What the work actually involves

You are handed a date and an evidence cutoff — say, a Senate race as of mid-September — and asked to write the analysis a strong practitioner would have written at that moment. That means a directional call, a probability or probability range, the polling and fundamentals you relied on, and an explicit account of the causal mechanism connecting the event to political and market outcomes. A second stream of work is evaluative: you read model-generated forecasts and grade them against your own standard, flagging where the model hindsight-leaks, misreads a house effect, treats a topline as a point estimate, or asserts a mechanism that does not survive scrutiny.

The binding constraint is point-in-time discipline. Anything you know because of what happened next is disqualifying, as is confidential polling, client-privileged risk work, or material non-public information. Reports are graded on whether the reasoning stands on the evidence available at the cutoff — a call that turned out wrong but was well-reasoned scores better than a lucky one that leaked the outcome.

What the platform screens for

  • Live judgment, not retrospective commentary. Screeners probe for calls you made before results were known, with dates, numbers, and what you got wrong.
  • Polling literacy under follow-up. Expect questions on likely-voter screens, weighting by recalled vote, house effects, mode effects, and how you treat correlated polling error across states.
  • Mechanism, not narrative. Saying an event "creates uncertainty" is not analysis; the screen wants the transmission channel and who is on the other side of the trade.
  • Calibration habits. Whether you have ever scored yourself, and what the evidence looks like.

Brand-name background — a national forecasting shop, a network decision desk, a presidential or Senate analytics team, a political-risk desk at a bank, fund, Eurasia Group, or EIU — is a strong first-pass signal, but the work sample decides. Roughly 8+ years of direct experience is the expectation, or unusually strong demonstrated performance in place of it.

Logistics

Fully remote and asynchronous, contractor engagement, with observed rates in the $150–250/hr band (stated as observed, not guaranteed; actual rate depends on assessed seniority and task type). Volume is project-driven and comes in waves around active election cycles and risk events. Most contributors work in blocks of several hours; a single report typically takes 60–120 minutes including sourcing. There is a paid or unpaid work sample before onboarding, and reports are reviewed, so expect revision requests early on.