It looks like a 10Y-3M inversion has predicted each of the last 3 recessions by about a year with no false positives in the last 35 year.

https://fred.stlouisfed.org/graph/?g=o4ro

Of course maybe this time is different...