Projects
Kalshi / Prediction markets

A different
read on
the odds.

An independent model of the match. A clearer view of the market.

Explore the model
Kalshi and FIFA project cover
The idea

Same match.
Different probabilities.

Compare an independent football model with the market. Surface the gap when the edge clears 3%.

  • Python
  • Poisson model
  • Kalshi
Find the edgeIllustrative example
Market
55%
62%
Model edge +7 ppAbove threshold

Move the model estimate to see the gap.

Read. Model. Compare.

Before kickoff
  1. 01

    Read the match

    Form, xG, Elo and player ratings.

  2. 02

    Model the score

    A calibrated Poisson scoreline matrix.

  3. 03

    Compare the odds

    Fair probabilities. A 3% edge threshold.

Inside the model

From match to market.

The build sequence
  1. 01

    Define the Market Scope

    Defined a narrow pre-match scope: 90-minute match predictions for 1X2, totals at 1.5/2.5/3.5/4.5, exact scorelines, and margins derived from a full scoreline matrix. Value signals are only surfaced when model edge is at least 3%, while live prediction, shootouts, and tournament-advancement markets are deferred.

  2. 02

    Build the Data Pipeline

    Built a cached sports-data pipeline for fixtures, form, H2H, player ratings, xG sources, Elo, live tournament xG, and resolved match outcomes. Market prices currently enter as manually supplied decimal odds, so this is an event/outcome pipeline with market comparison rather than a live order-book feed.

  3. 03

    Engineer Predictive Features

    Engineered football-context features including opponent-adjusted xG rates, Elo, blended player attack ratings, recent form, live in-tournament xG, rest fatigue, altitude, H2H history, and knockout conservatism. Odds are used to measure edge, not as model features, and target-exclusion checks help prevent future data leakage.

  4. 04

    Train and Calibrate

    The core model is a Dixon-Coles-corrected Poisson simulation, not a trained classifier, with calibration handled through a global scoring scalar and low-score correction. Candidate variants are compared through diagnostics and a defined promotion ladder, while recalibration is currently a manual post-matchday step.

  5. 05

    Backtest and Validate

    Built interim chronological walk-forward validation so parameters are fit only on earlier completed games and evaluated on later ones. Contract-level review reports proper scoring rules for 1X2 and totals, while execution costs, exposure limits, and formal promotion gates remain planned validation controls.

  6. 06

    Recalibrate and Roll Out

    The operating loop is designed around sealing predictions, logging results, rescoring completed matches, advancing the data epoch, and recalibrating before the next slate. New model candidates can run in shadow mode and only move toward user-facing output after passing predefined checks.

Measure what matters.

Evaluation framework
Probability Calibration

Brier

Built scoring evaluates completed snapshots with 3-way Brier score, log-loss, and raw hit rate, judging probability quality instead of only whether the favorite won. Totals evaluation also tracks goal-distribution metrics, with under-dispersion in high-scoring games identified as an active workstream.

Market Edge

ΔP

Given decimal odds, the model removes bookmaker overround and computes edge as model probability minus fair implied probability, flagging outcomes at or above a 3% value threshold. Realized edge tracking is built, but proof of market edge is still a hypothesis until more odds-backed games are scored.

Backtesting

Walk-Forward

The interim walk-forward evaluator sorts completed snapshots chronologically, fits calibration only on earlier games, and scores later games to avoid lookahead leakage. Provenance checks mark predictions as forward verified only when the target fixture is excluded from the consumed input data.

Risk Controls

Limits

Built controls include immutable prediction snapshots, input provenance, confidence-threshold gating, and a scalar cap that limits overcorrection on thin samples. Transaction-cost modeling, exposure caps, and staking recommendations are planned but intentionally held back until calibration is stronger.

Calibration and validation are ongoing. Execution costs and exposure controls remain planned.

Project results

Win / Loss Ratio

82W 22L
Hover or focus a game
79% Win Rate
Total profit$1,34653.84× deposit
Net profit$25 initial deposit
Model confidence
84%