I was incorrectly computing load on a distribution of read fractions. I have to compute the load for each fr separately and then weight them.