dose, page-13

Anyone who wants to argue that the first Phase 2b trial of AOD9604 was problematic will find a receptive ear with me. There is no doubt that structural inadequacies and other flaws in that trial gave us a result that unquestionably set the project back by two years and hurt many investors.

Also, I take no issue with the fact that random outliers and normal statistical uncertainties can introduce additional variation.

Is it possible that this kind of normal statistical variation could completely account for the differing results for the individual dosages tested in the 2004 trials? One cannot entirely rule out this possibility, but I personally ran an analysis of just this question, and I found that with the number of trial participants involved in each of the dosage groups, it was EXTREMELY unlikely that the result could have been mere statistical noise.

It is often argued that the results did not reach "statistical significance" of a 95% reliability, and therefore should be discounted.

But, failure to reach "statistical significance" does not equate to "statistical insignificance". Statistical insignificance involves another calculation entirely (albeit a related one) and measures the chance that the result could be nothing but noise. I am not going to share the precise result I obtained, but I can tell you that it did not support the view that there is nothing going on with the drug.

But, as you are willing to grant that the drug may work, I'll grant that it may not. But, I am sure enough that it does that I have placed a rather substantial amount of my own wealth at risk on just that proposition.

It won't be long before I find out whether I have been an astute genius of incredible foresight, or a fool.