Status: In Progress – Updated with each Google toolbar PageRank update.
Time Spent: 8 hours so far.
Money Spent: $13.09 for 11 .info domains.
Update: The sites got their pagerank assigned before the pagerank update. I think this is because a few blogs linked to the feeder siter and that made Google put focus on them. However I just came to know that nofollow links make sites lose as much linking power as if they were normal links. This is making the backlinks to this page mess up the PR calculations completely. I deleted the back links from the node sites so they should now get higher numbers at the next toolbar pagerank update. The results so far: feeder sites all pagerank 2. All nodes except node A got pagerank 1. Node A pagerank 0.
I have set up a simple network according to the example network found at Wikipedia’s page on PageRank. PageRank is a number that may or may not be very important for search engine optimization but it’s still fun to conduct an experiment with it. Each site is a .info domain and in total there are 11 domains,each setup with a basic WordPress install. I even made the header colors of the kubrik template match the colors of the sites in the image below. Any outgoing links on these sites will be removed, set as nofollow, or, in case I want to use the sites linking power identical on all sites.
As shown above, The sites use the same structure as the classic PageRank example at Wikipedia. There are no links from outside the network to any sites other than the feeder sites. I am aware that the strength of the network should be regulated by the number of feeder sites used, not the linking power to the feeder sites, but I chose to do it this way in order to be practical. The links to the feeder sites are made as identical as possible, with each page linking to them linking to them all, in a random order. I never thought I would ask anybody not to link to my sites, but please don’t, since it would mess up the experiment.
Similar but rewritten posts are placed on all the sites in order to make them all unique to the eyes of Google. They all use the standard sitewide bloggroll links to link to each other. The sites will also use Google Sitemaps and ping Google so that they will all be indexed at about the same time.
Since the Google toolbar PageRank is split into only 10 values, of which only 0-5 are realistic to get, each toolbar PageRank will not give a lot of data. However if I get proper links to the feeder domains and observe the experiment over a few PR updates, I may get useful data in order to make a guess on the logarithmical base that Google uses.
My humble guess is that with no inlinks to the feeder sites, nodes A, D, E and F would get a PageRank of 1 and nodes B and C would get a PageRank of 2. Twith the logarithmic base being 5, an increase in inlinks to the feeder sites making them PageRank 1 should then push the link power upwards and make A, D, E and F into PageRank 2 and B and C would become PageRank 3. However I don’t think the PageRank works exactly as in the theory described at Wikipedia. I suspect that Node C will get much less power than Node B, compared to the image above. This is because mathematically, a site would pass on all its link power in a single link and if there are two links, it would be split in half. In my humble opinion I think that it’s more likely Google values both links at perhaps 70% of the single one, making them more valuable in combination than the single link.
Here are the initial links to the feeder domains: Feeder 1, Feeder 2, Feeder 3, Feeder 4, Feeder 5