By using this site, you agree to our Privacy Policy and our Terms of Use. Close
donathos on 14 January 2010
Edit Thread
Kasz216 said:
donathos said:


Right.  They're not exact numbers.  But exact numbers, for our purpose, don't exist--NPD, in this month and every other, releases some estimation, based on whatever data they've access to and the formulas they use to extrapolate that data.

If we're going to assign a "winner," or a first, second, and third place... it ultimately has to be based on whatever numbers NPD comes out with, at the end of their calculations.  Do the NPD estimations necessarily match the "reality" of the situation?  If NPD says that PS3 sold more than 360, does that mean it's necessarily so?  Of course not.  But we'll never find out if they're wrong, because nobody seems to be in a better position to tally the sales.

What we *do* know is this: by whatever process NPD uses, for better or worse, PS3 calculated to have a higher number of sales than 360.  And that *is* victory, in this context, because there's no more sophisticated or more precise calculation to appeal to.

Consider: in an election, votes are tallied.  Mistakes are made during these vote tallies, and the final vote counts aren't 100% accurate; there's a margin of error.  And yet, the candidate with the higher vote tally--even if by one solitary vote--is declared the "winner," even if its within whatever margin of error, because that's the only system available to count the votes.

NPD didn't though.  That's the thing... when a tracking firm gives it's numbers... that you take the margin of error is implicit.

Also the fact that real numbers don't exist is why respecting the margin of error is so important.

 

Also no... actual elections don't have a margin of error.  When mistakes are within the margin of error they have recounts and go over disqualfied ballots and they're contested leagally.

Regarding NPD, I'd be more inclined to agree if either they stated their margin of error with their results, as for instance political polls tend to do, or if they released their results as a range of numbers.

But my point is that "victory" here is nothing more than whatever number is higher, according to NPD.  Maybe "victory" ought to be "whatever console actually sold the most units"... but again, we don't have a better way of calculating that, do we?  What I'm contending is that having a higher number according to NPD is the sole criterion for "victory" in this context.  Is it a useless victory?  A pointless victory?  Maybe.  But if we were all well-adjusted people, would we be arguing about console sales in the first place? :)

And regarding elections, sure they do have a margin of error.  Yes, within some percentile there are recounts and legal contests, etc.... and all of those processes have their own margins of error, if you will.  What they won't wind up with, no matter how many recounts are done, is some exact tally of the *actual* votes cast--any attempt on our part to count those votes is going to be a flawed attempt.  But that doesn't stop us from pronouncing a  "winner" (as in, for instance, the 2000 US Presidential election), nor should it.