# When Does Retrieval Help Time-Series Forecasting?

arXiv: https://arxiv.org/abs/2609.20193 · HTML: https://mlsanity.com/paper/2609.20193/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 32 of its 41 references that are in the map.
- First posted on arXiv 29 Jul 2026.

## Builds on (most cited first, 30 of 32)

- [A Time Series is Worth 64 Words: Long-term Forecasting with Transformers](https://mlsanity.com/paper/2211.14730/) (2022)
- [Are Transformers Effective for Time Series Forecasting?](https://mlsanity.com/paper/2205.13504/) (2022)
- [TimesNet: Temporal 2D-Variation Modeling for General Time Series Analysis](https://mlsanity.com/paper/2210.02186/) (2022)
- [Chronos: Learning the Language of Time Series](https://mlsanity.com/paper/2403.07815/) (2024)
- [A decoder-only foundation model for time-series forecasting](https://mlsanity.com/paper/2310.10688/) (2023)
- [Unified Training of Universal Time Series Forecasting Transformers](https://mlsanity.com/paper/2402.02592/) (2024)
- [TimeMixer: Decomposable Multiscale Mixing for Time Series Forecasting](https://mlsanity.com/paper/2405.14616/) (2024)
- [TimeXer: Empowering Transformers for Time Series Forecasting with Exogenous Variables](https://mlsanity.com/paper/2402.19072/) (2024)
- [GIFT-Eval: A Benchmark For General Time Series Forecasting Model Evaluation](https://mlsanity.com/paper/2410.10393/) (2024)
- [Sundial: A Family of Highly Capable Time Series Foundation Models](https://mlsanity.com/paper/2502.00816/) (2025)
- [Are Language Models Actually Useful for Time Series Forecasting?](https://mlsanity.com/paper/2406.16964/) (2024)
- [CycleNet: Enhancing Time Series Forecasting through Modeling Periodic Patterns](https://mlsanity.com/paper/2409.18479/) (2024)
- [catch22: CAnonical Time-series CHaracteristics](https://mlsanity.com/paper/1901.10200/) (2019)
- [Frequency Adaptive Normalization For Non-stationary Time Series Forecasting](https://mlsanity.com/paper/2409.20371/) (2024)
- [Retrieval Augmented Time Series Forecasting](https://mlsanity.com/paper/2411.08249/) (2024)
- [TS-RAG: Retrieval-Augmented Generation based Time Series Foundation Models are Stronger Zero-Shot Forecaster](https://mlsanity.com/paper/2503.07649/) (2025)
- [Retrieval-Augmented Diffusion Models for Time Series Forecasting](https://mlsanity.com/paper/2410.18712/) (2024)
- [Scaling Law for Time Series Forecasting](https://mlsanity.com/paper/2405.15124/) (2024)
- [It's TIME: Towards the Next Generation of Time Series Forecasting Benchmarks](https://mlsanity.com/paper/2602.12147/) (2026)
- [The Effectiveness of Discretization in Forecasting: An Empirical Study on Neural Time Series Models](https://mlsanity.com/paper/2005.10111/) (2020)
- [Context parroting: A simple but tough-to-beat baseline for foundation models in scientific machine learning](https://mlsanity.com/paper/2505.11349/) (2025)
- [Channel Dependence, Limited Lookback Windows, and the Simplicity of Datasets: How Biased is Time Series Forecasting?](https://mlsanity.com/paper/2502.09683/) (2025)
- [Predicting the Future by Retrieving the Past](https://mlsanity.com/paper/2511.05859/) (2025)
- [Enhancing Multivariate Time Series Forecasting with Global Temporal Retrieval](https://mlsanity.com/paper/2602.10847/) (2026)
- [TS-Memory: Plug-and-Play Memory for Time Series Foundation Models](https://mlsanity.com/paper/2602.11550/) (2026)
- [Channel-wise Retrieval for Multivariate Time Series Forecasting](https://mlsanity.com/paper/2604.05543/) (2026)
- [Why Do Time Series Models Need Long Context Windows?](https://mlsanity.com/paper/2606.01999/) (2026)
- [Multivariate de Bruijn Graphs: A Symbolic Graph Framework for Time Series Forecasting](https://mlsanity.com/paper/2505.22768/) (2025)
- [Retrieval Mechanisms Surpass Long-Context Scaling in Time Series Forecasting](https://mlsanity.com/paper/2605.08217/) (2026)
- [MFRS: A Multi-Frequency Reference Series Approach to Scalable and Accurate Time-Series Forecasting](https://mlsanity.com/paper/2503.08328/) (2025)

## Abstract (arXiv)

Retrieval plug-ins supply a deep forecaster with information its lookback window cannot carry. Published evaluations report consistent gains, and each credits its own mechanism. We show that the benefit belongs instead to the operating point: the relation between window length S and dominant seasonal period L, an axis the standard protocol never varies. Stratifying the evaluation by that relation exposes the regime. At S=12, a simple control that repeats the last observed period beats the six standard backbones, in aggregate, on four of seven benchmarks by 8% to 44% of MSE. It beats the strongest plug-in we run on ETTm1 and matches it on ECL. It is worse by up to 25% on the three datasets whose training-split spectra lack a concentrated, shared period. A controlled synthetic sweep of horizon, period, and window shows the benefit boundary tracks the period (correlation +0.71), not the horizon (-0.23). A paired control with no phase to recover nearly erases the effect, consistent with phase starvation. Zero-shot pretraining does not escape it: a foundation model trails trained backbones by 22% to 50% on the periodic benchmarks. Within our instrument, exact lookup matches graph diffusion: the payoff is consulting the record, not the machinery on top. Two interpretable statistics, a trend test and a staleness rate, predict the sign of the per-cell benefit at 0.76 accuracy under leave-one-dataset-out evaluation, a suggestive margin over the 0.69 majority rule, where a 22-feature stack manages 0.57. We propose no new plug-in. The contribution is the regime map, the protocol that reveals it, and two statistics that screen it before deployment. Code: https://github.com/KurbanIntelligenceLab/retrieval-regime.