Hmm, I dunno, you may be unnecessarily demanding here.
I would think, if something is proven to work then yes, X phenomena will be observed given variables Y and
Z are also present. So, if Y and Z are present in the right context, we can be as certain as we are ever going to be that X will occur.
If X doesnt occur as expected, then we have improperly identified the other variables, and/or there were hidden variables our crude measuring systems didnt quantify.
In other words, IF all the requisite variables are in place, creatine WILL work 100% of the time. By definition.
If creatine doesnt work, the requisite variables are not present. This does nothing to disprove creatines efficacy, in fact quite the opposite.