> It's not remotely clear why the actual value seems to be so small, especially since it's not exactly zero. It is very clear actually: one or both of the mode

> It's not remotely clear why the actual value seems to be so small, especially since it's not exactly zero. It is very clear actually: one or both of the models producing the conflicting result are wrong. Smart money's on both, because we have good other reasons to think each of them is wrong.