90%inside the likely range, 1 month ahead
71%inside the likely range, 3 months ahead
0.23 / 0.23average miss (points), ours / “no change”
Does not beat “no change” yet
48months tested, each with only the data known at the time
The aim is 80%: the real change should land inside our likely range 8 times in 10. Much higher means the range is too cautious; much lower, too bold.
Forecast vs what happened
our likely rangeour central forecast
official result, insideofficial result, outside