Can you help me win this argument?---->Where can the most accurate sales...

[nordlead]

well, first off read the methodology page http://vgchartz.com/methodology.php

Here are some of my thoughts (note, I have a minor in math, but I'm no expert on statistics).

VGChartz, just like NPD, samples a portion of the market for sales information. Since they can't poll 100% of the market, they use statistics to "guess" how many were actually sold based on what their partners reported. Now I saw guess in quotes, because all VGChartz and NPD do is make educated guesses.

NPD has an advantage over VGChartz in that their sample size is considerable larger. This provides NPD with more safeguards that prevent them from having major mistakes in their results, however they are still possible.

While VGChartz has a smaller sample size, anyone with knowledge of statistics will tell you, that a small sample size with an excellent formula can be better than a large sample size with a poor formula. The downside of this smaller sample size is when an anomaly comes through their data, it is harder to figure things out, so their results can be skewed one way or the other easier. This is where you need good judgment on the part of the person compiling the data to determine if the final results look realistic.

One other difference, is that VGChartz doesn't fudge their new numbers if they get new information. Instead they go back and correct the week that was wrong. NPD may also provides corrections, but they do not provide any to the public.