazon's rankings. Just something to get some kind of idea. But they need to improve accuracy and develop a track record a lot before I trust them more...
Kind of pointless to bury it in this post, but it also might be nice if they, stopped combining US and Canada sales, and dropped the attempt at predicting 5-10% over. When you are already 10 or 20% wrong, these just compound it terribly. And most people just want USA NPD's and dont include Canada, so it makes vg look less accurate than you actually are. I realize that's technically the viewers fault, but I just think an attempt should be made to predict USA NPD's only, without the overcount, and it would be a lot more accurate. Perhaps just do a seperate specific USA only NPD prediction (seperate from all the other vgchartz archives and stuff, just a side project). Keep the 5-10% over for VGCHARTZ official archive database and all, just make a seperate NPD prediction in ADDITION to that.