[LINK] BlueCava tracks individual devices, assigns creditworthiness

[rene]

On Thu, 02 Dec 2010 12:23:22 +1100, Jan Whitaker wrote:

> At 12:10 PM 2/12/2010, Steven Clark wrote:
>>> BTW, the figure of the plugin data changed from one check to
>>> another and I didn't change anything in the browser.
>>>
>>
>> *you* didn't ... which is part of the point :D
>>
> I figured it was the dynamic nature of the dataset they are comparing
> to. Still, my fingerprint is the same. So the relative level may or
> may not be 'telling' me anything. The stats of 1 in 640,000 like mine
> isn't useful unless you know the total number of cases. Or am I
> thinking about this wrong?
>
> Now it's down to 426,000. I don't get it. The changes are in the
> browser plug-in section.
>
> My head hurts. Can someone explain this to me in English?

My head hurts too, but imo it's the test that's arguably flaky/not very 
helpful. I ran it three times and first time it said "Within our dataset of 
several million visitors, only one in 640,### browsers have the same 
fingerprint as yours." (can't remember the ### numbers). Second time it 
said, 426,###, third time it said 213,418. 

Hence it appears to me that the more any particular user runs the test, the 
less unique it will say their browser is (because they have no idea how 
many users have run the test multiple times). That combined with the 
failure to state exactly how many browsers they reckon have run the test, 
at any given minute, imo makes the thing pretty useless.

Irene