Public research project

Prediction Market Execution Lab

Testing whether apparent short-horizon prediction-market edge survives spread, latency, fill probability, and settlement.

Explore dashboard View repository

Read the public reports

Public sample execution attempts tracked through order, acceptance, and fill states — Public sample only. Most execution attempts do not become filled exposure.

Evidence summary

What the public workflow shows.

The findings are sample-backed diagnostics, not live-performance claims.

Apparent edge

Tick replay can identify theoretical mispricing.

It does not establish that the price is executable.

Execution funnel

Many signals fail before they become filled exposure.

Spread is only one of several failure points.

Calibration

Extreme probability buckets are less stable in the public sample.

Sparse observations make tail confidence fragile.

Profitability

The sample does not show reliable realized PnL.

Public data is anonymized and downsampled.

What the evidence shows

From signal edge to executable exposure.

These figures are generated from anonymized public samples. They are not live trading performance or a complete historical backtest.

Public sample execution attempts grouped by recorded status — **Execution states**Recorded status explains where attempts stop.

Calibration curve comparing fair probability buckets with realized outcomes — **Calibration**Tail probability confidence is fragile.

Research method

Separate theoretical edge from executable edge.

The public workflow uses anonymized samples while excluding wallets, order identifiers, private ledgers, model artifacts, and live execution systems.

Read the full methodology

01
Public sample
Anonymized candidates, executions, settlements, and ticks.
02
Fair probability
Reference price, volatility, and time-to-resolution assumptions.
03
Theoretical edge
Compare estimated fair probability with market-implied probability.
04
Execution diagnostics
Inspect spread, latency, fills, rejected orders, and settlement.
05
Public artifacts
Export reproducible reports, notebooks, figures, and dashboard views.

Explore and reproduce

Follow the evidence at the level you need.

Scope and limitations

Research demo only.

This is not financial, trading, or investment advice. The project does not claim to predict BTC, provide a profitable strategy, or operate a live trading or betting system.

Public outputs use anonymized, downsampled samples and should not be interpreted as complete historical empirical performance.

Review all limitations