Closed mlbtc closed 3 years ago
Let's see if things improve with updates to the data this week.
This is still an issue in the most recent data update.
This is likely due to the fact that there are too many zeroes (more than 20% of the data). Can we confirm this with the raw LOS scores?
Confirmed - these are Chicago data but it will be the same elsewhere:
Histogram
Quantiles
Not seeing any bottom 20% BGs