Metraipolis › archive › dc_64985cde04bc

Forecast Accuracy Is Not Trading Profit: Evolving Small Recurrent Networks for Stock Return Prediction — arXiv:2610.07825

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.07825
License (as archived): CC0-1.0
Archived: 2026-10-07T11:42:40.427Z
SHA-256 of the archived text: d3c1b5561923fae6e97be20daff1ab45caf04f5e3e1e5ec91598c8edc7a77fc7
Stable URL of this page: https://metraipolis.com/documento/dc_64985cde04bc

How to cite

Forecast Accuracy Is Not Trading Profit: Evolving Small Recurrent Networks for Stock Return Prediction — arXiv:2610.07825 — arXiv (en), https://arxiv.org/abs/2610.07825 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_64985cde04bc (Metraipolis archive dc_64985cde04bc, sha256 d3c1b5561923fae6e97be20daff1ab45caf04f5e3e1e5ec91598c8edc7a77fc7)

BibTeX

@misc{dc64985cde04bc,
  title = {Forecast Accuracy Is Not Trading Profit: Evolving Small Recurrent Networks for Stock Return Prediction — arXiv:2610.07825},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.07825},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_64985cde04bc; sha256 d3c1b5561923fae6e97be20daff1ab45caf04f5e3e1e5ec91598c8edc7a77fc7},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.07825. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Time series forecasting models are typically compared on pointwise error, which scores a prediction in isolation from the decision it is produced for, and a lower forecast error does not imply a better decision downstream. A parallel debate asks whether modern transformer architectures forecast better than recurrent and other lightweight models. We compare linear, fixed recurrent, transformer, and mixing based architectures against recurrent networks evolved by neuroevolutionary architecture search, evaluating each on forecast accuracy and on the net return of a daily long/short strategy. All models are fit on a pooled panel, one network trained across the whole universe. Across four mid-cap portfolios and three trading years, the evolved networks rank first on both forecast accuracy and net trading performance, while the second most accurate model loses money once positions are formed and costs are charged. The advantage tracks a horizon match, since rank IC for the evolved networks rises from a one-day to a ten-day scoring horizon while every model above 300 parameters declines. They are also the cheapest end to end: a CPU-only search of 16 minutes yields 66-weight networks that predict in 10.8~$\mu$s on a Raspberry Pi Zero, against transformer baselines of up to 817,153 parameters that require GPU training.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Neural and Evolutionary Computing (cs.NE) ; Computational Engineering, Finance, and Science (cs.CE); Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.07825 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-07T11:19:24.921Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).