I did some experimentations with everyone's favorite domain, mydomain.com. I was hoping that the site would have less than 1000 pages so I could test the accuracy of the numbers we all use to calculate the SI index, and it does. What I found was that the numbers are unreliable at best.

I ran the first query to get the total number of pages in google's index for the mydomain.com domain (site:mydomain.com) and got 894 results. I then did a query to get the number of pages in the main index (site:mydomain.com -fake:mydomain.com) and got 152 results. This, with some math, indicates that mydomain.com has an SI ratio of 83%.

Now, I checked those numbers for accuracy by seeing how many pages are actually indexed by Google. I ran my main query again, and this time used a hack to view the last "actual" indexed page (added &filter=0&start=893 to the URL) and found that Google only has 621 pages in the total index. This means that mydomain.com's actual SI ratio is actually 75%.

I did some additional testing, and have determined that the supplemental index not be the only index that is blocked by the -allinurl: parameter. I did a few searches that my site comes up for in Google's blog search, and checked the pages that were returned. I then entered these URLs into the filtered normal search and they were not returned as being in the main index. This could indicate that the blog index is seperate from the main index. I tested this further by searching for some of Matt Cutts' blog entries, and they showed up as not being in the main index either. Anyone interested in helping me test this further?

__________________
The best way to learn anything, is to question everything.