Soccer predictions: actual data vs predicted data

Slate 20260918 · 2 match(es)

How to read this report. Each match shows the model's headline prediction, the stage-by-stage lambda (goal-rate) build, both teams' standing/form inputs, and every starter's projected line next to their real season baseline. A moderate / large deviation badge flags a projection that strays far from the player's season form — a quick way to spot when the inputs look right but a projection does not. When the real result is known an actual-vs-predicted banner appears.

All matches overview

Match	Score	Winner	H / D / A	O2.5	BTTS	Conf	λ H/A	Actual	Result
Brentford vs Chelsea Premier League	1-2	Chelsea	37.4% / 22.2% / 40.4%	69.0%	69.1%	0.56	1.74 / 1.82	-	pending
Espanyol vs Elche LaLiga	1-0	Espanyol	44.8% / 27.4% / 27.8%	42.9%	47.9%	0.58	1.37 / 1.02	-	pending

Match detail

Brentford vs Chelsea

Premier League · 20260918 · 4-2-3-1 vs 4-2-3-1

Predicted 1-2Winner: ChelseaConfidence 0.56λ home/away 1.74 / 1.82Most-likely P 8.2%

Actual vs predicted: result not available yet (upcoming fixture).

XGBoost Reasoning & FBref Dataset Tracing

Antagonist + XGBoost match analysis for Brentford vs Chelsea. Player statistics regressed using position-specific XGBoost models (G, D, M, F) with 2026-2027 FBref priors. Micro-duels resolved across 4-2-3-1 vs 4-2-3-1 (duel lambdas: Home 1.778, Away 1.680). Final calibrated Poisson lambdas: Home 1.742, Away 1.816 (most likely score 1-2).

Home Starters 9/11 XGBoost (10 FBref) Away Starters 10/11 XGBoost (10 FBref) Bench Modeled H: 7 / A: 7 Duel λ (H/A) 1.78 / 1.68 Final λ (H/A) 1.74 / 1.82

FBref Team Priors: Brentford (Gls/90: 2.0, GA/90: 2.0 → E[λ]: 3.5) vs Chelsea (Gls/90: 3.5, GA/90: 3.5 → E[λ]: 3.5) | Blend Weight: 0.192

Modeling Pipeline Stages (5 layers)

1. XGBoost player regressions for volume & event rates with market-value quality scaling
2. Multi-source prior blending (FBref 2026-2027 tables + Sofascore season history)
3. Antagonist formation duel matrix resolving 1v1 and zonal matchups
4. Team-level prior calibration (FBref possession and attack/defense efficiencies)
5. Bivariate Poisson matrix with Dixon-Coles low-score correction

Standing & form context

Metric	Brentford	Chelsea
League position	5.00	4.00
Points per game	1.67	2.00
Attack index	2.08	1.67
Defense (concede) index	1.46	0.83
Form multiplier	0.98	1.02
Standing λ multiplier	1.59	1.15

Lambda decomposition (how the goal rates were built)

Stage	Brentford λ	Chelsea λ
FBref Squad Prior E[λ] (weight 0.192)	3.500	3.500
Pre-mismatch	1.778	1.680
Pre-competitive-floor	1.742	1.816
Pre-calibration (Micro-duels)	1.778	1.680
Final λ	1.742	1.816

competition calibration scale 1.0competition observed total goals 0.0quality gap applied Truemismatch score pressure applied Truecompetitive underdog floor applied Falsestage discount 1.0

Scoreline distribution & markets

Top 5 scorelines	P
1-2	8.2%
1-1	9.1%
2-1	7.8%
2-2	7.1%
0-1	5.1%

Home win	37.4%
Draw	22.2%
Away win	40.4%
Over 2.5	69.0%
BTTS yes	69.1%

Pregame fair line — favorite Chelsea (40.4%), fair total 3.56, fair spread -0.07

Projected players vs season baseline

Proj = projected for this match · Season = this season's per-game average from match data · deviation badge compares projected xG to the player's season xG/game.

Brentford — projected line vs season baseline (4-2-3-1)

