That’s a valid point. But I’m not trying predict whether he will, let’s say, win another monument or b not because those legends did after 30. I was responding to someone saying he’s washed up and simply won’t win big again. I was trying to show various reasons (including his performances last year) why I think he still could, including examples of past riders who performed well after 30 in the classics. If I wanted make a statistical probability model of his future success based on other riders past results then I indeed would need to take into account all riders in similar circumstances. But that wasn’t my point.Just saying… if you are gonna use existing data you are gonna have to look at it from both sides.
Here’s an analogy:
Someone says to me, it’s never going to snow on Christmas again.
I say, it’s true it hasn’t snowed for several years but it did snow on that day in 2019, 2015, and 2009 and once every 5 or so years before that. I don’t know if it’s going to snow again but it seems likely it still will again. The fact that it didn’t snow those many other years doesn’t change my opinion that it still could.