There have been experiments (computer simulations) that tried to explain 'Pareto distribution' -- a power law distribution of wealth leading to 20 percent of population having 80% of the wealth. I refer to 1996 Axtell and Epstein book (google for Sugarcape).
Basically, they created a number of software 'agents' and placed them on a landscape full of resources. Each agent possessed a 'fitness function' that controlled its further survival. Having access to resources improved the fitness function since the agent could replenish its energy and could go on hunting for resources having some surplus for the rainy day.
What transpired from the simulation is that 'the Rich get Richer' axiom is basically true. First generation agent might have stumbled on a resource nearby by accident. It has improved its fitness function and he was able to successfully compete on the next round.
Meanwhile similarly fit agent who had bad luck in the first round (no resource nearby) was left chronically under-fit and his chances of 'making it to the top' diminished with every round.
This piece of empirical knowledge makes two important points about the society and the wealth.
Firstly, even in the free market society people always have unequal opportunities because this is an emergent quality of the system, not the result of some corruption, inability, personal passivity etc. The best and the brightest will stay poor unless they stumble on a resource which is not taken. (A new internet business idea for example).
Secondly, revolutions have their merits. They re-shuffle fitness landscape. Even when people do not directly 'expropriate' somebody's wealth by simply levelling the playing field revolutions create competing chances for the talented and the hard-working who would otherwise be forever destined to be relatively unfit comparing to those who already have access to resources.
Looks like extreme pro-market views are as primitive as extreme 'left' views. Social science really becomes fascinating when aided by computer science, doesn't it?