Investigating simple target-covariate relationships for Chronos-2 and TabPFN-TS

TL;DR

This paper evaluates how well recent Time Series Foundation Models, Chronos-2 and TabPFN-TS, capture simple target-covariate relationships through controlled experiments.

Contribution

It introduces a systematic assessment of covariate integration in TSFMs, revealing differences in their ability to model simple dependencies.

Findings

TabPFN-TS outperforms Chronos-2 in capturing target-covariate relationships.

Chronos-2's strong benchmark performance does not imply optimal covariate modeling.

Short horizon predictions benefit more from effective covariate integration.

Abstract

Time Series Foundation Models (TSFMs) have recently achieved state-of-the-art performance, often outperforming supervised models in zero-shot settings. Recent TSFM architectures, such as Chronos-2 and TabPFN-TS, aim to integrate covariates. In this paper, we design controlled experiments based on simple target-covariate relationships to assess this integration capability. Our results show that TabPFN-TS captures these relationships more effectively than Chronos-2, especially for short horizons, suggesting that the strong benchmark performance of Chronos-2 does not automatically translate into optimal modeling of simple covariate-target dependencies.