Abstract
Time-series forecasting models achieve strong benchmark performance but exhibit severe systematic bias in industrial deployments. This train--deploy gap is conventionally attributed to temporal-structural errors or distribution shifts. We characterize a complementary source that these explanations overlook: canonical losses embed fixed statistical priors, while industrial demand mixes benign and pathological regimes---zero-inflation, skewness, high variability---in which these priors are systematically violated. The induced bias persists even under perfect temporal modeling, remains in a distributional-shape component that normalization cannot remove, and creates an aggregation trade-off invisible to aggregate metrics. We turn these observations into an evaluation toolkit centered on the Regime-wise Relative Bias Vector (RBV): a metric-agnostic, regime-decomposed diagnostic that audits how pooled training allocates systematic mismatch across pathological subpopulations. A controlled attribution analysis decomposes RBV into a model-independent intrinsic floor, set by each loss's estimand, and an excess component attributable to training, tracing observed bias to the loss rather than the model. A large-scale study---13 loss objectives, 3 seeds, 60,000+ series spanning RetailShiftBench and M5, with random-split controls---shows that regime-aware diagnosis separates optimization-type from bias-type failure, and that regime-aware training resolves the pooling-induced bias that capacity scaling cannot, for mean-type losses. A formal structural observation, that risk under evaluation-distribution contamination is affine in the pathology mixture weight, grounds these findings. Our work complements model ranking with mechanism-grounded, regime-oriented evaluation.
Keywords
Subject
Publication details
- Journal
- Not available
- Open access
- Green open access
Cite this article
APA 7
Nie, P., Xu, C., Chen, Y., Ren, C., Hu, W., Yang, C., & Zhang, J. (2026). Beyond Model Ranking: Regime Diagnosis for Distributional-Statistical Misspecification in Industrial Time-Series Forecasting. https://omanscience.com/en/articles/beyond-model-ranking-regime-diagnosis-for-distributional-statistical-misspecification-in-industrial-time-series-forecasting
MLA 9
Nie, Pengyu, et al. "Beyond Model Ranking: Regime Diagnosis for Distributional-Statistical Misspecification in Industrial Time-Series Forecasting." https://omanscience.com/en/articles/beyond-model-ranking-regime-diagnosis-for-distributional-statistical-misspecification-in-industrial-time-series-forecasting.
Chicago (author–date)
Nie, Pengyu, Chenglang Xu, Yaoshi Chen, Chaogan Ren, Wei Hu, Chao Yang, and Jiangong Zhang. 2026. "Beyond Model Ranking: Regime Diagnosis for Distributional-Statistical Misspecification in Industrial Time-Series Forecasting." https://omanscience.com/en/articles/beyond-model-ranking-regime-diagnosis-for-distributional-statistical-misspecification-in-industrial-time-series-forecasting.
Harvard
Nie, P., Xu, C., Chen, Y., Ren, C., Hu, W., Yang, C. and Zhang, J. (2026) 'Beyond Model Ranking: Regime Diagnosis for Distributional-Statistical Misspecification in Industrial Time-Series Forecasting', Available at: https://omanscience.com/en/articles/beyond-model-ranking-regime-diagnosis-for-distributional-statistical-misspecification-in-industrial-time-series-forecasting.
Vancouver
Nie P, Xu C, Chen Y, Ren C, Hu W, Yang C, et al. Beyond Model Ranking: Regime Diagnosis for Distributional-Statistical Misspecification in Industrial Time-Series Forecasting. https://omanscience.com/en/articles/beyond-model-ranking-regime-diagnosis-for-distributional-statistical-misspecification-in-industrial-time-series-forecasting
IEEE
P. Nie, C. Xu, Y. Chen, C. Ren, W. Hu, C. Yang, and J. Zhang, "Beyond Model Ranking: Regime Diagnosis for Distributional-Statistical Misspecification in Industrial Time-Series Forecasting," https://omanscience.com/en/articles/beyond-model-ranking-regime-diagnosis-for-distributional-statistical-misspecification-in-industrial-time-series-forecasting.