Research process
BrierIQ organizes event-market analytics around calibration, model-market divergence, liquidity, market fit, catalyst context, and post-resolution learning.
Last updated: May 2026
Informational analytics only. Not financial advice. Not a recommendation to trade.
BrierIQ is designed to help users research event markets with more structure. It focuses on probabilities, calibration, market movement, and context rather than instructions or execution.
The methodology is research-oriented. BrierIQ helps organize evidence, compare probability estimates, and review historical outcomes so users can perform independent analysis.
Calibration is different from win rate. A forecast can be directionally right often while still assigning probabilities poorly. Brier score measures probability forecast quality after an outcome is known; lower Brier score is better.
Historical resolved flags can help evaluate whether probability estimates were calibrated across closed examples.
Model-market divergence compares market-implied probability with BrierIQ, model, or user probability estimates. Divergence is a research input, not a recommendation.
Divergence can disappear quickly when new information, liquidity, or market attention changes.
Apparent edge can be reduced by spreads, liquidity, fees, timing, and execution friction. BrierIQ evaluates whether a market is worth researching, not just whether a price moved.
BrierIQ may evaluate volume, notional movement, VWAP-style anchors, and other flow/context features. Momentum can be noisy, and momentum signals remain subject to readiness gates.
News and source context can help explain a market move, but news alone is not enough. Official and source integrations may expand over time as coverage is validated.
Not every market is equally modelable. Data availability, resolution clarity, liquidity, time to close, historical comparability, and adverse-selection risk all matter.
BrierIQ is designed to support criteria-triggered research alerts. User-visible alerting remains gated by validation, sample size, net-value, and readiness checks. Alerts are not recommendations or execution instructions.
Closed examples may be shown after resolution. They are not guarantees of future performance. They are used for transparency, calibration review, and post-resolution learning.
BrierIQ is in controlled beta. Some features are live, some are beta/manual/sample, and source coverage is expanding. The product should not be interpreted as fully automated across all markets.
For the current coverage breakdown, see Beta Coverage.
BrierIQ does not provide execution, direct market links, recommendations, guaranteed outcomes, or affiliation with Kalshi or any exchange.
FAQ · Beta Coverage · Terms · Privacy · Support