Gameweek 1's Biggest Fantasy Surprises
Five players who blew past PitchMetric's predicted points in the season's opening round — and why a Gameweek 1 model has almost nothing to go on.
Model v2 — identical to v1 for this gameweek. There's no current-season data for anyone yet in the season opener, so v2's price-based signal has nothing to work with (what changed in v2).
Every player here cleared 30 minutes on the pitch in Gameweek 1, which leaves 229 qualifying performances to rank by one number: actual points scored minus PitchMetric's predicted points (xP) for that player, that match. The five below are the biggest positive gaps of the round — the players the model most under-rated. A shared theme runs through all five, and it's a real limitation, not a bug: Gameweek 1 is the one week of the season where nobody, anywhere, has kicked a ball yet — no goals, no assists, no minutes for any player at their current club this season. With no current-season form to go on for anyone, the model falls back to treating almost every outfield player the same: a flat floor built only from how likely their team is to keep a clean sheet, not from anything about the individual. That's exactly why it's structurally blind to the kind of breakout individual moments that decide fantasy gameweeks. See the methodology for the full model.
Ranked among 229 players who played at least 30 minutes in gameweek 1, by actual points minus predicted xP. Average surprise across that pool: +1.84 (std. dev. 3.46).
Photo: Bryan Berlin, CC BY-SA 4.0, via Wikimedia Commons
.jpg){kind=link}
Box score: 1 goal, 1 assist, clean sheet, 2 bonus · xG 1.47, xA 0.21 · ICT 12.9
The single biggest surprise of the round, in either direction. De Cuyper only signed for Brighton this summer, arriving from Club Brugge in Belgium with a real goalscoring pedigree for a full-back — 29 goals and 37 assists across his career there. None of that reputation counts for anything in a league he'd never kicked a ball in before this match, so the model had nothing to price him on beyond a bare 1.53-point floor. It was also the season's very first game, so there was no current-season form for anyone else to lean on either. He repaid the blank slate with a goal, an assist, a clean sheet and two bonus points in a 4-0 rout of Aston Villa. It wasn't a fluke on the underlying numbers either: 1.47 expected goals and 0.21 expected assists (1.68 combined) means the chances he actually got were genuinely high-quality, not scraps.
Photo: Timmy96, CC0, via Wikimedia Commons
.jpg){kind=link}
Box score: 2 goals, clean sheet, 3 bonus · xG 1.41, xA 0.02 · ICT 14.5
Brighton's second Gameweek 1 shock, from the same 4-0 win over Aston Villa as De Cuyper above. Hinshelwood is a genuine homegrown talent, having come up through Brighton's own academy — but he'd never scored a Premier League goal before this match, and with it being the season's first game, there was no current-season form to separate him from anyone else in the squad either. He came off after 63 minutes with a brace, the game's top bonus haul, and 1.41 expected goals backing up the finishing — good for a predicted score of just 1.56 points, essentially the same bare floor every outfield player without a scoring history got that week. Two Brighton players this high on the same list, from the same clean sheet, says more about how little separated any two names in a squad before a ball was kicked than it does about either player individually.
Box score: 1 goal, clean sheet, 1 bonus · xG 0.28, xA 0.00 · ICT 8.5
Hull City beat Manchester United 2-0 on their return to the Premier League, and Mendy was one of two home defenders who scored in it. Hull hadn't played a top-flight match in years before this one, so there was no recent Premier League record for Mendy or almost anyone else in the squad — the model could only lean on a generic clean-sheet floor of 1.83 points for a newly promoted side playing at home. His goal came from a shot worth only 0.28 expected goals — a low-probability strike rather than a tap-in, which makes the outcome as much a one-off as a repeatable signal. No free-licensed photo of Nobel Mendy is available on Wikimedia Commons yet.
Photo: Kieran Clarke, CC BY 2.0, via Wikimedia Commons
_(cropped_x2).jpg){kind=link}
Box score: 1 goal, clean sheet, 2 bonus · xG 0.52, xA 0.00 · ICT 7.9
The second of two Hull City center-backs to score in that 2-0 win over Manchester United — same clean sheet as Mendy above, same 1.83-point floor, a different goal (0.52 expected goals this time, a somewhat more presentable chance). A newly promoted club with no recent top-flight matches produced two double-digit defensive returns from sixty-three shared minutes against one of the league's biggest names. There was no shame in United failing to break Hull down on the day — but nothing could have flagged in advance that it would be two Hull defenders, of all people, doing the damage at the other end.
Photo: Werner100359, CC BY 4.0, via Wikimedia Commons
.jpg){kind=link}
Box score: 2 assists, clean sheet, 3 bonus · xG 0.05, xA 0.33 · ICT 7.4
Two assists and a clean sheet in a 3-0 win over Spurs, from a midfielder who — like literally everyone else in the league that weekend — had no current-season form behind him yet, so the model priced him at a flat 1.95 points regardless of who he actually was. His 0.33 expected assists is the most encouraging underlying number on this list relative to the actual return — genuine chance creation, not just a stat-line that happened to him. Worth revisiting once he's built up a real run of matches for the model to actually learn from.
Four of these five points came out of just two matches — Brighton's and Hull City's, both wins built on a clean sheet plus a goal from an unheralded name. That's the Gameweek 1 pattern in miniature: the model correctly rates a fixture's clean-sheet chance, but has no way yet to know which individual will supply the goal on top of it. From Gameweek 2 onward, every one of these players now has a real observed data point, and the model's individual rates stop being flat — see how that plays out in Gameweek 2's surprise list.
Gameweek 1's biggest disappointments
Every gameweek also has an equal and opposite story: players our model rated highly who delivered far less. Same method, same 30-minute cutoff, ranked the other way — the five biggest gaps between what we expected and what actually happened, on the low side.
Box score: — · xG 0.00, xA 0.01 · ICT 1.2
Sent off before half-time in Villa's 4-0 opening-day mauling at Brighton — an early red card is close to the single worst thing that can happen to a predicted starter's afternoon, and no model input flags it coming. Nothing to learn about the model here; this is exactly the kind of pure variance a point-in-time prediction can't see.
Box score: — · xG 0.00, xA 0.00 · ICT 4.8
Played the full 90 in the same 4-0 defeat and put through his own net for good measure — an own goal costs a player the same 2 points a genuine defensive lapse would, on top of conceding four without a whisper of a clean-sheet bonus. Villa's back line simply had the single worst afternoon of anyone's gameweek.
Box score: — · xG 0.00, xA 0.00 · ICT 6.1
A yellow card and four goals conceded in the same match Lindelöf and Gomes went through — three of our five biggest GW1 misses came from the same back line on the same afternoon. When a defense implodes this completely, it isn't three independent surprises, it's one team-level event the model can't see building from a single player's history.
Box score: — · xG 0.00, xA 0.00 · ICT 7.9
The fourth Villa defender in our worst five — 81 minutes played, four goals conceded, nothing to show for a predicted return built on a fixture that looked far more even on paper than it played out on the pitch.
Box score: — · xG 0.00, xA 0.00 · ICT 8.4
A yellow card and two goals conceded away at Ipswich on newly-promoted Sunderland's opening day — a perfectly ordinary bad result for a team we (and most models) simply had too little data on yet to call with any real confidence.