Odd Unexpected Test Results Of Additional Functions

googlealchemist

For all of these tests I used chunks of 10k urls...different for each test but of course using the same exact 10k list for the side by side comparisons per test.

I ran a few tests using just the basic SER engines checked compared to also checking the relevant SER extended engines. 

The latter gave me an additional 2 or 3 extra % points each time which is worth it to me IF those extra urls turn out to be valid vs bunk...which I'll be testing next before moving forward with my main sorting. I'll update here.

That didnt come as a surprise though...what did surprise me was when I tested the other options as noted below and got the opposite results of what I expected.

--------

Attempt to recognize engine from file name

With just the straight SER engines checked: 793 checked vs 832 unchecked

With the SER relevant extended engines checked: 945 checked vs 994 unchecked

--------

Deep

With just the straight SER engines checked: 819 checked deep vs 824 unchecked deep

With the SER relevant extended engines checked: 1001 checked vs 1006 unchecked

--------

Wildcard

With just the straight SER engines checked: 851 checked deep vs 852 unchecked deep

With the SER relevant extended engines checked: both options gave the same 1044

--------

So it was pretty close all around but as you can see when I tested checking each of those functions individually I got LESS results WITH them checked vs without.

Just wondering what others experience is and if they can make sense of this or if I'm just missing something stupid as the point of extra checks was to get MORE urls vs less?

s4nt0s

Hmm so you definitely got better detection using the relevant SER extended engines, but the wildcard/deep didn't seem to make an improvement. I know the deep/wildcard should have a very minimal effect on detection, but I'm surprised to see its slightly worse with those options enabled.