EMH is a mental-model of equilibrium, and is useful when used that way. Financial systems have multiple equilibria [1], so EMH "violations" come in two flavors

EMH is a mental-model of equilibrium, and is useful when used that way. Financial systems have multiple equilibria [1], so EMH "violations" come in two flavors: 1. Small deviation from equilibrium. This is where Renaissance Technologies hangs out. 2. Market herd expects the wrong equilibrium. This is where Nassim Taleb hangs out. On the scale of single years, 2 happens so rarely than you can't get statistically significant measurements. So, we end up with endless philosophical debates about whether EMH is "true". [1] https://en.wikipedia.org/wiki/Diamond%E2%80%93Dybvig_model https://en.wikipedia.org/wiki/Diamond%E2%80%93Dybvig_model