I remember that, and I concurred. That was the path to take if you believed my "guess #1." The beauty of that scenario was the you could select a guess and then get the chance to substantiate your guess or adjust.

I did a little more digging and found a great post by BLAfarm on the DSMC analysis. Check out number 63113

Unblinding, as you suggested at the time, would have penalized us on the p-value a little bit, but would have told us everything we needed to know in order to course correct. (My understanding is that unblinding at the point, would have still allowed the trial to continue. It wasn't a full unblinding.)