> I expect two COVID infections would increase the risk like it seems to do with the vaccines.

Then the relevant metric should be infection vs vaccine+infection, given that natural immunity is better than vaccine-conferred immunity (which seems very bad with Omicron)