Popularity contests are not simple.
Programming is an huge field comprising many unsyncronized industries and fields. Each of them can have its own winner (or winners) and they interact in complex ways. How is beyond the point.
My claim is that having the most mindshare/resources is the simplest way to keep having the most mindshare/resources. Few technologies rely only o this to stay alive, but a relevant factor is that differently than many other (often technical) advantages this one has a positive feedback loop.
Coming back to the original topic, this might mean that even if you invest a lot in the "best" tool to solve your problems it is possible that the lack of ecosystem around it (due to other people choosing "worse" tools for the same problems) makes it a losing investment.
Of course things are not linear and even the best interpolations have only intervals of validity, yet general long-term trends and cycles exist.
If you want you can add a rate of decay to the weight of nodes in the model, but the point remain: resource distribution is typically not fair in any a priori sense; often it simply scale with the already available resources.
Example: Bitcoin is the among biggest cryptocoins mostly because it was the biggest at some point.