The Championship's fourth matchweek delivered a string of unexpected outcomes that left our forecasting model bruised. Wrexham's 3-0 demolition of Millwall at The Den and Derby County's 2-0 victory at Portsmouth were the standout shock results, while the model managed to correctly call only five of 12 match result forecasts. From goalless draws to goal bursts, it was a difficult round for the projections, particularly in total goals and exact-score markets.
How our forecasts did
The most glaring difficulty came in exact scoreline calls, where only one of 12 attempts was correct — QPR's 2-1 defeat of Cardiff matching the model's prediction to the letter. Half-time/full-time projections fared little better, with just three graded right. Total goals was another troubled category: only four of the 12 forecasts landed, the model repeatedly anticipating more than 2.5 goals in matches that stubbornly refused to deliver them. Match result and win-or-draw calls each hit five from 12, a below-par return, while first team to score also managed five correct outcomes. The brightest spot was both teams to score, which graded seven of 12 correctly and offered some respite in an otherwise chaotic round.
Tuesday
Lincoln 0-0 Blackburn
A complete miss for the model. Every one of the seven forecasts — from a Lincoln win and both teams to score, to over 2.5 goals and a 1-0 home scoreline — was wrong. Blackburn, billed as steady on the road, held firm in a stalemate that nobody saw coming.
Portsmouth 0-2 Derby
Another clean sweep of incorrect calls. The model had rated Portsmouth as medium-confidence winners and expected goals at both ends, but Derby’s first win of the season — with a clean sheet — overturned every projection.
Preston 1-3 Bristol City
A rare strong performance. The model correctly called a Bristol City win, the visitors to score first (with high conviction), over 2.5 goals, both teams to score, win-or-draw, and Bristol City ahead at the break en route to victory. Only the predicted 0-1 exact scoreline missed, with Preston grabbing a consolation.
Sheffield Utd 3-2 Bolton
The match result and win-or-draw calls landed, but the moderate-confidence forecasts for fewer than 2.5 goals and at least one clean sheet were undone in a five-goal thriller. Bolton twice levelled, but the Blades held on.
Swansea 2-0 Watford
The sole slip was both teams to score: the model expected a Watford goal that never came.
West Ham 4-2 Wolves
This high-scoring encounter vindicated the model’s match result, over 2.5 goals, both teams to score, first team to score, and half-time/full-time calls. Only the exact scoreline fell short as the Hammers ran riot, far exceeding the predicted 2-1 margin.
Birmingham 1-1 Southampton
The model backed a Birmingham win and over 2.5 goals; instead a solitary first-half strike for each side settled things. Both teams to score was the lone correct call from seven, the rest missing as two 4-2-3-1 systems cancelled each other out.
Stoke City 1-0 Norwich
For the second time on the night, all seven forecasts were wrong. Norwich were favoured to extend Stoke’s miserable start, but the Potters kept their first clean sheet and nicked a win that left the model’s analysis — which highlighted Norwich’s superior attacking threat — looking hollow.
Wednesday
Millwall 0-3 Wrexham
A result that upended almost every expectation. The model had called a Millwall win, fewer than 2.5 goals, and at least one side to stay scoreless. The clean-sheet part of that last call was technically correct — Millwall failed to score — but the visitors’ three-goal salvo wrecked the match result, total goals, and first team to score projections.
QPR 2-1 Cardiff
The round’s standout success. QPR’s win, over 2.5 goals, both teams to score, and even the exact 2-1 scoreline all landed. The model only misjudged the first scorer and interval lead, having expected Cardiff to open the scoring, but this was comfortably the most accurate forecast of the matchweek.
West Brom 1-1 Charlton
West Brom struck first, as predicted, but could not hold on, and the 1-1 draw — with under 2.5 goals — foiled the match result, win-or-draw, total goals, and half-time/full-time calls. Both teams to score and first team to score were the two correct outcomes.
Burnley 1-1 Middlesbrough
The model’s call for a Middlesbrough win proved incorrect, as did the over 2.5 goals and the half-time/full-time projection. However, both teams to score and Middlesbrough scoring first were on the money in a tight affair at Turf Moor.
What we got wrong
This was a rough round, and honesty demands we confront it directly. In three fixtures — Lincoln vs Blackburn, Portsmouth vs Derby, and Stoke vs Norwich — the model missed every single forecast. Across those 21 calls, the only ray of light was the both teams to score call in the Stoke game, which unfortunately also failed.
The Lincoln-Blackburn goalless draw was a particular low point. Our analysis noted Lincoln’s stalling momentum and Blackburn’s steadier away form, yet the model still tilted toward a narrow home win and more than 2.5 goals. The 0-0 scoreline, with neither side able to break through, exposed the overconfidence in both the result and the goal projections.
Portsmouth seemed well set: they had scored in every league match, Derby were winless and likely missing top scorer Sammie Szmodics, and a 2-1 home win looked entirely plausible. Derby’s 2-0 away victory, with a clean sheet, was a reminder that early-season data can be deceptive, especially when a side finds form overnight.
Stoke City’s 1-0 defeat of Norwich was another total miss. The model leaned on Stoke’s defensive frailty — nine goals conceded in three games — and Norwich’s superior xG numbers, expecting the Canaries to strike first and win. Stoke held firm and punished us for trusting a trend that collapsed at the first sign of resistance.
What this round changed
Matchweek 4 reinforced the volatility of a Championship season that is still finding its shape. A round this unpredictable is unlikely to define the season, but it serves as a timely reminder that no forecast, however well-reasoned, is immune to the chaos that makes the second tier so compelling.
