Fable
Two messengers in the city's small newspaper: one picks and chooses, one searches everything
The town newspaper hired two couriers specifically to gather news leads from the market. Old Shi was a veteran who had worked the market for over forty years. Every item he brought back was reliable, but among the mere five or six items he always missed some important ones. Xiao Gu was new and could bring back forty items in a morning, covering almost every corner of the market, but twelve or thirteen of them were either duplicates or inaccurate.
At first, the editor-in-chief only paid attention to Old Shi, because Old Shi's drafts did not need review and went straight to the page. He always read Xiao Gu's drafts once and deleted them once, complaining that it wasted time. But in one issue reporting counterfeit goods at the market, Old Shi reported three problematic vendors, while Xiao Gu reported eleven—and the later investigation confirmed that all eleven indeed had problems, while Old Shi had missed eight. One of those eight later had a poisoning incident, and the newspaper was sharply criticized.
Only then did the editor-in-chief understand that Old Shi's problem was not that he searched incorrectly, but that he never went to those corners at all. His precision was close to 100 percent, but his recall was less than 30 percent. Xiao Gu's recall was over 90 percent, but his precision was only 60 percent. Replacing two couriers with one meant either being safe but missing big things, or being comprehensive but reviewing manuscripts every day.
The editorial department held a three-day meeting and reassigned the two couriers. Old Shi handled verification—Xiao Gu brought back forty leads, and Old Shi picked out the most reliable parts that did not need further review, while marking others that needed a second round of confirmation. Xiao Gu only handled searching, with no limit on whether leads were repetitive or rough, but every lead had to include a source rating.
The new layout process also changed. The front page only published news verified by Old Shi, with a small list of all clues found by Xiao Gu attached behind it, marked as pending confirmation, confirmed, or ruled out. Readers could judge the source level themselves, but the newspaper was only responsible for the verified parts.
The newspaper's approach was discussed by peers for a long time, but half a year later no one was arguing about whether it should be done—everyone was arguing about what precision to set and what recall to set. Each newspaper chose a different balance point based on its own manpower, budget and reader tolerance.
The editor-in-chief later told new editors just one thing: precision is the money you spend to protect your credibility, recall is the money you spend to avoid missing things. These two sums of money cannot be each other's budget.