First we have the group of experimental compressors, designed with only one thing in mind, getting the best possible compression regardless the time it takes and the amount of memory it uses. ... The best programs in this group, i.e. the ones getting the best possible compression, are WinRK (using PWCM mode), PAQ8, PAQAR, Durilca and SLIM. At the moment WinRK (PWCM) and PAQ8 are the two best compressors of the five, but they are also much slower then Durilca and SLIM (to give you an idea; PAQAR in -8 mode does about 5-7 KB/s on a slightly overclocked AMD Barton 2800+). ...
Q. Why don't you add compression times?
A. As the sites' name already indicates, I'm looking for the maximum compression ratio. The time it takes is irrelevant (to me), getting the best switch combinations takes by far the most time... I agree time is useful to know if you are considering an 'every day use' compressor, but that's not what the site is about.
Apart from the fact it takes a lot of time to do all measurements there is a bigger problem. I try to find the best combination of switches to yield maximum compression (and this really can be an exotic combination for programs like 7-Zip, Durilca etc). That particular combination can lead to say 40.00% compression with a compression speed of say 20 Kb/s. A different set of switches can have 39.99% compression with a speed of 500 Kb/s. Is it fair to list this 20 Kb/s as, 99.9% of the people will not use the switches I used?. I think speed comparisions are only fair if you let the program use it's default settings. The multiple file compression test (MFC) does include (de)compression times are more 'regular' switches / options are used there.