Dated, probabilistic forecasts on Japan's fiscal and monetary institutions, geopolitics, and AI / industrial strategy — published before resolution and resolved against predeclared criteria. No silent rewrite. No hidden misses.
No hindsight commentary. Each forecast is published before its resolution deadline, with its forecast probability, resolution criteria, and reasoning fixed at publication — so the reasoning itself can be judged, not only the outcome.
No silent rewrite. Transparent correction. The forecast proposition, published probability, resolution deadline, and resolution criteria are not changed after publication. Material corrections — factual errors, misdescriptions, broken references — are added as dated Correction notes, additive to the original record.
Identify the behavioral principles of the actor in question — a ministry, a cabinet, a company. Institutions are more predictable than headlines suggest.
Build a falsifiable forecast with an explicit rationale, a forecast probability, predeclared resolution criteria, and a resolution deadline. Timing is the key variable: when something accelerates matters more than why someone feels a certain way.
When the deadline passes, the verdict is announced in public — including the calls that went wrong, and the ones that went right for the wrong reasons.
Each forecast has a proposition and predeclared resolution criteria. When the resolution deadline passes, the forecast is resolved into one Verdict. Causal reasoning is recorded separately as a Mechanism note, so that the resolution of a proposition and the quality of its reasoning are not conflated.
The outcome matches the proposition under the predeclared resolution criteria.
The outcome does not match the proposition. The miss is recorded in full, with a note on where the reasoning broke.
The predeclared resolution criteria cannot be applied cleanly to the outcome — for example, the relevant event did not occur in a measurable form by the deadline. The record is retained; the forecast is not counted as a Hit or a Miss.
Mechanism is recorded on a separate axis — ALIGNED, DIVERGED, or UNRESOLVED — describing whether the observed outcome arose through the causal chain stated in the forecast. A Hit with a DIVERGED Mechanism is still a Hit, and the divergence is documented as part of the record.
Protocol note. The vocabulary above is Protocol v1.1, in force from Forecast No. 002. Forecast No. 001 was resolved under Protocol v1.0, whose verdict vocabulary was Hit / Miss / Lucky; its record is preserved unchanged.
A judgmental 55% forecast on the Bank of Japan's target for the uncollateralized overnight call rate at the end of the resolution day. The base case is around 1.25%; the HIT / MISS boundary is 1.00%. Predeclared resolution criteria and mechanism hypothesis are recorded in full on the forecast page.
The guidelines (概算要求基準) are expected in late July 2026. The forecast: the Ministry of Finance gatekeeps access to the new investment envelope at the entry stage, rather than leaving it open-ended. Full rationale is recorded on the Japanese-language ledger, fixed at publication.
Editorial Correction — 22 Aug 2026: Social posts published on 3 August 2026 misdescribed the underlying proposition. The proposition, published probability, deadline, and verdict are unchanged. Read the correction →
Budget request guidelines, supplementary budgets, MOF behavior, and the institutional choreography behind Japan's fiscal year — read as a forecastable system, not a news cycle.
Japan's position between Washington and Beijing: alliance management, energy security, and the timing of pressure — when things accelerate, and what that acceleration reveals.
Capital flows, sovereign AI policy, and the strategic entry of global AI players into Japan — tracked at the level of institutions and incentives rather than product launches.
| Format | Frequency | What it does |
|---|---|---|
| 予測FORECAST | Monthly | A dated, probability-tagged claim with its full rationale and predeclared resolution criteria, published before resolution. |
| 判定VERDICT | Per deadline | HIT, MISS, or VOID — recorded when each forecast's resolution deadline passes (Protocol v1.1). Forecast No. 001 was resolved under v1.0 as Hit / Miss / Lucky. |
| 道標SIGNPOST | Monthly | A preview of the coming month's policy events and what to watch for. |
| 台帳照査LEDGER REVIEW | Quarterly | Self-assessment across the whole record: resolution rate, calibration, and lessons drawn from the ledger. |
The website is the public ledger. Each other channel has a defined role. If the channels ever disagree, the ledger — this site — is authoritative.
The public record. Every forecast in full: proposition, forecast probability, resolution deadline, and predeclared resolution criteria — fixed at publication.
Long-form Japanese record of each forecast and its resolution: original rationale, primary sources, and post-resolution notes.
The English-language digest for international decision-makers: what was published, what happened, and why it matters for Japan exposure.
Forecast declarations when they are made, and resolutions when deadlines pass — plus primary-source links as evidence surfaces.
Nihonbashi Forecast is a public forecasting project led by Akira Hara. Nihonbashi is the historic starting point of the Edo-period Five Routes and the site of Japan's national road-origin marker — a fixed zero from which every distance is measured.
The premise is simple: analysis of Japan is abundant, but publicly resolved analysis is rare. This site measures every claim from a fixed zero point — a public, dated, falsifiable record — so that readers can judge the forecaster by the same standard as the forecast.
予測は、真剣な遊びだ。
Forecasting is serious play. The record is kept because it is honest,
and kept in public because that is the point.