Just adding my voice to the chorus...
Don't read too much into the statistics. At best, we get a guesstimate from QT. I didn't start using QueryTracker until I'd already trunked a project for lack of interest. I've entered a bit of that data, but I really don't have the inclination to go through 100+ queries that got rejections. It was a 135,000 word science fiction novel that didn't have a space war in it. I could have written a great novel (it wasn't, we'll leave it at that), and a lot of agents would have passed on it just because of length and genre. After such a disaster, it was a bit of a shock to start getting requests on the next project I queried. What was the difference? Better writing, better query? Probably. I'd written two more novels, I was bound to get better. Was it more saleable material? It's an urban fantasy, so yes, I'd say that probably was a factor. Do I really know why? No. Will I ever no why? Nope. And there are other factors to consider, none of which we know.
So I wouldn't throw out agents that don't take on a lot of clients. While it's true that newer agents are more open to new writers, landing an established agent raises your chances of a sale, and I think the odds increase dramatically the bigger the reputation the agent has. Do I know what those odds are? Nope. Do I have any expectation of ever knowing those odds? No. So I've got someone in my query queue who only takes on 1 or 2 novels a year. I figure my chances are about as likely as a hundred degree day occurring in Minneapolis in January, but I won't know if I don't try.