Whenever there's a change in methodology, it's common practice to retroactively recalculate the entire series so that the data are consistent. Economists aren't that dumb.
This still allows for you to choose the metric that best benefits your narrative today, which essentially every bls report has done.
I still remember them removing the people who had been out of the job for xx months during the 2008 recession from the 'unemployed' category. 'oh look unemployment is down!'. '...but so is the labor participation rate...'