Based on my general interpretation of the systems, BPI does factor margin of victory (in a "diminishing returns" way, for lack of better description — it doesn't reward piling up blowouts), RPI apparently does not. So as much as I think ESPN overhyped BPI in the early stages, the committee applying more extensive metrics will help with accuracy.
Also we have to remember, valuing close losses better than losses by wider margins, it works the same way with close wins (e.g., the win over Indiana State almost counts as a "tie.")
There are a few reasons why RPI will never go completely away. It is being diminished for ranking purposes, and will continue to follow that trend, but to get INTO the tournament in the first place. I believe it will continue to be the baseline. Then they will use the other metrics to re-rank/seed teams with something more comprehensive.
First and foremost the 2 main reasons that RPI is sticking around is
1) It greases the wheels as far a scheduling goes. If you use its guidelines to schedule, it brings order to the cupcakes and makes teams schedule "appropriately." This makes the non-con games more desirable and most importantly more marketable. It flat out makes the NCAA money to have teams schedule appropriately. If you can eliminate the risk of playing someone you're going to blow out by 35/50 by taking the MOV into account then teams don't schedule interesting games in the Non-con.
2) The NCAA for a lot of reasons will never be okay with rewarding blowouts. One thing that the wins are wins theory measures better than Ken P and BPI is the intangible of making the pressure shot. If a win by one is only marginally better that a loss by 1 then you get McDermott weird results too. The reality is those last 2 points are the most important. You can always say what you want about if Stu makes a bucket its still worth 2. They are very different from the clutchness that was Georges in the last possession of a game. It matters and The advanced metrics diminish the importance of that winning bucket and the "winner" buck in close games refines and distills the "advanced metrics"
ps for the clueless folks that say winning hurts your RPI, please you don't know anything about the topic, please quit talking and do some research.