There was a study recently that used historic weather, solar irradiance data and energy use data and worked out that a certain mix of (over-provisioned) solar, wind and 3 hours of storage would have sufficed to provide energy during almost all hours of the studied decades.
I'm not sure where you're getting that "almost all hours" from? Here's the relevant figure, which seems to indicate hundreds to thousands of hours under your given scenario:
https://www.nature.com/articles/s41467-021-26355-z