It strikes me that the advantage of engaging in reciprocal linking essentially is a means of aggressively redirecting pagerank held within a pool between two sites towards the homepage (or other target pages) of each site. In a perfectly neutral arrangement, there would be a balanced amount of PR exchanged by each site. If this is done correctly then a series of links will be generated that will educate the SE's via anchor text as to what the targetted pages are relevant to.

So the question that follows is how to establish a comparable exchange.

Of course if all sites were at the same PR and had the same number of links per page, then this would be a silly question to pose. But they are not. So I'm hoping that we can establish an effective "conversion system" for SEO's both new and old to use. Certainly for many, this is automatically done in their heads. But I like numbers that I can play with to ensure accuracy and fairness. I'll start with what I've learned so far. I'm hoping that others will reply to this thread and suggest some helpful modifications!

First we start with the PR Alorithm:
PR(A) = (1-d) + d (PR(T1)/C(T1) + ... + PR(Tn)/C(Tn))

This algorithm provides us a basic concept of the PR of a page PR(A) based off of the PR of the pages that link to it (T1, T2, T3...Tn). A dampening factor (d) is applied which most consider to be 0.85. Finally, the weight of each referring page is divided by the number of links on that page C.

So an example would be:
PR(yoursite.com)= (1-.85) + 0.85(PR.of.Link.page1)/(number.of.links.on.page1) + 0.85(PR.of.Link.page2)/(number.of.links.on.page2) + Etc

There are two ways to continue from here. The first way would be to simlpy plug in the PR found on your Google toolbar as well as the number of links on each page. The other option is to correct for the PR that the google toolbar shows for "real pagerank." It's probably best to do the correction here so that you don't have to convert it later. Since PR is based off of a logarthmic scale, there is an exponential increase of value as you go. It has been reported to function off of base 6, 7, or 8 but I will use 10 for now since it is easier to keep track of the math. We would then convert to:
1-10=Pr1
11-100=2
101-1000=3
1001-10000=4
10001-100000=5
100001-1000000=6
1000001-10000000=7
etc...

Now comes the fun part. We can then integrate the above conversion factors into our basic algorithm and find that we now have a fair means for assessing link exchanges.

PR1 link with 20 total links on page: 0.1925=(1-0.85)+(0.85*1/20)*1
PR2 link with 20 total links on page: 0.6175=(1-0.85)+(0.85*11/20)*1
PR3 link with 20 total links on page: 4.4425=(1-0.85)+(0.85*101/20)*1
PR4 link with 20 total links on page: 42.6925=(1-0.85)+(0.85*1001/20)*1
PR5 link with 20 total links on page: 425.1925=(1-0.85)+(0.85*10001/20)*1

With this technique, we can now modify the variables to account for changes in PR, number of links per page, and number of pages linked to. For example:

200 PR1 links with 20 total links per page: 8.65=(1-0.85)+(0.85*1/20)*200
1 PR4 link with 100 total links per page: 8.65=(1-0.85)+(0.85*1001/100)*1


We now have a simple means to be assured that we are engaging in Pagerank neutral trades. This system assumes that the weight of a link is also directly related to the PR of the page listing it so that a PR5 page will provide more weight to the SERPS than a PR3 page would with all other factors being equal.

__________________
Find a roommate with a Roommates : Roommate Finder