OK so he was referring to the total alpha of 0.025. Is that one-sided or two-sided? I never doubted it was one-sided. But then he did refer to the single trial approval strategy twice, e.g.
"again, it's important to point out that because this trial, as a single trial, is looking for a p-value of 0.025, that makes it even harder for an interim look, because what ia is trying to do is to achieve that more stringent p-value in the end, so it's not surprising IA did not succeed".
Perhaps I am over-reading - if it were a one-sided alpha of 0.025, then there is nothing stringent about it. However, it would be if the alpha were two-sided.
Market Data 
Markets 
AI
