Mispricd
Menu
DataEvery live figure on one page MethodEvery rule, source and correction

Markets · What the markets show now

BoardEvery open Kalshi temperature bracket, read hourly GamesKalshi game markets: the price and the numbers, hourly Whale finderPolymarket weather wallets, ranked from their fills, weekly

Record · What held up once it settled

ScorecardThe settled record, week by week CalibrationDoes a 20c bracket pay one time in five? CitiesEach city's record: ruled out, the leader, calibration, when the day is decided DaysThe daily box score: every settled day, city by city CallsDated calls with a probability, graded on their date ReceiptsThe calls, the settled weeks and the retired desk The deskEvery desk's lights, from the event log, live

Research · What we found when we looked hard

StudiesFive questions, called before they were measured PiecesLong reads, decoded from the fills and the tape SymbolsGematria, graded on 22,990 games

Terms Privacy X Owner sign-in

Graded calls

What we said would happen, and whether it did.

Every call is a dated, falsifiable statement with a probability attached, a rule that decides it, and a date on which it is graded from the data. Hits and misses both stay on this page. The score at the top is the Brier score of everything graded so far: 0 is perfect, 0.25 is a coin flip.

Hypothetical results · the calls marked hypothetical, their claims and their resultsThese results are based on simulated or hypothetical performance results that have certain inherent limitations. Unlike the results shown in an actual performance record, these results do not represent actual trading. Also, because these trades have not actually been executed, these results may have under- or over-compensated for the impact, if any, of certain market factors, such as lack of liquidity. Simulated or hypothetical trading programs in general are also subject to the fact that they are designed with the benefit of hindsight. No representation is being made that any account will or is likely to achieve profits or losses similar to these being shown.
The calls
callthe claimwe saygrades onstatus
loading the calls…
How grading works

Each call names the file and the command that decides it, from the same tools that produced the pieces. On the grading date the command is run on the fresh export, or on the permanent tables it reads (the studies, C-007 on), the result is pasted in, and the tag flips to hit or miss. A call is never edited after it is posted; if the rule turns out to be ambiguous, it is graded against us. The probability is ours, stated before the fact, and the Brier score above is the mean of (probability − outcome)² over graded calls. Each grading's full printout, ending with the code's commit and, where it read an export, that file's checksum, is published as a plain-text receipt and linked from its call: a graded call's result links the receipt of its grading, and each call posted from 2026-10-16 on also links the receipt behind the figures it quotes before the call.