Most (all?) economic theories are based on models. Models make assumptions about the world. Sometimes these assumptions are reasonable and the model will make useful/accurate predictions. Sometimes these assumptions are unreasonable, and the model will make bad predictions.
One of the hardest things to predict/make assumptions about are how humans act. The most accurate way to test if assumptions about human behaviour are correct is to run an experiment on actual humans. However running an experiment on humans at a scale large enough to actually test economic theories is expensive, and has legal and ethical issues (although for some reason if a government does exactly the same thing but doesn't call it an experiment it's ok).
MMRPGs (Massively Multiple Role Playing Games) sound like a decent way to tests how humans behave at a large scale with low risks.
For example you could imagine a tycoon style game where players buy, develop, rent and sell plots of land. Then you could test what impact Georgian land taxes or harberger taxes have on the 'economy' of the game.
Would it be possible to create an MMRPG with an accurate enough model of the economy to test economic theories whilst still being fun to play?
What parts of the economy would need to be modelled accurately, and what parts could be simplified/abstracted away?
What economic theories would you test, and how?
What would the plot of your economic MMRPG be?
How would it be funded in such a way that the game developers are incentivised to accurately test economic theories?
What types of inaccuracies would you expect results from theories tested that way to contain?