## PageRank (or PR) leakage

Whether PageRank Leakage exists or not, is a question of semantics. The PageRank for a given page
is solely determined by the **in**bound links.(1)
However, as the spreadsheet will show,
an outgoing link can drain **the entire site** for PageRank.

We have just set up two separate sites - with one page (named Page5) linking to a page in the other site. A picture of the spreadsheet can be found here).

Then we studied how the linked-to site was able to
increase its PageRank.
We will now look at how the link**ing** site fares.

First iteration |

**guessed**at first, that the PageRank transferred from one site to the other would be 0.85 * (1/5) = 0.17. If we look at the first iteration (picture to the right), we see that Page5 has retained all of its PageRank - since PageRank isn't influenced by

**out**bound links. This is what makes many people declare that PageRank leakage doesn't exist.

The Page Ranks of the 4 other page have dropped, however, since they are now getting a smaller share of Page5's PageRank. The loss equals 0.425 per page for a total of 0.17 - exactly as we predicted.

Second iteration |

**next**iteration all 5 pages get a lower PageRank. The total PageRank of this mini-site is no longer 5 or 4.83, but 4.6855.

Third iteration |

Many iterations later |

And notice another curious thing: For every iteration the leaking page (i.e. Page5) has the **highest** PageRank in
this mini-site. The other 4 pages have even lower PageRank because Page5 isn't doing its part. This slight difference
in PageRank will probably not have an effect in the real world - but if anybody tells you their pages perform better when they
link out - give it a thought.

A final note: This does not mean that you shouldn't link out.(2)
PageRank is only **one** part of Google's algorithm. I'm personally convinced
Google have lots of other ingredients in their secret sauce, which will reward sites that links out.

And now for The Big Picture