[The `sklearn` documentation discusses calibration.][1] One of the example models (in orange) is a naïve Bayes model that has a descending calibration curve, meaning that observations with larger estimated probabilities of occurrence actually occur less often.

[![sklearn calibration ][2]][2]

A member of this community has posted an [example][3] where the same phenomenon occurs but is even more visually extreme.

[![The Great calibration][4]][4]

Those seem to be examples where the rankings are wrong.


  [1]: https://scikit-learn.org/stable/auto_examples/calibration/plot_calibration_curve.html#sphx-glr-auto-examples-calibration-plot-calibration-curve-py
  [2]: https://i.sstatic.net/yv4od.png
  [3]: https://stats.stackexchange.com/questions/570095/how-to-calibrate-models-if-we-dont-have-enough-data
  [4]: https://i.sstatic.net/ALnHt.png