The previous article discussing optimal representation was basing its results on a proportional power of the population. I think they were using a factor of 1/3 for their equation. So, if you raise the Population Number to 0.4 and divide by 3, you come up pretty close to what they thought optimal for the US. (I think it depends on what population numbers you use. I came up with 815 using a population of 295,734,134 (2005 est). Using their 807 figure indicates they are looking at a population of 288M)

Of course this doesn’t explain the discrepancy between France and Italy, unless the CIA has the wrong population figures for those countries.  My proportional calculations produced 433 and 425 reps respectively which is still a fair bit less than their numbers. So maybe the proportion is a variable as well?

If you use the 2000 US pop of 281M then the proportion is 2.97 to get the 807 result. But it does nothing to get closer to their France and Italy numbers.

Later.