I agree - I also have (big) doubts about the value of this test.
Actually, this test was done by Virus Bulletin - but its purpose was not to evaluate the speed of the scanners, but rather to test for false alarms; the table of times and speed was just an "additional information". So, you should be careful when interpretting these results.
- the article doesn't mention the scan settings - the scanner speed will differ
significantly for various scanner settings (for avast!, you may set if whole files, or just important parts will be scanned, if archives will be detected and unpacked, etc.)
- it's absolutely pointless to consider a scan time of "4 seconds" having any sense - if you scan just a few files, the main part of the time will be some overhead (scanner initialization, etc.), the scanner speed won't really affect the result. Besides, measuring the speed on whole seconds (in the range 1-10) and stating the resulting speed to 5-6 significant digits is... erm, strange
- this particular test is from VB June 2005. If you take a look at the similar tables in VB October 2005, you'll notice the e.g. avast! scanner speed when scanning ZIPped executables is
bigger than when scanning the (unpacked) executables themselves. That's also rather suspicious
, and brings more questions about the test methodology...