These metrics are totally meaningless. If some metric claims that there is a 50% chance of recession in the next 12 months, I'd expect it to be right roughly half the time it makes that confident a claim. Similarly, I'd expect a recession one third of the time that the metric says 33%.

The year leading into the Great Recession was mostly around 36%. One year before the 2001 recession it was 24%. And then we have spikes around > 40% in 1999, 2002, 2003, 2006, 2016, and 2018, none of which "paid off".

So how are we supposed to interpret these numbers? That a 40% chance of a recession has a 15% chance of actually predicting a recession?