PredictionToolkit logoPredictionToolkit
Language: English
Search tools
← Forecasting

Research tool

Polymarket Lens

Polymarket Lens is a public client-side dashboard for probability calibration, resolved-market surprises and active spreads. It groups historical probability proxies into bins and compares them with observed YES outcomes. Public charts, sample counts and source code make the calculation inspectable, but the proxy is not a clean fixed-horizon forecast dataset.

3.0/ 5Editorial ratingEditorial assessment
Open the Polymarket Lens research dashboardastuka.github.io
Available on
Web

Screenshots and interface

Product preview1 / 2

Calibration research and its plotted values.

What is Polymarket Lens?

The historical proxy subtracts oneMonthPriceChange from the final outcome price. That trailing change can include resolution, so its name does not establish a forecast made exactly one month earlier. Markets below $5,000 in volume and records without meaningful price change are excluded, changing the study sample.

Read each bin’s observed YES frequency against its probability and the diagonal reference, then check the number of supporting markets. Brier score, directional accuracy and average outcome surprise summarize the selected sample. The surprise table opens individual resolved questions, while the active table compares returned bid–ask gaps; neither table executes trades.

Who it is for

  • Readers learning how probability calibration is measured.
  • Researchers examining surprising resolved questions.
  • Developers studying a small public browser-analytics application.

Research coverage and methods

  • Calibration plot

    Compare each probability bin with its observed YES frequency and the diagonal reference line.

  • Per-bin sample counts

    Inspect how many markets support each point rather than treating all bins as equally supported.

  • Summary evaluation metrics

    Review Brier score, directional accuracy and average outcome surprise for the selected records.

  • Resolved-market surprises

    Find large gaps between reconstructed probability and the final outcome, then open the original question.

  • Active spread comparisons

    Compare the widest and tightest bid–ask gaps returned by current data.

  • Methodology and open source

    Read reconstruction, volume filters and exclusions on the page, or inspect and adapt the MIT-licensed source.

Research tasks

Interpret a calibration point

Select a bin, inspect its sample count and compare observed frequency with predicted level. Use the plotted values directly to determine the gap’s direction.

Investigate an unexpected resolution

Open a resolved-table entry and inspect the original question, timeframe and resolution.

Adapt the research method

An exact forecast horizon needs probabilities observed at known timestamps. Read the implementation before using its monthly-change proxy, and substitute timestamped historical observations when the study requires that timing.

Pricing and total cost

The public dashboard is free and its source is MIT licensed. Its client-side public-data requests require no account, API key or wallet connection.

Research inputs and access

  1. Open the dashboard, wait for public data and read Methodology before interpreting metrics.
  2. Inspect the calibration plot and bin counts, then open individual resolved markets from the surprise table.
  3. Use the repository to reproduce or change the calculation.

Inspect the MIT-licensed Polymarket Lens source

Useful links

What to check

  • The monthly-change proxy may include resolution and has no established exact forecast horizon.
  • Excluding markets below $5,000 volume and those without meaningful change alters the sample.
  • Generated overconfidence/underconfidence labels can contradict the displayed gap direction; use actual bin values.
  • A wide spread may reflect low liquidity and does not by itself demonstrate a profitable trade.

A closer look

Ratings & reviews

Editorial rating
3.0/ 5
0 user reviews
PredictionToolkit editorial

Our take on Polymarket Lens

Public charts, counts and code make this small analysis inspectable and provide a route to reproducing it. Reconstructed probabilities lack an exact forecast horizon, materially limiting what the calibration results can establish.

Assessment criteria
Task usefulness, documented scope, setup clarity, and access and cost transparency.
Assessment date
Documentation reviewed

Documentation-based assessment. How we rate

User reviews 0

The main score switches to user ratings after 5 published reviews.

Be the first to share your experience

A specific example helps the next person choose the right tool.

Related tools

Browse category
PolyBackTest historical Up/Down research with a provider backtest example
PolyBackTest
Dataset3.5Editorial

Replay short crypto rounds against recorded liquidity and keep Data API and AI-backtest entitlements separate.

  • Access Web · API

Use timestamped historical books for execution research.

Prophet Arena question input and model leaderboard entry
Prophet Arena
Research tool3.5Editorial

A live AI forecasting benchmark with resolved-event scores, comparable-event standings and inspectable research traces.

  • Access Web

Compare model forecasts and scored outcomes across public questions.

Alphascope market-research entry point and forecasting tools
Alphascope
Research tool3.0Editorial

Request an AI forecast from a link, question or image, then compare market odds and news context.

  • Access Web

Also in Forecasting and Research Tools.

Frequently asked questions about Polymarket Lens

Does Polymarket Lens need my wallet?

No. It requests public data in the browser without an account or API key.

How is the historical probability reconstructed?

It subtracts the market’s oneMonthPriceChange field from the final outcome price. This is a proxy, not a clean fixed-horizon observation.

Is the source available?

Yes. The linked GitHub repository uses the MIT license.

Does the spread table trade?

No. It displays returned bid–ask gaps for research.