# Anteproof — week ending 2026-09-20

Every forecast below was logged before its outcome was knowable, and none has been edited since. Period: 2026-09-13 to 2026-09-20.

## Resolved this period (1)

Brier score **0.384** (a coin flip scores 0.250; lower is better), 0 of 1 on the right side of 50%.

| Our forecast | Outcome | Score | Question |
|---:|:---:|---:|:---|
| 38% | YES | 0.384 | 🇺🇸 Will the Fed raise interest rates in September 2026? |

## New forecasts (596)

| Our forecast | Resolves | Question |
|---:|:---|:---|
| 31% | 2026-12-14 | Will jazzband/tablib publish a release by 14 Dec 2026? |
| 7% | 2026-12-14 | Will eliben/pycparser publish a release by 14 Dec 2026? |
| 84% | 2026-12-14 | Will jupyterlab/jupyterlab publish a release by 14 Dec 2026? |
| 4% | 2026-12-14 | Will mightyiam/get-own-enumerable-property-symbols publish a release by 14 Dec 2026? |
| 10% | 2026-12-14 | Will VoidStarKat/half-rs publish a release by 14 Dec 2026? |
| 72% | 2026-12-14 | Will pybind/pybind11 publish a release by 14 Dec 2026? |
| 22% | 2026-12-14 | Will jab/bidict publish a release by 14 Dec 2026? |
| 75% | 2026-12-14 | Will twisted/twisted publish a release by 14 Dec 2026? |
| 22% | 2026-12-14 | Will GoogleChrome/workbox publish a release by 14 Dec 2026? |
| 14% | 2026-12-14 | Will gitpython-developers/smmap publish a release by 14 Dec 2026? |
| 5% | 2026-12-14 | Will jonschlinkert/is-plain-object publish a release by 14 Dec 2026? |
| 34% | 2026-12-14 | Will pillarjs/path-to-regexp publish a release by 14 Dec 2026? |
| 32% | 2026-12-14 | Will npm/npm-package-arg publish a release by 14 Dec 2026? |
| 11% | 2026-12-14 | Will shellscape/webpack-manifest-plugin publish a release by 14 Dec 2026? |
| 12% | 2026-12-14 | Will Tinche/aiofiles publish a release by 14 Dec 2026? |
| 3% | 2026-12-14 | Will jonschlinkert/isobject publish a release by 14 Dec 2026? |
| 11% | 2026-12-14 | Will eslint/eslintrc publish a release by 14 Dec 2026? |
| 7% | 2026-12-14 | Will sindresorhus/onetime publish a release by 14 Dec 2026? |
| 22% | 2026-12-14 | Will petgraph/petgraph publish a release by 14 Dec 2026? |
| 78% | 2026-12-14 | Will elastic/elasticsearch-py publish a release by 14 Dec 2026? |
| 57% | 2026-12-14 | Will getpelican/pelican publish a release by 14 Dec 2026? |
| 97% | 2026-12-14 | Will sindresorhus/p-limit publish a release by 14 Dec 2026? |
| 32% | 2026-12-14 | Will pest-parser/pest publish a release by 14 Dec 2026? |
| 12% | 2026-12-14 | Will google/brotli publish a release by 14 Dec 2026? |
| 48% | 2026-12-14 | Will python/importlib_metadata publish a release by 14 Dec 2026? |

…and 571 more, all in the public log.

## Most at risk this week

The 10 repositories with the lowest open maintenance forecasts, among source repositories of packages with at least 1,000 dependent repositories (ecosyste.ms dependent_repos_count). Lowest of the two families first; a probability is the latest sealed forecast, rounded to a whole percent.

| Package(s) | Repository | P(keeps commit pace) | P(release in 90d) | Dependents | Receipt |
|:---|:---|---:|---:|---:|:---|
| `npm:rxjs` | ReactiveX/rxjs | 3% (515+ commits) | 29% | 2,616,591 | [#271573](https://anteproof.com/questions/271573) · [#271662](https://anteproof.com/questions/271662) |
| `npm:js-yaml` | nodeca/js-yaml | 3% (166+ commits) | — | 1,075,506 | [#275704](https://anteproof.com/questions/275704) |
| `npm:builtins` | juliangruber/builtins | 3% (11+ commits) | 5% | 929,408 | [#276019](https://anteproof.com/questions/276019) · [#278557](https://anteproof.com/questions/278557) |
| `npm:through2` | rvagg/through2 | 3% (24+ commits) | 5% | 849,057 | [#275850](https://anteproof.com/questions/275850) · [#277683](https://anteproof.com/questions/277683) |
| `pypi:chardet` | chardet/chardet | 3% (130+ commits) | 7% | 284,023 | [#275834](https://anteproof.com/questions/275834) · [#277675](https://anteproof.com/questions/277675) |
| `pypi:isort` | PyCQA/isort | 3% (124+ commits) | 30% | 126,691 | [#276066](https://anteproof.com/questions/276066) · [#276665](https://anteproof.com/questions/276665) |
| `pypi:mistune` | lepture/mistune | 3% (90+ commits) | 11% | 64,256 | [#275838](https://anteproof.com/questions/275838) · [#276392](https://anteproof.com/questions/276392) |
| `pypi:asn1crypto` | wbond/asn1crypto | 3% (31+ commits) | — | 34,080 | [#276049](https://anteproof.com/questions/276049) |
| `pypi:statsmodels` | statsmodels/statsmodels | 3% (1,303+ commits) | 56% | 23,993 | [#275604](https://anteproof.com/questions/275604) · [#276094](https://anteproof.com/questions/276094) |
| `crate:num-bigint` | rust-num/num-bigint | 3% (78+ commits) | — | 20,513 | [#275754](https://anteproof.com/questions/275754) |

*A low probability is a forecast that a stated question resolves NO, not a statement that a package is unsafe or should be replaced. The commits_90d and release_90d families' resolved track record is withheld: those questions resolve by counting the archive series our source-health guard does not trust, so scoring them would not mean anything. A probability is not advice.*

Maintenance families: No track record is published for these questions. Their outcomes were counted from a third-party archive of GitHub events that we have since stopped trusting, so a score would measure that archive rather than the projects. We withheld it instead of publishing a number we cannot stand behind.

## Launch gate

Latest evaluation (2026-09-06): **PASSES**. n = 900 · Brier skill +0.112 · bias z +0.50 · Spiegelhalter Z +0.15

Against prediction markets: on 331 resolved market questions our Brier was 0.180 vs the market's 0.165 at the same moment — behind.

---
Anteproof is a product of Recoilix LLC.