That's a tough question. Xatu, Omastar, and Kabutops are the three highest rated pokemon in terms of overall weight, and while they are good, I don't think anyone would argue that they are literally the most powerful pokemon in all of Ubers. I do think that large differences in performance can be useful in determining a pokemon's worth the higher up in usage you go, because there is more of a sample size. For example, Rayquaza checks in with a poor weight of .688. Because Rayquaza is used so often, this poor rating indicates that many teams that use it struggle. Ferrothorn, however, has an average weight of .82. This shows that teams that use this are more often than not successful.
Though not every team with Rayquaza stinks, nor is every Ferro-based team good, the sample size for the two is more than enough to indicate that Ferro is more effective in this meta than Rayquaza is. The further down you go, the less information would be gleaned from weight. Suppose that for some reason that Arceus-Ghost was only used 5 times this month. In those matches, it went 2-3, and therefore will receive a low weight. Now, I recently made a trollish team with an Arceus-Grass check (toxic/light screen METANG!) and went 4-1. Does this make Metang better than Ghostceus, or just about anything other than Magikarp? Of course not.