{"slug": "tw3cast-a-frozen-router-of-lightly-fine-tuned-foundation-models-for-time-series", "title": "TW3Cast: A Frozen Router of Lightly Fine-Tuned Foundation Models for Time-Series Forecasting on GIFT-Eval, Selected Entirely on the Training Split", "summary": "TW3Cast, a time-series forecasting system with no agent or language model, ranked 3rd of 130 entries on the GIFT-Eval benchmark by mean MASE rank as of 2026-09-14, behind two agentic-category entries. TW3Cast uses a routing table computed once on the training split and frozen, serving one of four modes per each of 97 dataset, frequency and horizon configurations, with experts drawn from lightly fine-tuned public foundation models Chronos-2, TiRex and Toto. The full router reaches a mean MASE rank of 19.4, versus 33.8 for the best base model served alone and 38.0 for the tournament served on every configuration, and the routing table, expert index, pinned base-model revisions, submitted score file and dated public-score snapshot are released so every leaderboard number regenerates by one script.", "body_md": "arXiv:2609.28506v1 Announce Type: new \nAbstract: TW3Cast is a time-series forecasting system that reaches position 3 of 130 entries on the GIFT-Eval benchmark by mean MASE rank, as of 2026-09-14. The two entries above it belong to the leaderboard's agentic category, multi-step systems that use agents or language models to reason about, generate or select forecasts. TW3Cast runs no agent and no language model. Its selection is a table computed once on the training split and then frozen, and its experts are public foundation models lightly fine-tuned on those training splits. For each of the 97 dataset, frequency and horizon configurations, the table serves one of four modes: a specialist, which is a LoRA or full fine-tune of Chronos-2, TiRex or Toto whose training data was cleaned and enriched by explicit rules; a quantile blend that contains a specialist; a blend of base models; or a selection tournament played on a backtest carved from the training split. Every decision in the table was taken on that backtest. A specialist is admitted the moment it beats the tournament there, so a candidate costs a few megabytes and minutes of GPU time, and a failed candidate changes nothing. Three guarded mechanisms protect the selection from its own biases: a dual accuracy and calibration criterion, an asymmetric margin against candidates that saw the series during training, and conservative per-window gates. The selection rules themselves were chosen inside a temporal meta-backtest. The best base model served alone reaches a mean MASE rank of 33.8, the tournament served on every configuration reaches 38.0, and the full router reaches 19.4. The routing table, the expert index, the pinned base-model revisions, the submitted score file and the dated snapshot of the public scores are released, and every leaderboard number in this paper regenerates from them by one script.", "url": "https://wpnews.pro/news/tw3cast-a-frozen-router-of-lightly-fine-tuned-foundation-models-for-time-series", "canonical_source": "https://arxiv.org/abs/2609.28506", "published_at": "2026-09-25 04:00:00+00:00", "updated_at": "2026-09-25 04:29:07.625766+00:00", "lang": "en", "topics": ["machine-learning", "ai-research"], "entities": ["TW3Cast", "GIFT-Eval", "Chronos-2", "TiRex", "Toto"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/tw3cast-a-frozen-router-of-lightly-fine-tuned-foundation-models-for-time-series", "markdown": "https://wpnews.pro/news/tw3cast-a-frozen-router-of-lightly-fine-tuned-foundation-models-for-time-series.md", "text": "https://wpnews.pro/news/tw3cast-a-frozen-router-of-lightly-fine-tuned-foundation-models-for-time-series.txt", "jsonld": "https://wpnews.pro/news/tw3cast-a-frozen-router-of-lightly-fine-tuned-foundation-models-for-time-series.jsonld"}}