The framework race React wasn’t invited to
Chart last week’s npm downloads for six hand-picked frontend frameworks and Vue leads the web by more than double. Add the one framework the chart left out and the story rescales: React out-downloads the entire chosen lineup five times over.

The claim
Vue is the web’s favorite framework — 12.9 million npm downloads last week, more than double its closest rival. Developers have made their choice.
The trick
Every number is real and current — the editorial work happened before the first bar was drawn, when someone decided who was in the race. Vue really did log 12.9M downloads that week, really is more than double Angular’s 5.5M, and really does lead this lineup. But the lineup is the trick: React logged 146M downloads the same week — eleven times Vue, and five times the six charted frameworks combined — and simply isn’t on the chart. A comparison set chosen after seeing the numbers can crown almost anyone: drop the leader and second place inherits the title, with every remaining bar arithmetically correct. The chart never claims React doesn’t exist; it just never mentions who’s missing, and a leaderboard looks complete by default. (This exhibit is our own demonstration of a pattern endemic to vendor benchmarks and “state of the ecosystem” slides, drawn from live registry data rather than redrawn from one specific published chart.)

The honest version
Put everyone in the room and let the axis do the talking: with React’s 146M drawn in, Vue’s bar becomes a sliver and “the web’s favorite” is revealed as “the favorite among the frameworks we picked.” The tell to look for is a missing heavyweight — when a comparison crowns a winner, ask who would obviously belong and isn’t there, and be suspicious of qualifiers doing quiet work (“among modern frameworks”, “excluding legacy players”). The honest chart also confesses what the metric can’t say: Preact’s 23.4M mostly arrives as a transitive dependency of other packages, and downloads count CI robots and mirrors as happily as humans — so even the complete chart is a download chart, not a developer census. A fair fight needs both: the full field, and a metric that measures what the headline claims.