Pacy
About Pacy
Who runs Pacy, how our live scores are checked, how our predictions are made and graded, and how every article is reviewed before it is published.
Pacy is an independent sports desk that publishes live scores, results and match predictions, as accurate and as fast as we can make them. It is run by a small team, not a content farm. Results are reconciled across the sources we collect before they are final. Every prediction is scored once the match has been played. No article goes live until an editor has approved it.
Seven sports
Football, US soccer, basketball, American football, ice hockey, baseball and tennis, from top leagues to lower divisions.
Live, every few seconds
Live matches are refreshed continuously. If the data stops arriving, the page says so instead of showing an old score as live.
Predictions we grade
Every pre-match estimate is kept and scored against the final result. The model is tuned against that record every night.
Editor-approved writing
Previews, recaps and news are checked for accuracy and credited sources. Nothing goes live without a human's approval.
What Pacy is
Pacy started as a scoreboard built to answer one question faster than anywhere else: what is happening in this match, right now? That is still the centre of it. The score is the largest thing on every page. The state of play is written in words rather than shown only as a colour. Every time is shown in one stated time zone, so two kick-offs can be compared without doing the arithmetic.
Around the scoreboard we have added the things a fan looks up next: fixtures and results by competition, team and date; a page for every match that stays up after it finishes; pre-match win probabilities; and written previews, recaps and news.
How a result reaches the page
- Collected. Fixtures, live scores and match events are gathered continuously from established sports data sources. Live matches are re-read every few seconds, upcoming ones on a slower cycle.
- Matched. A match reported by several sources becomes one record, not several. Teams and competitions are matched through a maintained catalogue of their names and IDs, not by guessing from similar spellings. That stops two clubs with similar names being merged.
- Checked. When accounts disagree, we show the one we can corroborate rather than the one that arrived first. A source that proves unreliable is dropped instead of patched around.
- Settled. A finished match is settled and left alone. Postponed and cancelled games are labelled as such, not left looking like upcoming ones.
- Corrected in the open. If a score turns out to be wrong, it is fixed and marked as corrected. See the corrections policy.
How our predictions work
Pacy's win probabilities come from a rating model. Every team carries a rating within its competition, and the rating moves after each result by how surprising that result was. The difference between two ratings, plus home advantage, gives the chance of a home win, a draw (where draws exist) and an away win. How fast ratings react is tuned per sport. A baseball rating moves more slowly than a football one, because a single baseball game says less about a team.
The model is refitted every night on recent results. We found that recent form predicts better than a long history, so old seasons are not allowed to anchor a team's rating. A refitted model is used only if it beats a simple baseline on matches it has not seen. Once a sport has enough graded games, its probabilities are also calibrated against that record. The aim is that a 60% chance really means the favourite wins about six times in ten.
Once a match is played, every prediction is scored against the result. Only the last estimate before kick-off counts, so a forecast updated many times is still graded once.
A probability is an estimate, not a promise, and it is labelled as one wherever it appears. Pacy does not give betting tips, does not publish odds as a recommendation and takes no money from bookmakers. The details are in the data methodology.
How our articles are made
- Built on facts we hold. Previews, recaps and round-ups start from Pacy's own match data and model estimates. News about players, transfers and injuries is taken from credible news organisations and official league and club sites, and each is named and linked.
- Drafted with AI assistance. We use AI tools to prepare first drafts. We say so here because readers deserve to know how the writing is produced.
- Checked before anyone sees it. Each draft is checked for factual grounding and for originality. A draft that repeats another outlet's wording, or makes a claim not backed by our data or a named source, is rejected.
- Approved by an editor. Nothing is published automatically. An editor reads every article in the review queue and decides whether it goes live. What they reject stays unpublished.
- Kept honest after publication. Errors are corrected with a dated note on the article, never by quietly editing or deleting it. See the sourcing policy and editorial standards.
What Pacy does not do
We do not invent facts to fill a gap. If we do not have a lineup, a venue or a timeline, the page leaves it out or says so. We do not rewrite other outlets' articles, sell coverage, or take payment to feature a team, a competition or a bookmaker.
How Pacy is funded
Pacy is independent and not paid for by advertising placed next to its scores. Analytics are first-party and only run if you agree to them. They carry no advertising identifiers, and if you decline you are not tracked and lose nothing. See the privacy page.
Talk to us
A person reads everything sent to [email protected]: corrections, tips, data problems, or questions about how we work. The contact page has the details.