And a stick tang is the narrowest section. If the tang is the same thickness as the blade flats, as in a saber grind, then I don't know what your argument is, since stiffness is linear in width and cubic in thickness. Cherry-picking anecdotal results from one set as opposed to other results doesn't make either side right. Knives break in the blade, at the ricasso, and in the tang. Depends on many factors, there's no absolute authority on the matter.
The only authorities I appeal to here are physics and empirical evidence... not sure how much more "absolute" one can get without delving into theology, but you are right that it has to do with many factors, the primary being
use. One CAN break a knife in the tang, but empirical evidence says more knives & swords have broken in the blade than in the tang in use. The well known destruction videos posted by noss bear out the same result, but they are only referred to because of their popularity, accessibility, and the variety there presented. History/archaeology, and the vast majority of modern anecdote tell the same story. Physics states that fracture will most likely occur at the weakest points nearest the apex of stress, and most of the stress on a knife or sword occurs in the blade, specifically the ground edge. Most of the fractures occurring in a blade in use are due to lateral stress and the edge chips out - this makes sense as there is the least material support present there and also because "stiffness" can be detrimental when sufficient force is applied to cause the blade to either flex or fracture. When prying laterally with a blade of any grind, the apex of stress is usually well ahead of the handle & tang. Also, so long as the handle and tang are securely attached, the tang's thickness is compounded by the handle, reducing likelihood of any flex that could stress the tang to its limits of ductility, and I also suspect that the tangs of forged knives and swords tended to possess greater ductility than the blade, but I'd need a metal-worker to confirm that... Anyway, only in very specific cases is sufficient stress placed on the tang of a ("well-made") knife or sword to induce fracture - the knife or sword is far more likely to fracture in the blade (which is thinner on average than the tang and esp. the tang+handle) or the ricasso (unusual, as the ricasso is thicker & stronger than the blade which is close enough that, even if the ricasso is the apex of stress, the blade should flex and/or snap first). The secondary importance of the tang allows for a great variety of tang designs, including that of folding knives, hollow-ground and skeletonized tangs, tapered tangs (e.g. loveless knives), stick tangs, etc. So long as the handle is well made and the tang has no anomalies like welds or inclusions, there should be no issues with stress in the lateral dimension. A stick tang is characterized by reduced breadth compared to the blade. This could mean that the blade has more metal supporting it against stress in the vertical (edge-to-spine) direction than the tang has (particularly in saber-ground knives and swords). This is also the case in hollow-ground and skeletonized tangs. However, the blade again also endures a greater portion (many times over) of this stress in normal use, and along more of its length, and usually at some distance from the handle, increasing likelihood of fatigue and fracture prior to that in the tang. The increased breadth and stiffness of the blade are not an advantage over the skinny tang when the force is always focused there - the tang may flex to endure its small portion of the stress-load, but the stiff blade may simply shatter apart as it bears the majority of the load. Thus, tang construction is of secondary importance in durability, blade construction is primary. I'm not saying that the tang is UNimportant ... just less important, because it need only be strong enough to outlast the blade.
What "cherry-picking" are you referring to? And I'd gladly be schooled on this subject, I just haven't come across objective information to the contrary. *shrug*