Naive forecast baselines

supporting-measurementforecast

Question

Establish naive/persistence forecast baselines before any model is built.

Verdict

Published before any model exists, per the project's standing rule that a null chosen after seeing model results cannot be distinguished from a rationalization. Supplies the persistence/naive nulls spread-forecast.md is graded against.

In plain language (1 term)

This site's own explanation of the words above. The verdict itself is published exactly as the research index records it and is not rewritten here.

null: The dull baseline a model has to beat, such as assuming tomorrow looks like today. It is chosen and published before any model runs, so it cannot be picked afterwards to make a result look better.

Population: Point-in-time tradeable universe, per item, aggregated equal-weighted and notional-weighted.
Resolution: Daily-resolution forward spread targets (y_fwd_spread_dm_{k}d).
Cutoff: Development span only; sealed period not read.
Baseline: Is itself the baseline (climatology / deviation persistence) supplied to spread-forecast.md.

Limitations

Not itself a hypothesis verdict; a precondition artifact published before any model exists, so the null cannot be chosen after seeing model results.

Notes

Not itself a verdict on any hypothesis; a precondition artifact for spread-forecast.md and any future forecasting work.

Source

Citation: po-research:forecast-baselines:ae6ab6319ef113ff
Write-up: docs/research/forecast-baselines.md
Data: docs/research/forecast-baselines.json, docs/research/forecast-baselines-per-item.csv
Artifact: research-index.json v1 — sha256:cb3d976779a5cbcb4188039a8d26d5e558a2d844030a638c2f16fe5a8eb0a2fe
Build data cutoff: 2026-08-12
Issues: #39, #46