What's puzzling is how analysts underestimated the complexity of SPA and assigned such weight to it
In general, most investment banks have crap research. Either the analysts follow too many stocks and cant devote too much time to each issue they follow, which in turn, leads to laziness in real assessment of the situation, or, in the case of the boyz at Roth, Rodman, Lazard, they'll say ANYTHING to get the banking business.
I don't mean this as an insult at all, in fact, it's quite the compliment: I have seen better analysis on companies on this message board, and, in this case, arguments about this SPA, than almost any research piece from an investment bank.
Without that, I suspect they will be less reluctant to invest in a 10- million dollar non-pivotal study.
How do you know 306 is not potentially registrational?
Pain is a valid clinical target for an NDA submission. It may be harder to hit, and impossible to get an SPA for, but that does not mean it can not be a basis for an NDA.
The commercial value of this might on it's own be questionable, but as you say they already could be on the market so the broader indication could be of more value than a stand alone pain NDA.