Saved in:
Bibliographic Details
Main Authors: Landsgesell, Jonas, Knoll, Pascal, Wenzel, Tizian
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.29928
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909012398702592
author Landsgesell, Jonas
Knoll, Pascal
Wenzel, Tizian
author_facet Landsgesell, Jonas
Knoll, Pascal
Wenzel, Tizian
contents Tabular foundation models such as TabPFN and TabICL already produce full predictive distributions, yet prevailing regression benchmarks evaluate them almost exclusively via point-estimate metrics (RMSE, $R^2$). This discards precisely the distributional information these models are designed to provide - a critical gap for high-stakes domains where not all kinds of errors are equally costly. We introduce ScoringBench, an open and extensible benchmark that evaluates tabular regression models under a comprehensive suite of proper scoring rules - including CRPS, CRLS, interval score, energy score, and weighted CRPS - alongside standard point metrics. ScoringBench covers 97 regression datasets from diverse domains, supports transparent community contributions via a git-based leaderboard, and provides two complementary ranking protocols: an ordinal Demsar/autorank approach and a magnitude-preserving z-score ranking approach. Evaluating several models - spanning in-context learners, fine-tuned foundation models, gradient-boosted trees, and MLPs - we find that model rankings shift substantially depending on the scoring rule: models that excel on point-estimate metrics can rank poorly on probabilistic ones, and the top-performing model under one proper scoring rule may rank noticeably lower under another. These results demonstrate that the choice of evaluation metric is not a technicality but a modelling decision - and, for applications where e.g. tail errors are disproportionately costly, a domain-specific requirement with direct consequences for model deployment.
format Preprint
id arxiv_https___arxiv_org_abs_2603_29928
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle ScoringBench: A Benchmark for Evaluating Tabular Foundation Models with Proper Scoring Rules
Landsgesell, Jonas
Knoll, Pascal
Wenzel, Tizian
Artificial Intelligence
Tabular foundation models such as TabPFN and TabICL already produce full predictive distributions, yet prevailing regression benchmarks evaluate them almost exclusively via point-estimate metrics (RMSE, $R^2$). This discards precisely the distributional information these models are designed to provide - a critical gap for high-stakes domains where not all kinds of errors are equally costly. We introduce ScoringBench, an open and extensible benchmark that evaluates tabular regression models under a comprehensive suite of proper scoring rules - including CRPS, CRLS, interval score, energy score, and weighted CRPS - alongside standard point metrics. ScoringBench covers 97 regression datasets from diverse domains, supports transparent community contributions via a git-based leaderboard, and provides two complementary ranking protocols: an ordinal Demsar/autorank approach and a magnitude-preserving z-score ranking approach. Evaluating several models - spanning in-context learners, fine-tuned foundation models, gradient-boosted trees, and MLPs - we find that model rankings shift substantially depending on the scoring rule: models that excel on point-estimate metrics can rank poorly on probabilistic ones, and the top-performing model under one proper scoring rule may rank noticeably lower under another. These results demonstrate that the choice of evaluation metric is not a technicality but a modelling decision - and, for applications where e.g. tail errors are disproportionately costly, a domain-specific requirement with direct consequences for model deployment.
title ScoringBench: A Benchmark for Evaluating Tabular Foundation Models with Proper Scoring Rules
topic Artificial Intelligence
url https://arxiv.org/abs/2603.29928