I partly share your intuition. Still, e.g., AU has highest migration/capita, yet ranks middle. Perhaps mobility is higher in countries with less inequality, which supports upward mobility, and where income quartiles are closer, making it easier to move from the bottom half to the top quartile?

Funny how to understand social learning in these externally influenced networks, one would probably have to start from... those basic mechanical-diffusion or naive-learning models (like the simplest ones used to model the spread of infectious diseases).

Assuming increased volume and quality of flows bring us closer to sufficiency, Ben's claim about *some* industries seems valid. Must it hold *across* most ind. for dominance? By analogy with requiring an underlying matrix to be either irreducible (some) or primitive (across), I lean towards "yes".

I'm rusty and might be missing something but... h is an info set ("horizontal" set of indistinguishable decision nodes). h^\preceq is a history ("vertical" set of decision nodes: one from each "eligible" h, where the partial order \preceq would have to be set up to determine the "eligibility").

Oh. Ok. From one of your earlier replies, it seems you may also need to keep track of histories. How about: h for an info set. H for the collection of infosets. H_i are i's info sets. H^t the set of possible histories at t, and h^t a specific history.

Pausing to carefully consider a notation choice... nice! For what it may be worth, I'm echoing Ben: iota for info sets and I for players --which is also consistent with MWG's excellent notation for extensive form games (pg. 227).

🙋

🙂 - thank you.