Using that logic then biotech investments would only be left to late stage 3 - by that time MOST of the value of the drug is built in to the stock price
This is clearly not true - given the huge pops that follow some successful ph iii's. If RPRX or DNDN get approved they will be much higher than they are now.
And note that your response is tangential to Dew's statement that, to put it more bluntly, MoAs are a dime a dozen. Virtually EVERY drug has an MoA touted by their developer. Even after it fails most management teams come up with a new or modified MoA. It says more about the management team than the science.
To restate what I believe to be Dew's point - MoA is clearly hypothesis generating. But it is too easily generated post hoc and thus should be treated with HUGE scepticism - hard data on commonly accepted (or completely pre-specified) endpoints is much less easy to game. And note that good companies often get such data well before ph iii.