"Group statistics have little predictive power for a single member of the group."
That isn't true. It isn't even true for your coin flip example. The comparison is properly made between membership in groups with different expectations. Anavex is far better off in an 85 percent expectation group than a 15 percent expectation group. You are better off in a 50 percent "coin flip" group than a one percent group.
The reason why the EMA news is a big deal is because blarcamesine is now known to have moved into a very favored group, even with all known drawbacks of the 2b/3 trial. Relitigating familiar 2b/3 issues at this point is therefore akin to double counting.
Further, within a group with a high expectation of success, like 95.7 percent (the EMA approval rate without considering withdrawals) for example, there is much less opportunity for meaningful expectation differences among subgroups. Say there are ten equally sized subgroups of nine among ninety fully processed submissions per year (approximating EMA annual opinion totals). Once one subgroup of nine experiences just four failures, the other subgroup expectations all move to 100 percent; 4 of 9 failures, 55 percent, is a hard subgroup expectation floor, just by arithmetic. (And if that were the case, you would see those qualitatively distinct numbers separately described.)
I realize there is resistance here to this good news for a variety of reasons that I can identify, such as bad faith, ego, and other psychological factors, but these are not actually refutations of the EMA basic math. You are giving that a try, but there are serious arithmetic limits to your counter argument, and even then you are making assumptions about blarcamesine's differences within its group for purposes of argument that have not been identified in practice.
Frankly, I do not believe there is anyone present here qualified to make this type of differentiation. It would require broad knowledge of EMA internal decision making. That cuts down the necessary expertise to involved EMA staff, or perhaps a professional consultant assisting in multiple submissions (if such people exist, which is doubtful.) Without the access to context, there is no opportunity to make a comparison between Anavex's chances and the chances of others in its current EMA group, leaving us, again, at 85 percent.
I make the necessary concession that we are not quite at the MAA submitted stage; I also have been unable to identify any obstacles to reaching that point, or seen any such obstacles identified here by anyone else. Missling, of course, has stated his intent to now apply. So my concession is not much of a concession as far as I am concerned.