F1 score

Many tasks care about two things at once: not missing the right answers (recall) and not returning wrong ones (precision). The F1 score combines both into one number by taking their harmonic mean, which punishes a system that is lopsided on either side.

It is a common way to say 'which system did better' in a single figure. On the RAG page, a method losing '28.8 F1 points' means it scored that much worse on this combined accuracy measure.