History is a single, continuous, seamless, uninterrupted whole. Everything matters. Everything. Not the sum of things, not single things, or individual things, but every thing. To understand only a little of it is to understand almost nothing. It's impossible to single out any given point and to explain it without reference to any others. There is nothing that exists within the great tapestry of history that exists on its own. To put it in a semiotician's words, history is a semantic network of events that relate to one another causally and temporally. We simply cannot divide it into its constituent parts without losing some of the meaning of that part. The smaller the part, the more meaning lost!
For this reason, every action is important. Every individual action forms the rich cloth of history, which means that the simplest of things really do matter. Material culture, from the mattock used to break up the dirt in the field to the seats upon which people sit, is part of this tapestry as much as any actions of individuals are. Great movements are the amalgamation of individual wills, and the "great people" of history surface only from the undifferentiated sea of meaning by degree not by kind.
What does this mean for roleplaying games? Not much if you're not interested in trying to emulate something approximating reality. If there's no desire for verisimilitude, it makes little difference. If, on the other hand, there is a hope that you will approach something that feels like the real world and functions more or less like it, it can mean a great deal.
For one, the idea that PCs can't be important because they are not really any different from the great unwashed masses, is ridiculous. They are no different, that is true, in essence. The true difference between them is that they strive to stand out and be movers and shakers (again, a difference of degree, not of kind). The opposite is also true: they cannot work their will wholesale upon the world, because its a complex place and all manner of things will occur that will prevent them from achieving the goals they want. Even when they believe they've done something good, it may affect the hopelessly complex system (for the functioning network of meaning that is history is a working system in the present moment, an insane Rube Goldberg machine wherein all the parts affect all the other parts) and result in outcomes wholly unintended.
This would also argue for the presence of rich history, that is fleshed out history, which the players can interact with. The more that you, as a DM, know about the setting the more it makes sense and the easier it is to approximate this complicated web. One of the reasons that random tables are so useful is that they can give the approximation of the complex system without relying on the rich history and mind-based computation powers of someone attempting to run it based on logic. But that hardly precludes them from working together, which is what I find easiest and most rewarding in balance: rich history for when I can conceive of how all the moving parts in a situation may work and complex tables and charts for when I am unsure of the vagaries.
As a DM, I am most comfortable when I know the most about a setting. I feel lost or ungrounded in vague settings until I do some footwork and flesh them out. I know this isn't true of many people, but I am prepossessed by that strange dictate which desires a rich and deep history to move forward. Rather than retcon things into making sense, or exercising logic that must extend backwards and make sense of previously nonsensical things, I prefer have access to all my decision making tools up front.
Therefore, I am myself an advocate of the deep, blanket history—the understanding of a setting that allows you to wade into that marvelous and horrifically complex system that is time and the present, to stand at the heart of that strange orrery, and to make it your own.