Scorecard: the paper's registered predictions
The paper's registered predictions and where each stands on this build: the registration, the resolver and its rules, the analysis versions, and each prediction's window, check date and status in the paper's words.
Registered on 2026-10-22: OSF ↗ · doi:10.17605/OSF.IO/9Q86N
Resolved by resolve.py, rules osf-rules-1, read on 2026-10-22: resolve.py · SHA256SUMS
Analysis versions: methodology-versions.json · claim-status.json
| # | prediction | set | p | window from | check date | status |
|---|---|---|---|---|---|---|
| F1 | The council stays below two independent voices at one day | A | 0.50 | 2026-11-02 | 2026-11-30 | not yet due |
| F2 | Market co-movement does not explain the likeness | A | 0.50 | 2026-11-02 | 2026-12-31 | not yet due |
| F3 | Errors are shared too a field to be published: N_eff of the error indicators, computed by the filed resolver | B | 0.74 | 2026-11-02 | 2026-12-31 | not yet due |
| F4 | Models are less sharp than the crowd | B | 0.40 | 2026-11-02 | 2026-12-31 | not yet due |
| F5 | Shared links mean shared calls | A | 0.56 | 2026-11-02 | 2026-12-31 | not yet due |
| F6 | All wrong together | B | 0.92 | 2026-11-02 | 2026-12-31 | not yet due |
| F7 | A right minority is more common than a unanimous error | A | 0.63 | 2026-11-02 | 2026-12-31 | not yet due |
| F8 | The table fills on schedule | A | 0.88 | 2026-10-19 | 2026-11-30 | not yet due |
| F9 | Independence is greater at the month than at the day | B | 0.30 | 2026-11-02 | 2026-12-31 | not yet due |
| F10 | Averaging beats following the leader | B | 0.64 | 2026-11-02 | 2026-12-31 | not yet due |
| F11 | No single model shows a detected daily advantage | B | 0.79 | 2026-11-02 | 2026-12-31 | not yet due |
| F12 | The crowd beats the council | B | 0.83 | 2026-11-02 | 2026-12-31 | not yet due |
| F13 | The combination beats the crowd | B | 0.30 | 2026-11-02 | 2026-12-31 | not yet due |
| F14 | Models are overconfident | B | 0.80 | 2026-11-02 | 2026-12-31 | not yet due |
| F15 | The council does not beat a simple rule by a clear margin | B | 0.80 | 2026-11-02 | 2026-12-31 | not yet due |
| F16 | No horizon carries a detected signal over a season | B | 0.63 | 2026-11-02 | 2027-04-30 | not yet due |
| F17 | Relative to climatology, the day is the hardest horizon | B | 0.37 | 2026-11-02 | 2027-04-30 | not yet due |
| F18 | Country is a small factor | B | 0.80 | 2026-11-02 | 2026-12-31 | not yet due |
| F19 | The US side scores better over a season | B | 0.57 | 2026-11-02 | 2027-04-30 | not yet due |
| F20 | The same news means less dispute | A | 0.70 | 2026-11-02 | 2026-11-30 | not yet due |
| F21 | Search adds little accuracy unresolved on the files of 21 October 2026: every Closed Book call was asked after the open and has no outcome from its own reference; it resolves once closed calls sealed before the common reference, or each closed call's own reference price, close and outcome, are published | B | 0.72 | 2026-11-02 | 2026-12-31 | not yet due |
| F22 | Discussion finds real risks | A | 0.30 | 2026-11-02 | 2026-12-31 | not yet due |
| F23 | Heeding the critique pays | A | 0.42 | 2026-11-02 | 2026-12-31 | not yet due |
| F24 | A model disagrees with itself a lot | A | 0.60 | 2026-11-01 | 2026-11-30 | not yet due |
| F25 | Leaders do not last | B | 0.62 | 2026-11-02 | 2027-04-30 | not yet due |
| F26 | The index is hard to beat | B | 0.60 | 2026-11-02 | 2027-04-30 | not yet due |
| F27 | Good forecasters are not the best traders | B | 0.83 | 2026-11-02 | 2027-04-30 | not yet due |
| F28 | Access to the models before the first forecast does not improve readers' Brier score by 0.01 or more | A | 0.46 | 2026-11-02 | 2026-12-31 | not yet due |
| F29 | Access moves readers toward the council unresolved on the files of 21 October 2026: the readers' first forecasts of the experiment are not published as rows (a field to be published and documented in arena-dataset.json) | B | 0.41 | 2026-11-02 | 2026-12-31 | not yet due |
| F30 | Style gives the model away needs the addendum to the registration naming the fixed guesser (filed before 2 November 2026) and the answer lines of the blind season with the post-season reveal | B | 0.46 | 2026-11-02 | 2027-04-30 | not yet due |
Cite this data
The data are open under CC BY 4.0. Cite the data release; the registration and the paper are listed with it.
- Data release: AI Council · Markets: open league files as of 22 October 2026 (immutable release for "Sealed Before the Open") · Chernets, Vadym (ORCID) · doi:10.5281/zenodo.23552700
- Registration of the predictions: Sealed Before the Open: registered primary endpoint and 30 dated predictions on AI forecasts of financial markets, 2026 to 2027 · OSF · doi:10.17605/OSF.IO/9Q86N
- Paper: Sealed Before the Open: A Verifiable Standard for Measuring AI Forecasts of Financial Markets · paper-revisions.json
CITATION.cff
Attribution line:
Chernets, Vadym (2026). AI Council · Markets: open league files as of 22 October 2026 (immutable release for "Sealed Before the Open") [Data set]. Zenodo. https://doi.org/10.5281/zenodo.23552700 · CC BY 4.0
BibTeX
@dataset{chernets_2026_23552700,
author = {Chernets, Vadym},
title = {AI Council · Markets: open league files as of 22 October 2026 (immutable release for "Sealed Before the Open")},
year = {2026},
publisher = {Zenodo},
doi = {10.5281/zenodo.23552700},
url = {https://doi.org/10.5281/zenodo.23552700},
note = {CC BY 4.0}
}
@misc{chernets_2026_osf_9q86n,
author = {Chernets, Vadym},
title = {Sealed Before the Open: registered primary endpoint and 30 dated predictions on AI forecasts of financial markets, 2026 to 2027},
year = {2026},
publisher = {OSF Registries},
doi = {10.17605/OSF.IO/9Q86N},
url = {https://osf.io/9q86n/}
}