Strange Download Counting at Download.com

In May, my free Windows tool "Free RAR Extract Frog" has been ranked #4 at the weekly download.com "Most popular File Compression downloads" charts. This happened roughly one month ago. The program was right behind WinRAR, WinZip and 7-Zip. The weekly charts stated that within seven days more than 12.000 new users decided to download and install Free RAR Extract Frog - that's more than one new user every minute (Screenshot taken 2009-05-21): At the time, as the following screenshot shows, more than 450.000 users had downloaded the freeware program from download.com (also taken 2009-05-21):

The strange thing is: In the following weeks, Free RAR Extract Frog constantly ranked #4 in the weekly download charts, but the numbers did not match the total data displayed by the downloads counter as shown in the image above. As of today, this counter is at 542,980 (screenshot taken 2009-06-24):

542,980 - 453,263 equals 89,717. These 89,717 downloads were counted between 2009-05-21 and 2009-05-24, exactly 33 days, roughly four weeks. Since the weekly download charts stated each week, that the program had 12 thousand and some downloads, the overall download counter should not have gone far beyond 500.000 (12.000 * 4 = 48.000; 453,263 + 48.000 = 501,263). What happend here? Is the overall download counter not working properly? Are the download charts somewhat buggy or fake?

Let's take a closer look at the detailed weekly download charts at download.com named "Most popular in Utilities & Operating Systems" (screenshot taken 2009-06-20):
This is part of the top 20, showing "position this week" and "position last week", then "weeks on chart" and so on and so forth. What do the charts look like a few days later? The same part of the top 20 named "Most popular in Utilities & Operating Systems" (screenshot taken 2009-06-24):

Not much has changed. The overall download counters went up, the "last week" download counters changed a little bit, but not in a way that matches the overall counters. The "position last week" data doesn't reflect last week's positions. To see that there is really something terribly wrong with this list, take a look at the whole top 20 and count how many downloads have the same number in the field "position last week" (screenshot taken 2009-06-24):

Any more questions? Ask download.com!