freejung - 3:02 pm on Jul 28, 2010 (gmt 0)
As previously noted, getting hard data on this sort of thing is very difficult as there are too many variables in the "wild" to control.
However, I think there is some value in people's "gut feelings" when they have a lot of experience watching complex data patterns. Humans are very good pattern matchers -- too good, sometimes, as we often see patterns in genuine chaos.
I have long felt that there is hysteresis in the system. Rankings feel like they have velocity and momentum. This can be seen fairly clearly with your own internal link structure: two very similar pages with the same internal backlink profile seem to have very different ranking potential depending on how long the page has been up. There are several possible ways this could arise, but historical re-evaluation of backlinks would definitely explain it.
So my feeling is that historical data is probably being considered, and that it is probably highly context-dependant.
Generally, it feels like Google is taking into account the overall shape of the link graph, including its temporal variation, in a manner that is probably fairly simple mathematically but to a human observer with limited information it would be very difficult to predict.
Is that actionable? Well, sort of, but it probably just means what we already pretty much know -- that the best links are those that genuinely constitute an editorial endorsement from an authoratative source, because Google is getting better and better at determining algorithmically exactly what that means.
Every now and again, someone mentions a "trick" that I cannot believe is being discussed in a public forum.
I think that depends on your attitude. If you are paranoid and insular, yeah, talking about such things is foolish. I tend to think that symbiosis is the most effective evolutionary strategy, and thus sharing of information can be generally beneficial. Anyway, I don't think it takes a rocket scientist to figure out that promoting those who promote you is probably a good idea -- symbiosis again.