Bullet proof is just an analogy here. In any ra

Bullet proof is just an analogy here.

In any randomized trail, there is always a factor of luck. In CD12, we got unlucky so that relatively more 65 year old/older patients in the treatment group (2:1 ratio treatment/placebo), so the efficacy of the drug looks bad without doing age adjusting analysis.

We could have gotten super lucky in CD12, having relatively more 65 year old/older patients in placebo group making the efficacy of the drug looks stronger than its real efficacy. But that didn't happen.

Since we already know that the age group distribution could be a factor of luck for it could potentially negatively impact our trial outcome, we can introduce age stratification in the randomization process so that age group distribution is equal in treatment/placebo groups.

The age stratification effort is to eliminate age as a factor to skew efficacy test outcome negatively/favorably. The effort is to make the trial result more trustworthy. So it is not a process of shoot first then put the target where the bullet hit.

There could be other unknown factors for any trials. To eliminate/mitigate those unknown factors, one way is to have a huge trial size to reduce the effect of luck.