Walker --
He recognizes drug approval is an art cloaked in science. Why, for example, have we chosen p=0.05 as the threshold for approval? Why does it matter that we choose a primary endpoint? Why not just choose a range of endpoints, as long as they are all measured off the ITT group? (Don't bother to try and prove this to me with a math equation, I'm perfectly familiar with the equations.)
Too many people forget that while everyone can verify the math associated with biostatistics is accurate, it does not mean the conclusion they imply is accurate. Biostatistical "proof" is only as accurate as the artfull assumptions it is based upon.
Walker is one of those people who understands that just because some fancy equation tells us something is "successful" or "failed" doesn't mean we should slavishly follow those conclusions if logic and common sense suggest a different interpretation.
Some dismiss such radical thinking as "lowering standards" or "compromising integrity of the process" but they completely miss the point. We can have standards and integrity and even a defined process without abandoning our common sense.
The other thing Walker understands well is how it is particularly important we allow a place for logic and common sense to override assumption-laden math when we are talking about human lives. We do not know so much about the human body. We need to allow for the fact we can be surprised at an outcome of a clinical trial and still derive something of immediate benefit to patients.
Even if the math tells us otherwise.
Market Data 
Markets 
AI
