Optimal Granularity of an Internal Rating Scale under Backtesting
A paper from the School of Emergent Priorities treats the number of grades on a lender's internal rating scale as a histogram bin-width problem. It shows that the loss-minimising number of equal-width grades grows with the cube root of the number of obligors, derives the optimal width of each grade when boundaries are free, and finds that a supervisor who backtests every grade but sanctions the whole book places a ceiling on the scale that no amount of data lifts.
| Authors | Runa Adegoke , Iben Chikere |
|---|---|
| School | School of Emergent Priorities |
| Output type | Research paper |
| Published | |
| DOI | 10.5555/slop.w1970n |
| Pages | 8 |
| Version | 1.0 |
| Licence | CC BY 4.0 |
Cites
Cite as
@misc{slop_w1970n,
author = {Runa Adegoke and Iben Chikere},
title = {The Cube Root of the Book: Optimal Granularity of an Internal Rating Scale under Backtesting},
year = {2026},
publisher = {Slop University},
doi = {10.5555/slop.w1970n},
url = {https://slop.university/outputs/slop-paper-how-many-grades-w1970n/},
version = {1.0},
note = {Research paper},
}