Predictions & Matches

PitchNexus uses recorded results to estimate three possible outcomes for an upcoming match: a home win, a draw, or an away win. The Match centre also shows experimental possible scorelines when there is enough scoring history. Every percentage is an estimate, not a promise about the result.

Which match history is used?

Each team's prediction history is built from completed matches within the relevant event or competition listing. Teams are identified by their recorded IDs; similar names are not automatically combined across events or providers. A forecast can therefore use fewer games than the team's combined history elsewhere on the website.

Completed results build the ratings and performance statistics using the available dates. Matches still count when their date or kickoff time is missing, or when multiple games share a recorded time. Those schedule details remain flagged for review. The Match centre's Games used, scoring averages and recent form come from the same saved analysis as the forecast.

Records are shown as Wins – Losses – Draws (W–L–D). A record of 3–1–2 means three wins, one loss and two draws. Points per game uses three points for a win, one for a draw and zero for a loss.

What goes into the outcome forecast?

Performance adjustments receive less weight when either team has little relevant history. The signals are combined into a difference in estimated team strength. A probability formula converts that difference into home and away chances, with a separate draw estimate that increases when the teams appear evenly matched. The three probabilities add up to 100%, apart from display rounding.

How to read the percentages

Two different kinds of information on each card

How possible scorelines are estimated

The Match centre needs at least three analyzed games for each team and usable scoring and conceding averages to show score estimates. An outcome forecast may still be available when score estimates are unavailable.

1. Start with scoring averages. Average the home team's goals scored per game with the away team's goals conceded per game, and do the reverse for the away team. A small 0.05-goal floor handles zero averages.
2. Build possible scores. A Poisson distribution uses those averages to give an initial probability to each scoreline.
3. Align with the outcome forecast. Scorelines within each outcome are reweighted so that all home-winning scores together match the published home-win chance, all draws match the draw chance, and all away-winning scores match the away-win chance.

Score predictions select the most likely exact score within the favored outcome, so the predicted score agrees with the win/draw forecast. If outcomes are tied, no single score prediction is selected. Other possible scorelines are grouped by outcome; all exact-score percentages are chances for the whole match.

Home score first

Each score is a chance for the whole match. Outcome totals include other scores.

“Both teams score,” “3+ total goals” and each team's clean-sheet chance are calculated from those same score estimates. These scenarios overlap and do not add up to 100%; a 2–1 result satisfies both “Both teams score” and “3+ total goals.”

Score estimates are experimental. They have not been backtested or calibrated as an exact-score model. They do not separately adjust for bracket strength, player availability or match duration.

How the outcome models are tested

The research process compares a baseline model with versions that add home/away results, shared opponents, or both. The current production model combines both enhancements.

1. Replay historical matches in time order. Build the prediction from earlier match history, then add the result after making the prediction.
2. Compare four successive test windows. This walk-forward approach checks performance across different periods rather than relying on a single split.
3. Check both picks and probabilities. Accuracy measures how often the highest-probability outcome occurred. Brier score measures how far the probabilities were from the actual outcome; log loss particularly penalizes confident mistakes. Lower is better for both probability measures.

Historical validation does not use final standings or the finished Team Analysis scores to predict earlier games. Research results are reviewed; rerunning validation does not automatically replace the production model. Historical test performance is not a guaranteed accuracy rate for every age group, competition or future Spain update.

When forecasts are available and refreshed

New forecasts become available after collection, analysis and publication have completed successfully. Starting a collection does not immediately update the website's predictions. Until the refreshed analysis is published, the site continues to use the last published forecast.

The Match centre shows current forecasts for unplayed, scheduled matches. Completed matches and cancelled, postponed or otherwise ineligible fixtures do not show a current forecast there. An archived pre-match forecast is not currently available on completed match pages.

A forecast can also be unavailable while analytics are pending, or when the match and prediction records conflict or no longer agree on the teams or date. Missing data is not treated as a zero-percent chance.

Schedule needs verification: Predictions are paused until the current schedule is verified.

Missing dates, missing kickoff times and overlapping recorded times do not prevent a forecast when the teams and other match information are confirmed. Dates and times are never invented to fill those gaps.

What the model cannot know

It does not currently use player availability, injuries, starting lineups, training progress or weather. It only knows the match history recorded in the system; missing results and limited history can affect the estimate. Home designation may not reflect a meaningful advantage at a neutral venue.

Teams without recorded results start with neutral rating defaults, so an outcome forecast can still appear with minimal evidence. Evidence strength is not a guarantee of accuracy, and a strong favorite can lose.

Rankings and predictions answer different questions. Rankings compare overall performance within a group. Predictions estimate the outcome of a specific matchup, so a ranking position alone does not determine the favorite.

These are published analysis snapshots, not live in-game forecasts. This explanation was checked against the current outcome engine and Match centre score estimates on October 6, 2026.