Methodology

What GoalPost is

GoalPost is an independent football discovery and intelligence platform. The scores, fixtures, and statistics you see are sourced from a licensed sports-data provider. Everything described on this page — GPR, GPS, match predictions, and Fair Market Value (FMV) — is built on top of that raw data by GoalPost itself. None of it is licensed from a bookmaker, a scouting agency, or another ratings provider, and none of it is sourced from betting markets as a primary input.

GPR — GoalPost Power Rating

GPR is GoalPost’s composite strength rating for a club or national team, expressed on a 1–99 scale. It blends a team’s underlying attacking and defensive output with its recent form, and applies a strength-of-schedule adjustment so a rating earned against weaker opposition is weighted differently from one earned against strong opposition. Teams with a very small sample of matches (early in a season, or a newly promoted side) are damped toward a more conservative estimate rather than allowed to swing on a handful of results.

You’ll see GPR on club and national-team pages, in the Team Screener, and on the Desk as Team Power. On the home page it is the Team Power Index (TPI). It is one number meant to answer “how good is this team right now” — not a prediction of any single match on its own.

GPS — GoalPost Player Score

GPS is GoalPost’s composite rating for an individual player, also expressed on a 1–99 scale. It’s built from position-relative percentiles: a player’s underlying actions (progression, chance creation, defensive work, discipline, availability, and more, depending on role) are measured against other players in comparable positions, not against the entire league at once. That percentile profile is then combined into a single score so players at different positions can still be compared on one scale.

GPS drives the Player Screener and appears on player pages. It is the season-long measure; one game is scored by the match rating below. A 7-day movement indicator (GPS Δ) shows whether a player is trending up or down, not just where they sit today.

Match ratings

Every player who plays gets a GoalPost match rating from 1 to 10. It starts from the match rating in our data provider’s feed and adjusts it for what matters at his position: each of his key numbers against the others at his position in the same match (passing for a midfielder, saves for a keeper), a clean sheet for keepers and defenders, a hand in the winning goal, a full 90 minutes in a knockout, title or relegation game, and deductions for a red card, an own goal, or a penalty missed or conceded. Gains shrink as a rating nears 10, so a 10 stays nearly out of reach.

The Player of the Match is the highest match rating in the game, 6.0 or more. The GoalPost XI (Team of the Week) is a round’s highest ratings in the shape those players fill.

Manager ratings

A manager’s match rating, also 1 to 10, weighs the result against what was expected before kickoff (30%), goals that came soon after his changes and how early he made them (20%), dead balls: penalties scored and corners and free kicks won (15%), goals conceded against the chances his side gave up (15%), rotation (10%) and discipline (10%). Manager of the Month is the best average in a league over three or more matches.

The Tactics Board

A match page’s Board tab breaks the game into six parts: how the sides lined up, who had the ball, where the threat came from, how they defended, how it turned, and the managers. It is built from the match record (the team sheets, every event’s minute, team and player statistics) and from each side’s own earlier matches. Our data provider gives no pitch coordinates, so nothing on the board shows where on the pitch anything happened.

• Shapes. Players stand in their slot on the team sheet, in the formation it lists. A substitute steps into the slot of the player he replaced.
• “Usually.” Each number is set against the side’s average over its previous ten competitive matches in GoalPost’s record, leaving out any match the record has no figure for. Club friendlies, pre-season tournaments and super cups don’t count, as GoalPost’s ratings leave them out; a national team’s friendlies do. Counts are per 90 minutes, so a match that went to extra time is scaled to 90. Shares (possession, pass completion, duels won) move in points. A line of text quotes a usual only from five of those matches, or all ten for a national team, whose matches span months against very uneven opposition, and it always says how many.
• By unit and by side of the team sheet. Defenders, midfielders and forwards are grouped by the position the team sheet lists for each player. A row’s outermost slots, in a row of three or more, are its left and right. That is where a player was listed, not where he played. Both need players’ own statistics, which the record holds for most recent matches but not older ones.
• Pressing. Opponent passes per tackle, interception or foul, across the whole pitch. The usual pressing measure counts only the opponent’s half; without coordinates GoalPost can’t split the pitch, so the label says “whole pitch.”
• How it turned. Each side’s chances minute by minute on the score and the clock, from GoalPost’s pre-match expected goals, with red cards. Unlike the live bar on a match page, it doesn’t weigh each side’s shots as they came, so it reads how the score moved against the pre-match picture. Shots by ten-minute spell come from GoalPost’s live poll, for matches it polled live. Expected goals in the threat part are the data provider’s, where it gives them.
• The lines. Every sentence the board writes is a template filled from its own numbers. Nothing on it is written by a model.

