Forecast

Put a number
on it.

A public forecasting desk: upme publishes its own probability on every question before asking for yours, and when a question resolves, both calibrations — ours and the community’s — are scored and published. No money, no prizes; just a track record that starts, like everything else on this site, at zero.

Questions open
6
Resolved
0
Brier scores
none yet
As of
2026-07-31

0 questions resolved · calibration publishes with the first resolution

Every question on this page was published with three things attached: a resolution rule precise enough to be argued with, a named source it will be resolved against, and our own probability — the desk number — printed before we ask for yours. A number that only appears after the outcome is not a forecast; it is a story. The desk numbers are frozen at publication, and if we ever revise one, the revision gets a dated entry in that question’s public history rather than a quiet edit.

To take the other side, log a whole number from 1 to 99. Zero and one hundred are refused by the form, because they are certainties rather than probabilities, and certainty is the one position this page does not accept from anyone — including us. Your email address identifies your forecast and does nothing else: it is stored as a keyed hash, we cannot mail it, and submitting again with the same address simply replaces your number, any time before the question’s deadline.

When a question resolves, every stored forecast and the desk number it was logged against get a Brier score, and both calibrations — ours and the community’s — are published on this page. Until then, the honest version of this desk’s record is the strip above: nothing has resolved, so nobody here has proven anything yet. The same discipline as the payout ledger, applied to our judgement instead of our money.

ny-v-kalshi-2027 By 2027-12-31

NY v. KalshiEX ends in Kalshi’s favor — dismissed, preempted or settled without a state win — by 2027-12-31

New York sued KalshiEX in July 2026 seeking $36B and a TRO; the CFTC moved to block the state. The Third Circuit has already said CFTC jurisdiction is likely exclusive; Massachusetts state courts disagree.

Our call — frozen at publication

60%

Federal appellate momentum (Third Circuit) favors preemption, but state courts have shown they will not fold, and 44 states are on record against the CFTC.

Resolution rule

Resolves YES on a docketed dismissal, a preemption ruling for Kalshi, or a settlement with no finding of illegal gambling. Resolves NO on a final state-court win or consent judgment against Kalshi. Unresolved by the deadline resolves NO.

Source: N.Y. Supreme Court docket; federal docket if removed. Reported by Bloomberg Law / CNN.

Desk history: 2026-08-02 · 60%

cftc-event-rule-2027h1 By 2027-06-30

The CFTC finalizes its event-contracts rule (amended Reg 40.11 / Appendix F) by 2027-06-30

The June 2026 NPRM proposes a three-step public-interest test that would largely permit sports-outcome contracts. Comments closed 2026-07-27.

Our call — frozen at publication

55%

The agency is motivated and the proposal is written, but a 44-state opposition letter and active litigation give it reasons to wait.

Resolution rule

Resolves YES if a final rule is published in the Federal Register on or before the deadline. A re-proposal, withdrawal, or anything short of a final rule resolves NO.

Source: Federal Register / cftc.gov press releases.

Desk history: 2026-08-02 · 55%

prop-firm-forecasting-2027 By 2027-12-31

An established prop-trading firm (FTMO, Topstep, FundedNext, The5ers, Apex or FundingPips) announces a forecasting or event-contract product by 2027-12-31

No evaluation firm offers one today. Webull added event contracts to paper trading in July 2026; brokers are converging on the category from the other side.

Our call — frozen at publication

35%

The category is loud and the firms are imitative, but their evaluation economics do not need it and compliance appetite is low.

Resolution rule

Resolves YES on an official announcement or launch by any of the six named firms of a prediction-market, event-contract or scored-forecasting product. Marketing surveys and blog posts do not count.

Source: Official firm announcements; two independent trade-press reports qualify.

Desk history: 2026-08-02 · 35%

state-blocks-sports-2027h1 By 2027-06-30

Five or more additional U.S. states obtain court orders blocking sports event contracts by 2027-06-30

As of publication: Massachusetts has an injunction (stayed on appeal), Maryland denied Kalshi relief, and roughly nine states are in crossfire litigation.

Our call — frozen at publication

45%

The states have numbers and motivation; federal preemption rulings cut the other way. Coin-flip territory, slightly under.

Resolution rule

Counts state or federal court orders (preliminary or final) in force at the deadline that block a CFTC-registered venue’s sports contracts in that state, beyond Massachusetts and Maryland. Five or more resolves YES.

Source: Court dockets, tracked via Bloomberg Law and trade press.

Desk history: 2026-08-02 · 45%

polymarket-15b-2027q1 By 2027-03-31

Polymarket closes a funding round at a $15B+ valuation by 2027-03-31

ICE has $1.6B invested; Bloomberg reported in April 2026 that Polymarket was seeking $400M at $15B.

Our call — frozen at publication

50%

The demand is reported and the category is hot, but reported talks fail as often as they close at the reported number.

Resolution rule

Resolves YES on a completed round (primary or secondary lead) at a post-money valuation of $15B or more, confirmed by the company or two independent reports.

Source: Company announcements; Bloomberg / The Information / CoinDesk.

Desk history: 2026-08-02 · 50%

upme-rulebook-v1-first By first paid challenge

upme publishes rulebook v1.2 before its first paid challenge opens

Our own commitment, listed here so it can be scored like everything else. The rulebook is pre-1.0, not in force.

Our call — frozen at publication

97%

It is our own gate and the whole brand depends on it — but nothing is ever 100%, and saying 100% would be exactly the kind of number this site exists to avoid.

Resolution rule

Resolves YES if rulebook v1.2 (or later) is live on /rulebook/ before any paid challenge can be purchased. Resolves NO if a challenge opens first — and that miss gets published like any other.

Source: This site’s own /rulebook/ and /transparency/ pages.

Desk history: 2026-08-02 · 97%

How the scoring works

A Brier score is the squared gap between a forecast and the outcome: log 70% on something that happens and you score 0.09; log 70% on something that does not and you score 0.49. Lower is better, zero is perfect, and hedging everything at 50% scores a flat 0.25 — the scoring makes that strategy visible rather than safe.

Your forecasts can be revised until a question’s deadline — resubmit with the same email and only the latest number counts. Our desk numbers cannot: they are frozen at publication, and any revision we make gets a dated public entry in the question’s desk history. When a question resolves, the outcome, the rule it resolved under, our score and the community’s distribution are published here together.

SCORED FROM THE FIRST RESOLUTION · DESK NUMBERS FROZEN · REVISIONS PUBLIC · NO MONEY, NO PRIZES