How we pick stories
No secret sauce. Here is exactly what decides what you see.
One event, one story
When nine outlets report the same court ruling, that is one story with nine sources, not nine headlines. Grouping happens before anything is ranked, which is why the front page doesn’t show you the same news over and over. The source count is shown because it is real information: eleven outlets on one story tells you something one outlet cannot.
Size means importance, not recency
A story’s size on the page reflects how much it matters, not how recently it landed. The weighting combines an importance score, how many independent outlets are carrying it, how severe it is, and how fresh it is. Newest-first is how you end up with a wall that tells you nothing.
The red banner is rare on purpose
Most days there isn’t one. It is reserved for the genuinely grave — assassinations, major deaths, catastrophes — and requires a high importance score, a severity classification, several outlets, and recency all at once. A banner that fires every day is decoration, and once readers learn to scroll past it there is no earning that attention back.
When we don’t know, we don’t pretend
Not every story has been scored yet. Rather than treat an unscored story as though it had been judged, the ranking leaves it to rise on the things it genuinely knows — corroboration, severity, recency. A guess dressed as a judgement is worse than an honest gap.
A human can overrule all of it
An editor can promote a story to the top, keep one off the front page, or nudge its weight. The automatic ranking is deliberately conservative, so it misses things; a person is expected to disagree with it, and when they do, the person wins.
Getting it wrong
We will. If a story is ranked absurdly, a source is misattributed, or a line of commentary misstates what happened, tell us — see Corrections.