Beevee's Owner/Giant Slayer
19,381 posts
BV is obsessed with my definition of survivorship bias because he doesn't know what it is, so he needs me.
Survivorship Bias:
Survivorship bias is the logical error of drawing conclusions (something that BV does very well) from the things that made it while ignoring the things that didn’t. This could be due to (but not limited to) failures that are invisible, discarded into BV's jackhole, or
inconvenient to the story being told.
The underlined reason is usually what BV does, as all of you already know.
Now let's compare this to his definition:
"it is the mortality of the data."
"mortality of the data” is a fascinating definition, mostly because it applies to almost everything, which means it explains almost nothing.
By that logic:
- My unread emails suffer from survivorship bias
- Dinosaurs committed survivorship bias by going extinct
- My gym membership proves survivorship bias because only my couch survived
Data “dying” is not survivorship bias.
Selection based on survival is survivorship bias. If mortality alone qualified, then every dataset ever would be guilty, which is less a definition and more a poetic shrug.
Blackvegatble's hypcorisy summed up in one post:
[/size]
Blackvegetable » 7 minutes ago » wrote: ↑7 minutes ago
Very simple questions...
From which you are running...