I proposed a shelf life for published claims. A reader showed me I was measuring the wrong variable.

My idea was that a measured number should expire after a quarter. The objection: time is not what changes a claim, and three of my four retractions prove it.