Not sure I agree with the two of you. Legit companies with legit drugs will presumably be able to differentiate their drugs from others in the marketplace with their superior drugs. The marketplace will be the ultimate arbiter.
Suggest you aren't understanding. Without rigorous data it is, literally, impossible to sort it out. As a way to think about this, you yourself do not want to hold companies through their phase 3 binaries because 90% of phase 2s are completely inadequate to determine whether the drug is clinically useful. And the 'reformers' essentially want to move to something much much closer to phase 2 'proof' as adequate for approval.
(Examples of such proof are: single arm comparison to historicals (do you consider such phase 2s adequately dispositive to hold through a phase 3?), biomarkers (do you consider such ph 2s adequate to hold through a clinical endpoint ph3?)). And you undoubtedly are better at sussing out real phase 2 efficacy than 90% of MDs.
PS Sorry for being a little pointed, but really don't think people are thinking this through. Over the Xmas break I'll screed on some related topics because there are some things we could probably do to improve the process. But only around the edges. Essentially we are already fairly close to the Shannon's Limit of biomedical research (Shannon's Limit is a about how much information you can get through a system).