Suppose that there are p groups of people that do make up statistics, and q that don't. This makes for a total of p+q groups. Assuming that any statistic is equally likely to be given out by each of these groups, the chance that a statistic is made up is p/(p+q). It is a property of fractions that this number can be made as close to any given probability as we desire by suitable choice of p and q. Therefore, when we say that "x percent of statistics are made up," as long as the number is between 0 and 100 percent inclusive, we are always right somehow.
Confuddling people with maths can be a scintillating experience.
You mad, you lose.
People do not like to think. If one thinks, one must reach conclusions. Conclusions are not always pleasant. —Helen Keller Le doute n'est pas une condition agréable, mais la certitude est absurde. —Voltaire