Reminds me of stats that were reported about the number of students who successfully landed co-op jobs each semester.
If a kid didn't get a job before their study semester was over, they'd just take their next study semester early and hope to land a co-op job in the next semester.
The school wouldn't include their headcount in the number of students who didn't get a job that semester and the employment rate looked wildly successful.
It took me a while to understand how survivorship bias related to the original post.
What I think it's saying is: if the previous owner installed heavy armored plating in their kitchen, and I moved in and noticed all of the fork marks in the bedroom carpet, I'd make the false assumption that they only dropped kitchen utensils in the bedroom.
i.e. too much forking in the bedroom, and none in the kitchen.
So if I replace all the flooring, I wrongly install heavy armoured plating in the bedroom and not the kitchen.