I see this often, but I am starting to think this is a fallacy. A company can be further along down the wrong path, making it harder to undo the bad things it had done up to that point. A company can be based on an unsustainable idea and no amount of work put into it will ever make it viable. I have a feeling that real-world data is much more interesting than this. I imagine the risk curves are not strictly decreasing with progress from idea to product with users, and I bet YC could shed some light on this by showing when companies actually close down vs when they start at YC.