The paper audits the impact of temporal leakage on financial-news direction prediction across 49,799 articles and 16 feature-model combinations, including TF‑IDF, MiniLM, FinBERT, and fine‑tuned RoBERTa‑large / DeBERTa‑v3‑large, as well as zero/few‑shot and LoRA probes of Llama‑3 and Qwen2.5. Random train‑test splits inflate MCC scores by 1.1× to 6.5×, with larger models and richer features showing greater gains, while end‑to‑end FinBERT fine‑tuning actually increases the gap. Only the mergers and acquisitions (M&A) category shows a positive locked‑test signal under near‑temporal chronological evaluation, with the signal localized to 2024‑2025 European‑tilted M&A semantics and not transferring to a 2009‑2020 U.S. corpus.
By Chenhao Xue, Raslen Guesmi, Siwei Feng, Yucheng Gong, Jacob Xavier Sundram, Jordan Pang, Lan Wang, Julian Kaljuvee
The study investigates which inputs influence a language model’s financial forecasts by creating synthetic earnings announcements that independently reverse 20‑day momentum, commentary tone, and EPS surprise. Six instruction/chat checkpoint conversions from four model families (0.49‑7.61 B parameters) were tested with 2,016 prompts each, covering 384 primary scenarios. Results show that the contrast in primary EPS ranges from –0.19 to 86.80 percentage points, with signed tone and momentum contrasts often exceeding EPS effects; numerical diagnostic accuracy varied from 51.0 % to 95.8 %. The experiment measures observable input reliance and numerical recognition but does not establish internal reasoning mechanisms or return predictability.
By Alina Khaybullina
The paper introduces DisclosureBeta, a theory that treats large language models (LLMs) as noisy measurement channels for a firm’s latent risk characteristics, integrating this noise into the asset‑pricing error budget. It establishes identification and consistency of regime‑conditional beta loadings within a piecewise‑stationary Fama‑French five‑factor framework, provides matching lower bounds, and proposes an adaptive estimator that blends text‑based and rolling‑window approaches, improving precision when price histories are short or regime‑breaks occur. The work also outlines a pre‑registered empirical evaluation on firms with thin price histories.
By Ping Kuen Wong
arXiv:2610.29578v1 Announce Type: cross
Abstract: Using a near-census of 802,863 resolved markets listed on Polymarket between January 2023 and May 2026, and only information available at listing (th...
By Ahmed Mahrous, Roberto Di Pietro
arXiv:2609. 20193v1 Announce Type: new Abstract: Retrieval plug-ins supply a deep forecaster with information its lookback window cannot carry.
By Mert Onur Cakiroglu, Elham Buxton, Mehmet Dalkilic, Hasan Kurban
arXiv:2608. 10433v2 Announce Type: replace Abstract: Temporal reports are increasingly emitted alongside numerical forecasts and are often interpreted as statements about the computation producing those forecasts.
By Qipeng Qian, Yuntao Qian
The paper demonstrates that the outcome of a forecasting leaderboard is largely determined by the evaluator’s design choices rather than the models themselves. By fixing the data, horizon, and period, the authors varied three key evaluation decisions—unit of analysis, error pooling, and scoring metric—and showed that each can reverse or eliminate the apparent superiority of any forecasting method. The study also evaluates the practical impact of these choices on a deployed system, revealing that the selection rule captures a significant portion of the potential performance gain, and confirms the findings on an external public dataset.
By Md Rezwanul Islam, Wael Mohammed
arXiv:2607. 12248v1 Announce Type: cross Abstract: Large pretrained time-series models such as TimesFM are attractive for financial forecasting, but raw directional accuracy is a misleading scoreboard in equity markets.
By Taizhen Cheung, SA Kwon
arXiv:2605. 30363v2 Announce Type: replace-cross Abstract: Regime shifts in financial markets reorganise the joint dynamics of asset prices and macro variables, breaking any single-regime calibration.
By Mingxuan Yi, Vidal Mehra, Jing Chen, John Cartlidge
The paper introduces RLR, a stable matching repair method that reuses valid observations while respecting matching capacity and handling delayed, out‑of‑order evaluations. RLR combines sound outcome certificates with batched matching schedules to manage unequal sample deficits, achieving a sharp issue‑round bound and separate delay and evidence‑lifetime bounds. The authors also present RLR‑P, a variant that monitors issue‑order prefixes without altering the sampling plan, and demonstrate through simulations that it reduces mean capped update‑certification time compared to a width‑based baseline, albeit with higher oracle CPU cost.
By He Sun, Junyuan Mao, Wenhao Jiang, Mingjun Xiao, Shinan Liu
arXiv:2606. 24950v1 Announce Type: new Abstract: Financial decision-making is contextual: forecasting prices, valuing companies, and assessing event exposure weigh price history, accounting fundamentals, macroeconomic regime, and contemporaneous text.
By Patara Trirat, Jin Myung Kwak, Jay Heo, Heejun Lee, Sung Ju Hwang
FinalityBench is an executable benchmark that tests how agents decide on shipping, re‑capturing, refunding, or waiting when a merchant’s payment processor, ledger, ERP, and bank feed receive delayed, duplicated, dropped, or reordered messages, causing contradictory beliefs about an order. The benchmark uses a hidden canonical event log and faulted delivery streams to generate system views, scoring each episode by the merchant’s terminal economic position relative to a privileged reference. It contains 321 tasks, including 45 twin pairs where all four views are identical yet the correct disposition differs, and evaluates nine programmatic policies, revealing that a ship‑on‑first‑sign policy performs best by accuracy but worst by paired loss, while a runtime‑gated irreversible‑action policy achieves 85.4% accuracy without losing money.
By Abhishek Sharma