Pro soccer Advanced

Every shot valued by how likely it was. Deepest in the marquee competitions.

• xGExpected goalsWhat an average team scores from the same chances.
• xGoTExpected goals on targetMeasured after the strike, so placement counts.
• npxGNon-penalty expected goalsPenalties stripped out.
• xGAExpected goals againstChance quality conceded. An early warning.
• xGDExpected goal differencePredicts results better than goal difference.
• xG/90Expected goals per ninety minutesA rate, so fewer matches still compare fairly.
• Adj xGOpponent adjusted expected goalsSo a soft run of fixtures does not read as form.

Two scoring rates become a full distribution over scorelines.

• λ home / λ awayModelled scoring ratesThe two inputs everything else is built from.
• ρLow score correlationNil nil and one one happen more than chance implies.
• 1X2Home, draw and away probabilityFormed before any market price is looked at.
• O/U 2.5Over under two and a half goalsThe most traded total in soccer.
• BTTSBoth teams to scoreFrom the same distribution, not a rule of thumb.
• Correct scoreScoreline distributionThe rawest thing the model produces.

One number per side, so any two teams can be compared.

• EloElo ratingRises with wins over strong teams, match by match.
• Elo ΔRating gapThe best single number preview of a fixture.
• SPI offOffensive ratingExpected goals against an average opponent.
• SPI defDefensive ratingExpected goals conceded against an average opponent.
• Matches ratedSample behind the ratingHow much to trust it yet.

How high a side defends, and which end gets played in.

• PPDAPasses allowed per defensive actionLow means aggressive pressing.
• Field tiltShare of final third possessionA better read on control than possession.
• Poss%Average possessionMostly context for the two above.
• DirectnessDirectness ratioPatient build up against a long ball side.

Two places matches are quietly decided.

• Set-piece shareShare of goals from set piecesTravels between opponents better than open play.
• GPGoals preventedSaves against the quality of shots faced.
• xGoT facedExpected goals on target facedThe baseline goals prevented is measured on.
• CS%Clean sheet rateIts own market, and a defensive form signal.

Volume on both sides, plus what happens when these styles meet.

• Sh/matchShots per matchThe simplest attacking volume measure.
• Big chancesBig chances created per matchClear openings, worth well above ordinary shots.
• Dangerous attacksDangerous attacks per matchA proxy for sustained pressure.
• Clash ΔMatchup shot differentialAgainst defences of this type, not the league.
• XI strengthPredicted lineup strengthSo a rotated side is not rated on its best eleven.
• H2H goalsHead to head goals per matchSometimes stubbornly unlike their season rates.
• H2H BTTS%Head to head both teams to score rateIn past meetings of this fixture.

The market's own opinion with the margin taken out.

• Implied %Implied probabilityBefore any margin is removed.
• No-vigMargin removed probability
• ShinShin de-vigAssumes some money is informed. Better on longshots.
• ConsensusSharp consensus priceWhere the books that move first agree.
• Line ΔLine movementOften carries more than the price itself.
• RLMReverse line movementMoving against the side taking the bets.
• DispersionBook disagreementHow widely the books are split.

The physical context, which shows up in second halves.

• Rest daysDays since last matchThe strongest of the scheduling effects.
• CongestionFixture congestion flagThree matches in eight days.
• Travel kmDistance travelledBites hardest in continental competition.
• Altitude ΔAltitude changeHome elevation against this venue's.
• ElevationVenue elevationThin air lengthens passes and shots.
• RoofClosed roof venueIndoors takes weather out entirely.

Plain language tags, so a preview can be argued.

• PressPressing tagHigh press or passive block, from the numbers.
• Build-upBuild up tagOut from the back, or long.
• Set-piece relianceSet piece dependence tagScoring that leans hard on dead balls.
• Deep blockDeep block flagDefends deep and cedes the ball.
• EdgeEdge direction and sizeWhich side our number prefers, and by how much.