Player	Pos	Min	Proj G	Proj xG	Season xG/g		Proj Sh	Season Sh/g	Proj Rtg	Season Rtg	Rec	Model / FBref Prior
Caoimhin Kelleher	G	95	0	0.00	-	n/a	0	0.0	7.5	54.2	4	FBref Prior G_service
Michael Kayode	D	75	0	0.32	0.14	129%	2	1.5	7.3	53.8	4	FBref Prior D_service
Sepp van den Berg	D	90	0	0.04	-	n/a	0	-	7.2	-	0	Prior D_service
El Hadji Malick Diouf	D	90	0	0.04	-	n/a	0	0.0	7.2	-	2	FBref Prior D_service
Nathan Collins	D	70	0	0.00	0.17	100%	2	1.7	7.1	66.7	3	FBref Prior D_service
Mamadou Sangare	M	76	0	0.17	0.10	69%	1	1.5	7.6	51.2	4	FBref Prior M_service
Dango Ouattara	M	51	0	0.15	0.02	755%	2	0.8	7.3	51.0	4	FBref Prior M_service
Kevin Schade	M	95	0	0.23	0.19	ok	2	2.0	7.5	54.5	4	FBref Prior M_service
Yehor Yarmoliuk	M	32	0	0.08	0.02	245%	2	0.7	7.2	65.0	3	FBref Prior M_service
Mikkel Damsgaard	M	60	0	0.09	0.01	1216%	2	0.5	7.4	55.2	4	FBref Prior M_service
Igor Thiago	F	95	1	0.38	0.58	34%	2	2.2	7.4	49.0	4	FBref Prior F_service
Team totals			1	1.74			18			7.3

Bench Candidates (7 modeled substitutes with empirical playing time)

Player	Pos	Exp Min	App Prob	Substitute Modeling Details
Hákon Rafn Valdimarsson	G	4.4m	10.9%	Market-value substitute prior
Kristoffer Ajer	D	14.5m	23.6%	Market-value substitute prior
Keane Lewis-Potter	D	14.5m	23.6%	Market-value substitute prior
Jaidon Anthony	M	23.4m	51.5%	Market-value substitute prior
Vitaly Janelt	M	22.6m	32.0%	Market-value substitute prior
Gustavo Nunes	F	25.5m	57.0%	Market-value substitute prior

Chelsea — projected line vs season baseline (4-2-3-1)

Player	Pos	Min	Proj G	Proj xG	Season xG/g		Proj Sh	Season Sh/g	Proj Rtg	Season Rtg	Rec	Model / FBref Prior
Mike Penders	G	90	0	0.00	-	n/a		0.0	7.1	-	3	FBref Prior G_service
Reece James	D	75	0	0.00	0.04	100%	2	1.2	7.3	50.0	4	FBref Prior D_service
Levi Colwill	D	89	0	0.00	0.00	100%	1	0.3	7.4	41.0	3	FBref Prior D_service
Maxence Lacroix	D	95	0	0.00	0.01	100%	0	0.2	7.4	50.2	4	FBref Prior D_service
Jorrel Hato	D	78	0	0.00	0.00	100%	0	0.2	7.2	41.2	5	FBref Prior D_service
Morgan Rogers	M	95	0	0.33	0.23	38%	2	1.8	7.6	47.2	5	FBref Prior M_service
Moisés Caicedo	M	16	0	0.14	-	n/a	1	0.0	7.4	66.0	1	FBref Prior M_service
Cole Palmer	M	95	1	0.35	0.21	68%	3	2.4	7.7	46.6	5	FBref Prior M_service
Estêvão	M	12	0	0.15	0.01	1019%		0.3	7.6	43.0	3	XGBoost M_service
Pedro Neto	M	60	0	0.18	0.12	58%	2	1.0	7.7	53.8	4	FBref Prior M_service
João Pedro	F	95	1	0.41	0.47	ok	2	2.5	7.8	56.8	4	FBref Prior F_service
Team totals			2	1.82			14			7.5

Bench Candidates (7 modeled substitutes with empirical playing time)

Player	Pos	Exp Min	App Prob	Substitute Modeling Details
Robert Sánchez	G	5.2m	14.7%	Market-value substitute prior
Malo Gusto	D	14.8m	24.3%	Market-value substitute prior
Wesley Fofana	D	14.7m	23.9%	Market-value substitute prior
Enzo Fernández	M	25.0m	72.9%	Market-value substitute prior
Geovany Quenda	M	15.5m	72.2%	Market-value substitute prior
Emmanuel Emegha	F	26.3m	60.3%	Market-value substitute prior

Antagonist / tactical summary

Midfield is evenly contested — Mamadou Sangare vs Morgan Rogers will be the pivotal duel. Total antagonist xG — Home: 0.63, Away: 0.68. Home danger zone: Striker vs Two Centre Backs — 1v2 Aerial/Physical Battle. Away danger zone: Striker vs Two Centre Backs — 1v2 Aerial/Physical Battle.

Midfield — possession 49.1% / 50.9%, dominance ratio 0.46, key duel: Mamadou Sangare vs Morgan Rogers

Kevin Schade xG 0.16Mikkel Damsgaard xG 0.06Dango Ouattara xG 0.06Pedro Neto xG 0.10Cole Palmer xG 0.19Morgan Rogers xG 0.13

Summary

Antagonist matchup analysis for Brentford vs Chelsea. Most likely scoreline: 1-2. Lambdas derived from 5 home and 5 away individual player duels.

