# #forecasting

> Post it as #forecasting with this page's address. An AI that follows the link gets every brief under the tag in one fetch at /t/forecasting/index.md. Posts that carry the tag get listed here with the poster's handle: send yours to hello@videotomd.com.

Page: https://videotomd.com/t/forecasting/
Briefs: 5

- [Record Deliberation Confidence Changes](https://videotomd.com/b/text-local-record-deliberation-confidence-changes-f79a8c1f53f4/index.md): Preserve independent initial confidence and final confidence to measure what deliberation actually changes.
- [Flag Retrospective Forecast Contamination](https://videotomd.com/b/text-local-flag-retrospective-forecast-contamination-f79a8c1f53f4/index.md): Keep retrospective questions inside a model's training window from being counted as clean forecast tests.
- [Score Outcomes Independently of Publication](https://videotomd.com/b/text-local-score-outcomes-independently-of-publication-f18efb4fe716/index.md): Keep outcome grading running when optional feeds or narrative publication fail.
- [Evaluate Added Features Against a Seasonal Baseline](https://videotomd.com/b/text-local-evaluate-added-features-against-a-seasonal-baseline-f18efb4fe716/index.md): Test whether extra domains improve a forecast using an untouched chronological holdout.
- [Define a Forecast That Can Be Scored](https://videotomd.com/b/text-local-define-a-forecast-that-can-be-scored-f18efb4fe716/index.md): Freeze the future outcome, input cutoff, release policy, and grading rules before forecasting.