Great post at WMW

I had to share this. Probably one of the best analogies of the Google sandbox:

Look at it this way. It's spring. Competitive soccer tryouts are here. Let's say there's a competitive league for boys 8-12. This is not your run of the mill league; it's a very competitive league. Not all players on a team get to play equal amounts of time. The least capable play very little. But in order to always have talented players in the wings, each year the team drafts a few 8 year olds and a few 9 year olds. By the time they are 11 or 12, they'll be great. They get to practice with the big boys, etc.

So, you're the coach. You know that for the most part, the 12 year old boys are best. Then the 11's, the 10's, and so on. You automatically assume that the new 8 year olds and the returning 9 year olds won't be good enough to start, or even see much game time, except at the end, if there's a big lead.

All the kids are playing by the same rules, and all are judged by the same criteria. But the 8 year olds almost never start, and when they do, they have to be far superior to almost all of the other 8 year olds.

The current algo works in much the same way. The sandbox filter says that for an 8 year old to play (i.e., new site), it has to be judged to be of a quality comparable to a 12 year old (four year old site). The 8 year old has to show something special. But if he can, then he can start. It just doesn't happen very often.