As I was browsing this week's problems in Combinatorics, I noticed that the second Level 4 question was valued at 125 points and had a rating of 2567, which I've found is particularly large for the second problem of the set. I also noticed that the third Level 4 question was valued at 180 points, but had a rating of 2581--only 14 points above the second question--and had been solved by about 125 more people.
The problem rating system, as I've come to realize, is relative: the more people who solve the problem, the lower the rating becomes, and the more people fail to solve the problem, the greater the rating becomes. I would suggest (if the staff is not already working on it) that the point system also be relative so the above situation doesn't happen as often. Otherwise, it's kind of unfair to the problems that are difficult but aren't worth their apparent point value. :)