Cabel, I think you have painted one scenario but there are a lot of moving parts. One big mover for RRR though is this is a trial that is based on “time to event” during the trial. My understanding of this is, if for example you compare one event in the placebo group that happens at 3 months with one event that happens at 6 months in the test group, these events are not statistically equivalent. The event in the test group that took longer to get to is statistically favourable to the event in the placebo group. That is my understanding. My guess would be that if your raw events scenario is correct then the RRR would be greater than 31% because if the drug is working it will likely be delaying event occurances in the treated group.  Just my guess

I have done a number of quick calculated scenarios similar to yours using event rates ranging from lower than ELIXA to similar to EXAMINE.  I think your add back of 5% for dropout participation is a little low but that is just my opinion. I have been using add backs of 20-30% meaning that I think it is reasonable to assume that the average dropout probably made it through 20-30% of the trial and contributed that proportate amount to total patient years of data accumulation. Another moving part and as I said just my guess. We shall see where the truth lies in awhile.