A simulated collection | Step 1: check the split | Step 2: internal versus external performance | Step 3: performance against distance to training | Step 4: intervals that respect clustering | Key points

Step 1: what is in the data? | Step 2: performance per predictor | Step 3: do outbreak clones flatter the results? | Step 4: intervals that respect clustering | Step 5: comparing the two meropenem rules on the same isolates | Step 6: stability over time | Step 7: a report for the paper | Caveats