IMO, HD's version of RPI, even though based in reality, should be changed. It gives double the amount of weight to what opponents do versus if a team actually wins or not. The math, in this case, is misguided. Now, it might be ok for RL use, where the selection committee uses it (at least this is what they ALWAYS say) only as a guide to differentiate teams that are ALREADY closely matched using OTHER criteria. In HD, it's a great deal more closely correlated to inclusion and seeding. So if HD is going to use such a formula, I think it should simply put more weight into actual win-loss percentage, relative to what opponents do. That is still important, but twice as important as actual win-loss percentage? I'm not buying it.
Btw, this is not a defense of the original poster, who seems to lack understanding of how RPI is going to continually adjust throughout the year, as previous opponents' win/loss percentages change dramatically, comletely independent of what the team does.