Thread: 1001Skins
View Single Post
Old 9th May 2017, 01:51   #118
garetjax
Candyass
(Major Dude)
 
garetjax's Avatar
 
Join Date: Apr 2001
Location: Benton, AR
Posts: 2,321
Send a message via ICQ to garetjax Send a message via AIM to garetjax Send a message via MSN to garetjax Send a message via Yahoo to garetjax
The download counts where flawed. A very high percentage of them were initiated downloads, without a completion event.

This mean that the download counters across most of the websites that do downloads counts were likely a problem too. While a couple of years ago I switched to using a handler to initiate downloads and then after final byte, did an tick for the download the damage was already done.

Example: SlickSeries2.wsz | 8.6 Million Downloads | 823.6K Actual = 10.4% actually downloaded it. I spent some time looking at everyone overall counters for downloads and they all have the same fundamental flaw. A counter, based on the click to the download; some are better at prevention and my method was fairly fool-proof for accurately setting the download counter; however, I said...it was after the damage was done.

So, we are now longer using this method; we use a combination of things (not disclosing to prevent gaming the system again) and we do not show the # so as to prevent a need to game the system or give a user a measure of their count vs other skins.




Our rating system has also been recalibrated and uses multiple systems to prevent similar outcomes.

1001skins |
That's not a skin, it's some god awful piece of skinner gunk. - Mr. Jones
garetjax is offline   Reply With Quote