Tactical parity between 4-2-3-1 and 4-2-3-1. Match outcome driven primarily by individual player quality. Midfield is evenly contested — Mamadou Sangare vs Morgan Rogers will be the pivotal duel. Total antagonist xG — Home: 0.63, Away: 0.68. Home danger zone: Striker vs Two Centre Backs — 1v2 Aerial/Physical Battle. Away danger zone: Striker vs Two Centre Backs — 1v2 Aerial/Physical Battle.

Espanyol vs Elche

LaLiga · 20260918 · 3-5-2 vs 4-2-3-1

Predicted 1-0Winner: EspanyolConfidence 0.58λ home/away 1.37 / 1.02Most-likely P 12.4%

Actual vs predicted: result not available yet (upcoming fixture).

XGBoost Reasoning & FBref Dataset Tracing

Antagonist + XGBoost match analysis for Espanyol vs Elche. Player statistics regressed using position-specific XGBoost models (G, D, M, F) with 2026-2027 FBref priors. Micro-duels resolved across 3-5-2 vs 4-2-3-1 (duel lambdas: Home 1.373, Away 0.935). Final calibrated Poisson lambdas: Home 1.373, Away 1.020 (most likely score 1-0).

Home Starters 10/11 XGBoost (9 FBref) Away Starters 9/11 XGBoost (9 FBref) Bench Modeled H: 7 / A: 7 Duel λ (H/A) 1.37 / 0.94 Final λ (H/A) 1.37 / 1.02

FBref Team Priors: Espanyol (Gls/90: 1.6666666666666667, GA/90: 1.67 → E[λ]: 1.235) vs Elche (Gls/90: 1.0, GA/90: 1.0 → E[λ]: 1.237) | Blend Weight: 0.228

Modeling Pipeline Stages (5 layers)

1. XGBoost player regressions for volume & event rates with market-value quality scaling
2. Multi-source prior blending (FBref 2026-2027 tables + Sofascore season history)
3. Antagonist formation duel matrix resolving 1v1 and zonal matchups
4. Team-level prior calibration (FBref possession and attack/defense efficiencies)
5. Bivariate Poisson matrix with Dixon-Coles low-score correction

Standing & form context

Metric	Espanyol	Elche
League position	12.00	19.00
Points per game	1.00	0.25
Attack index	1.25	1.04
Defense (concede) index	2.50	1.04
Form multiplier	1.04	0.96
Standing λ multiplier	1.62	1.03

Lambda decomposition (how the goal rates were built)

Stage	Espanyol λ	Elche λ
FBref Squad Prior E[λ] (weight 0.228)	1.235	1.237
Pre-competitive-floor	1.373	0.935
Pre-calibration (Micro-duels)	1.373	0.935
Final λ	1.373	1.020

competition calibration scale 1.0competition observed total goals 0.0quality gap applied Truemismatch score pressure applied Falsecompetitive underdog floor applied Truestage discount 1.0

Scoreline distribution & markets

Top 5 scorelines	P
1-0	12.4%
1-1	12.9%
0-0	9.2%
0-1	9.2%
2-1	8.8%

Home win	44.8%
Draw	27.4%
Away win	27.8%
Over 2.5	42.9%
BTTS yes	47.9%

Pregame fair line — favorite Espanyol (44.8%), fair total 2.39, fair spread 0.35

Projected players vs season baseline

Proj = projected for this match · Season = this season's per-game average from match data · deviation badge compares projected xG to the player's season xG/game.

Espanyol — projected line vs season baseline (3-5-2)

Player	Pos	Min	Proj G	Proj xG	Season xG/g		Proj Sh	Season Sh/g	Proj Rtg	Season Rtg	Rec	Model / FBref Prior
Marko Dmitrović	G	88	0	0.00	-	n/a	0	0.0	7.0	57.0	5	FBref Prior G_service
Omar El Hilali	D	89	0	0.00	-	n/a	0	0.0	6.9	55.4	5	FBref Prior D_service
Quilindschy Hartman	D	16	0	0.00	-	n/a	0	0.0	7.0	17.2	4	FBref Prior D_service
Vanja Drkušić	D	56	0	0.00	-	n/a	0	0.0	6.7	33.5	4	FBref Prior D_service
Urko González	M	56	0	0.07	-	n/a	1	0.0	6.9	65.8	4	FBref Prior M_service
Tyrhys Dolan	M	87	0	0.18	0.04	338%	2	0.8	6.9	53.0	5	FBref Prior M_service
Gabriel Moscardo	M	52	0	0.07	0.01	726%	1	0.2	6.8	27.6	5	FBref Prior M_service
Javi Puado	M	90	0	0.16	-	n/a	1	-	6.9	-	0	Prior M_service
Pol Lozano	M	16	0	0.07	0.01	359%	1	0.2	6.7	33.2	4	FBref Prior M_service
Roberto Fernández	F	89	1	0.39	0.32	ok	2	1.8	6.8	59.6	5	FBref Prior F_service
Bryan Zaragoza	F	15	0	0.26	-	n/a		0.0	6.5	70.0	1	XGBoost F_service
Team totals			1	1.37			10			6.8