Match predictions

Pre-match win/draw/loss percentages come from a statistical scoreline model (a Poisson-based approach with a low-score correction, following the well-established Dixon–Coles method from the football analytics literature), not from copying a bookmaker’s odds. The model is fed by:

• each team’s season-long attacking and defensive rates, adjusted for venue (home/away);
• recent form, blended in more heavily when the in-competition sample is still thin;
• head-to-head history between the two sides, weighted by how much recent data exists;
• a downward adjustment for confirmed injuries or suspensions to key players; and
• when an outside model publishes a full win/draw/loss forecast of its own, a smaller blended weight toward it, used as an ensemble anchor rather than the primary signal (its coarse placeholder splits are left out).

The output is a full probability split (home win / draw / away win) plus an implied scoreline distribution, shown before kickoff on covered fixtures. Once the match is live, the win-probability bar updates using the score, the clock, each side’s live shot and expected-goals rate, and red cards shown, against that pre-match baseline — it does not explicitly model tactical substitutions or in-game management beyond what a team’s own live scoring rate already reflects.

Wherever GoalPost shows the betting market beside its forecast — a match’s Market Check, the matches page, the Desk and the Ledger — the market is the bookmaker’s pre-match win/draw/win prices with the bookmaker’s margin taken out in proportion, so the three chances add up to 100% like GoalPost’s own. Turned straight into chances, a bookmaker’s prices add up to more than 100% (that surplus is its margin), which would overstate every outcome. A price taken after kickoff isn’t used. The market is the benchmark GoalPost is measured against, not a pick to beat.

Predictions, GPR, GPS, and FMV are all statistical estimates for informational and entertainment purposes. None of them are betting advice, financial advice, or a guarantee of any outcome — see our Terms of Service for the full disclaimer.

FMV — Fair Market Value (GoalPost Value)

FMV is GoalPost’s own estimate of what a player would be worth in today’s transfer market. It is not an official price, an agent’s asking figure, or a confirmed transfer fee — it’s a model estimate. It starts from the player’s GPS rating; adds premiums for rarity at his position, for the deep-lying midfield roles the market under-prices, and for recent major individual honors; discounts minutes he doesn’t play when fit and output built in weaker leagues; and turns the result into euros on a curve by age and role. Player pages and GoalPost cards show it as GoalPost Value, with its range and the terms that moved that specific number, not just the headline.

How we grade ourselves — The Ledger

Estimates are only useful if they’re checked. Every scored prediction is compared against the actual result after the final whistle and logged automatically — nothing is scored by hand, and nothing is removed from the record after the fact. As real transfers land with reported fees, they’re compared against what GoalPost’s FMV said at the time, on the same public page.

Editorial commentary

Match previews, post-match recaps, and milestone storylines on GoalPost are grounded in GoalPost’s own statistical data (the actual final score, the actual standings movement, recent form) rather than written free-form. Points are always derived from the real score rather than invented, and head-to-head references are phrased as “on record” or “in the last N meetings” rather than implying a complete all-time count, since our historical data only goes back so far.

Limitations, honestly

• GPR, GPS, and predictions are statistical estimates. Upsets happen; models don’t predict them.
• Live in-play win probability starts from the pre-match model and updates with the score, the clock, each side’s live shot and expected-goals rate, and red cards shown — it does not explicitly model in-game tactical adjustments beyond what a team’s own live scoring rate already reflects.
• Coverage depth varies by competition. Top-tier leagues and major international competitions get full statistical coverage; smaller competitions get a lighter dataset, which can mean a thinner sample behind any given rating.
• FMV is a model estimate, not a market-confirmed value — treat it the same way you’d treat any single analyst’s valuation, alongside other sources, not as the last word.

Questions about how a specific number was calculated? See The Ledger for the accuracy record, or read our Corrections Policy if you think something is wrong.