Urgent.News

What's breaking now, across hundreds of outlets.

Tech

Building a Cryptographically Verifiable Sports Prediction Ledger (SHA-256 + Brier Score) and Dixon-Coles xG Simulator in Go

Most online sports prediction dashboards suffer from an obvious credibility problem: hindsight editing. Predictions get quietly revised after kickoff, or losing streaks disappear from marketing banners. While building AsianScore , we wanted our football and NBA quantitative models to be publicly auditable like an engineering system rather than a black-box tip sheet. Here is the architecture we…

In the pursuit of creating transparent and auditable sports prediction models, a team developed a system built on cryptographic techniques and statistical methods using the Go programming language. By implementing SHA-256 hashing and the Dixon-Coles xG simulator, they aimed to provide a publicly-verifiable ledger of football and basketball match predictions.

The core of the architecture involves locking the model probabilities for each match within a specific time frame before kickoff. Subsequently, a SHA-256 fingerprint is generated using the match ID, kickoff timestamp, and probabilities of home, draw, and away wins. This input hash serves as a cryptographic commitment to the model's initial prediction, ensuring the forecast cannot be altered once set.

Once the match concludes, an automated worker reconciles the actual scores with the predicted probabilities to calculate the Brier Score. This strict scoring rule penalizes incorrect predictions more heavily, providing a precise measure of prediction accuracy. By tracking this metric across a live ledger containing 302 locked matches and 206 settled outcomes from various leagues, the team achieved an impressive 61.65% accuracy rate with a mean Brier score of 0.166.

The Dixon-Coles Bivariate Poisson technique addressed the underestimation of low-scoring draws by football models. By incorporating a low-score correlation factor, this method adjusts the predicted probabilities more accurately, offering a better representation of match outcomes. The implementation of this approach required minimal dependencies, allowing for easy integration into the overall system.

To ensure transparency, the team provided public endpoints that allow anyone to inspect the raw hashes, league-specific breakdowns, and interact with the simulations directly. This open approach enables users to verify the predictions' integrity and fosters trust in the model's reliability.

Rewritten by urgent.news; may contain errors — check the original.

Read the original at dev.to →

Report a problem with this story

More in Tech

I Got Tired of Noisy Terminal Output, So I Built a Tiny CLI in C++

I've been working with different languages and frameworks for years, and there's one thing that has always bothered me a little: how much unnecessary output our development tools produce.

  • Muffle is a lightweight CLI written in C++17 to filter command output
  • Users can run muffle -- npm test to get concise test results
  • Muffle logs complete output to files and preserves original exit codes

What Checking Every Scheduled Run, Not Just the Latest, Actually Caught

42 scheduled runs so far, 5 of them failed outright, and not one of those 5 ever surfaced on its own. No email, no notification, no red mark anywhere I'd naturally look.

  • 42 scheduled runs conducted, but 5 failed without alert
  • Failures went undetected until routine manual checks
  • Status check with Get-ScheduledTaskInfo caught issues

Sitemap could not be read: the four causes, and the command that names yours

Search Console's "Sitemap could not be read" is one message for four different failures. That is why the usual advice — resubmit it, wait, resubmit again — works for maybe a quarter of the people who…

  • The "Sitemap could not be read" message has four possible causes.
  • Use curl to check the sitemap's format; text/html indicates an issue.
  • Verify sitemap URLs match the host domain; mismatched loc lines cause problems.

More from Saturday 21 November →