jfm,
Your a riot!
"Only significantly (non parametric definition) more deaths/events than can be expected or markedly more than the placebo throw a flag. Then they intervene."
Cutting thru all the flowery language, you finally see it my way. I have stated for some time that I believe the science supports the notion that death will have the highest RR of the components. If you read as well as you write, you know my opinion already. A reduction of 40 or 50% would get these guys to look at what is going on. The DMC meet quarterly...I believe HG mentioned 16 or 18 times so far. So it starts at roughly equal deaths in both groups, but say by meeting 10,(does it really matter) they start to notice a significant difference in death(steadily climbing). Ok, they are coded, I mentioned as much before. That means nothing if they see group A has many more deaths than group B. They have the ability to uncode at any meeting if they see it necessary. In my opinion, that would qualify as necessary. I just don't believe that they would see a large reduction in death and wait until 60% readout. Remember...safety first. Now, maybe you see death as being more like a 15% difference between the arms and as such, the DMC saw no safety issues and only first uncoded at the 60%. I certainly hope that is not the case. Let's see how it plays out.
On a lighter note, take it from someone who has owned more than his share of Porsche's...the 911 Carrera 4S should be a consideration. Have you picked a color? They will paint one any color you want...all you have to do is bring a sample and they will paint it that color at the factory.