Bench Candidates (7 modeled substitutes with empirical playing time)

Player	Pos	Exp Min	App Prob	Substitute Modeling Details
Angel Fortuno	G	4.4m	10.9%	Market-value substitute prior
Clemens Riedel	D	13.6m	21.8%	Market-value substitute prior
Unai Núñez	D	13.6m	21.7%	Market-value substitute prior
Edu Expósito	M	22.1m	30.8%	Market-value substitute prior
Álex Calatrava	M	22.0m	30.8%	Market-value substitute prior
Marcos Fernández	F	18.5m	71.9%	Market-value substitute prior

Elche — projected line vs season baseline (4-2-3-1)

Player	Pos	Min	Proj xG	Season xG/g		Proj Sh	Season Sh/g	Proj Rtg	Season Rtg	Rec	Model / FBref Prior
Alejandro Iturbe	G	90	0.00	-	n/a		0.0	6.8	-	4	FBref Prior G_service
Kevin Lomónaco	D	90	0.02	-	n/a	0	-	6.9	-	1	Prior D_service
Roy Revivo	D	9	0.15	0.12	ok	1	1.0	5.9	62.0	1	XGBoost D_service
Buba Sangaré	D	85	0.00	0.01	100%	0	0.2	6.4	48.6	5	FBref Prior D_service
Víctor Chust	D	89	0.00	0.07	100%	0	0.4	6.9	52.8	5	FBref Prior D_service
Facundo Buonanotte	M	50	0.18	0.07	170%	2	1.4	6.8	37.6	5	FBref Prior M_service
Germán Valera	M	85	0.15	0.00	5896%	1	0.2	6.8	50.8	5	FBref Prior M_service
Martim Neto	M	42	0.09	0.00	2003%	2	0.2	6.7	31.8	4	FBref Prior M_service
Marc Aguado	M	65	0.05	0.01	311%	0	0.2	6.7	50.7	4	FBref Prior M_service
Lucas Cepeda	M	24	0.12	0.05	138%	2	1.8	6.9	67.8	4	FBref Prior M_service
Ezequiel Ponce	F	40	0.09	0.05	84%	2	0.8	6.6	50.2	4	FBref Prior F_service
Team totals			1.02			10			6.7

Bench Candidates (7 modeled substitutes with empirical playing time)

Player	Pos	Exp Min	App Prob	Substitute Modeling Details
Matías Dituro	G	4.4m	13.1%	Market-value substitute prior
David Affengruber	D	14.3m	23.1%	Market-value substitute prior
Rubén Sánchez	D	13.6m	29.9%	Market-value substitute prior
Federico Redondo	M	22.0m	30.7%	Market-value substitute prior
Grady Diangana	M	22.0m	49.9%	Market-value substitute prior
Fernando Niño	F	25.3m	33.4%	Market-value substitute prior

Antagonist / tactical summary

Midfield is evenly contested — Urko González vs Facundo Buonanotte will be the pivotal duel. Total antagonist xG — Home: 0.45, Away: 0.47. Home danger zone: Striker vs Two Centre Backs — 1v2 Aerial/Physical Battle. Away danger zone: Left Wing vs Right Wingback — Compact Flank.

Midfield — possession 50.8% / 49.2%, dominance ratio 0.52, key duel: Urko González vs Facundo Buonanotte

Roberto Fernández xG 0.22Urko González xG 0.04Pol Lozano xG 0.03Facundo Buonanotte xG 0.15Lucas Cepeda xG 0.08Martim Neto xG 0.06

Summary

Antagonist matchup analysis for Espanyol vs Elche. Most likely scoreline: 1-0. Lambdas derived from 5 home and 5 away individual player duels.

Tactical parity between 3-5-2 and 4-2-3-1. Match outcome driven primarily by individual player quality. Midfield is evenly contested — Urko González vs Facundo Buonanotte will be the pivotal duel. Total antagonist xG — Home: 0.45, Away: 0.47. Home danger zone: Striker vs Two Centre Backs — 1v2 Aerial/Physical Battle. Away danger zone: Left Wing vs Right Wingback — Compact Flank.

Generated 2026-09-18 13:35:28 UTC · Soccer prediction reporting