Saved in:
| Main Authors: | , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2512.00836 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915645561503744 |
|---|---|
| author | Howerton, Emily Lessler, Justin |
| author_facet | Howerton, Emily Lessler, Justin |
| contents | Counterfactual scenario modeling exercises that ask "what would happen if?" are one of the most common ways we plan for the future. Despite their ubiquity in planning and decision making, scenario projections are rarely evaluated retrospectively. Differences between projections and observations come from two sources: scenario deviation and model miscalibration. We argue the latter is most important for assessing the value of models in decision making, but requires estimating model error in counterfactual worlds. Here we present and contrast three approaches for estimating this error, and demonstrate the benefits and limitations of each in a simulation experiment. We provide recommendations for the estimation of counterfactual error and discuss the components of scenario design that are required to make scenario projections evaluable. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_00836 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Assessing model error in counterfactual worlds Howerton, Emily Lessler, Justin Artificial Intelligence Counterfactual scenario modeling exercises that ask "what would happen if?" are one of the most common ways we plan for the future. Despite their ubiquity in planning and decision making, scenario projections are rarely evaluated retrospectively. Differences between projections and observations come from two sources: scenario deviation and model miscalibration. We argue the latter is most important for assessing the value of models in decision making, but requires estimating model error in counterfactual worlds. Here we present and contrast three approaches for estimating this error, and demonstrate the benefits and limitations of each in a simulation experiment. We provide recommendations for the estimation of counterfactual error and discuss the components of scenario design that are required to make scenario projections evaluable. |
| title | Assessing model error in counterfactual worlds |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2512.00836 |