Comment by shw1n

I personally think this is the right approach, where you can assign probabilities to the unknowns (the "thinking in bets" section)

Because then the discussions/research switch toward data and evidence, with the results downstream of those

Overall when people can agree "I understand stance 1 if the data says X, or stance 2 if the data says Y", and then all the energy goes into the data analysis, I consider that a successful conversation