The Most Unlucky Teams in Premier League History (by xG)

Football has a way of punishing the deserving and rewarding the fortunate. Every season, a handful of teams create enough chances to finish comfortably in the top half but end up scrapping near the bottom. And every season, a couple of teams overachieve spectacularly, winning games they had no statistical right to win.

xG data lets us quantify this. When a team's actual goals scored falls significantly below their xG total, they've underperformed — they've created the chances but haven't converted them. Over a full season, a big gap between xG and actual goals is one of the clearest signals of bad luck (or, less charitably, terrible finishing) in football.

Here are some of the most unlucky teams in recent Premier League history, measured by the gap between what they created and what they got.

Brighton 2022/23 — The Gold Standard of xG Heartbreak

No conversation about xG underperformance is complete without Brighton under Roberto De Zerbi. This is the team that practically made xG a mainstream topic.

Brighton's underlying numbers in 2022/23 were astonishing. They created enough high-quality chances across the season to be in the conversation for European qualification based on xG alone. Their buildup play was among the most sophisticated in the league — progressive passes, chances from open play, xG per shot — they were elite across the board.

And yet the goals didn't come. Not at the rate they should have, anyway. Brighton's actual goal tally fell well short of what their xG suggested. They'd dominate possession, create three or four clear-cut chances, and either miss them, hit the woodwork, or find the opposition goalkeeper in inspired form.

The gap between their xG and actual goals was one of the largest in Premier League history for a single season. It was a textbook case of a team that was far better than their league position suggested. Anyone watching Brighton play could see they were outstanding. The results just didn't reflect it week after week.

What makes Brighton's case particularly interesting is what happened in the broader picture. The quality of their play under De Zerbi attracted attention from bigger clubs, with the manager himself eventually moving on. The underlying data wasn't lying — this was a genuinely excellent team whose results didn't match their performances.

Liverpool 2020/21 — The Defending Champions' Nightmare

Liverpool's title defence in 2020/21 was one of the strangest seasons any champion has endured. They finished third, which sounds fine on paper, but the journey was anything but smooth. There was a period in mid-season where they went on an alarming run of home defeats — Burnley, Brighton, Manchester City, Everton, and Chelsea all won at Anfield.

The xG data told an interesting story. Liverpool's chance creation remained strong throughout much of the season. They were still generating high-xG opportunities at a rate consistent with a top-four side. But their conversion rate, particularly in that mid-season slump, fell off a cliff.

Part of this was the absence of Virgil van Dijk and the resulting defensive reshuffle, which is a genuine quality issue rather than luck. But in attack, Liverpool were still creating the chances — they just weren't finishing them. Their xG underperformance during that awful Anfield run was severe. On another timeline, where those chances went in at a normal rate, the narrative of that season would have been completely different.

Newcastle 2018/19 — Rafa's Impossible Job

Rafael Benitez's Newcastle were perpetually set up to frustrate. Playing in a low block with limited resources, they ground out results through defensive organisation and the occasional counter-attack. The conventional wisdom was that Newcastle under Rafa were a dull, limited side who got what they deserved.

The xG data paints a more nuanced picture. While Newcastle were never going to challenge for the top six, their xG numbers suggested they were creating more than enough to be comfortable in mid-table. Their actual goal return was notably lower than their xG for the season.

This was a team constrained by circumstances — underinvestment from the ownership, a small squad, and a manager who had to prioritise not losing over trying to win — but when they did create chances, they created decent ones. They just didn't convert them at a normal rate.

Newcastle survived comfortably in the end, but the xG data suggested they should have had a significantly less stressful season than they did. Benitez was building something better than the results showed, which made his eventual departure — and the club's initial decline after it — all the more frustrating for Newcastle fans.

Leicester 2019/20 (Second Half) — The Great Collapse

Leicester's 2019/20 season is remembered for one of the most dramatic second-half collapses in recent Premier League history. They were flying at Christmas, sitting comfortably in the top four and looking nailed on for Champions League qualification. Then it all fell apart. They won just three of their last 14 league games and finished fifth, missing out on the Champions League.

What's often overlooked is that Leicester's xG numbers didn't decline as dramatically as their results did. They were still creating decent chances throughout that awful run — they just stopped converting them. At the same time, their xG against remained relatively stable, but they started conceding from lower-quality opportunities.

The xG data suggests that Leicester's collapse was driven significantly by variance — things that had been going their way in the first half of the season stopped going their way in the second half. Their actual performance level dropped less than their results suggested. This doesn't fully explain the collapse (fatigue, injuries, and psychology all played a role), but it suggests the drop-off was amplified by bad luck on top of a genuine but smaller decline in performance.

What xG Underperformance Actually Tells Us

When a team underperforms their xG, there are several possible explanations. It's usually a combination rather than a single factor.

Pure luck

Football involves a lot of randomness. A shot that's 0.15 xG goes in about 15% of the time. If a team takes 100 such shots over a season, they'd expect roughly 15 goals. But the actual number could be anywhere from 8 to 22 without anything unusual happening — that's just how probability works. Some seasons, the bounces go your way. Some seasons, they don't.

Finishing quality

Some teams genuinely lack clinical finishers. If your main striker is hitting shots straight at the goalkeeper or blazing over from good positions, your xG will be high but your goals will be low. This is a quality issue rather than a luck issue, but it still shows up as xG underperformance.

The tricky thing is distinguishing between a striker having a bad run (which will correct itself) and a striker who simply isn't good enough to convert at the expected rate (which won't). This is where watching the matches, not just the data, remains important.

Goalkeeper heroics against them

Sometimes the opposition goalkeepers just have blinding games against you. It's not that you're finishing badly — it's that you keep running into the best performance of the other keeper's season. This is pure bad luck and it absolutely affects xG underperformance.

Shot placement within high-xG positions

Two shots can have identical xG values because they're taken from the same position, but one is placed into the bottom corner and the other is blasted straight at the goalkeeper. The xG model doesn't know where the shot is going — it just knows where it's taken from. A team that consistently hits shots centrally from good positions will underperform their xG even though the model thinks they're creating good chances.

Post and crossbar

Every season, some teams hit the woodwork significantly more than average. Hitting the post is essentially a near-miss that xG doesn't account for — the shot was on target-ish but didn't go in, and it won't show up in any adjustment to the stats. Teams that hit the woodwork a lot are genuinely unlucky.

Does xG Underperformance Predict Future Improvement?

This is the million-pound question, and the answer is: usually, yes.

Research consistently shows that teams who significantly underperform their xG in one season tend to improve their conversion rates the following season, even without major squad changes. This is because the luck component — goalkeeper heroics, woodwork hits, and random variance — tends to even out over time.

However, the regression isn't always complete. If a team's underperformance is partly driven by genuine quality issues (poor finishing, inability to beat set defences), those factors may persist. The key is separating the random variance from the systematic issues, which requires watching the matches as well as studying the numbers.

For prediction purposes, xG underperformance is one of the strongest signals available. A team that created chances worth 55 goals but only scored 45 is almost certainly going to score more next season, assuming a similar level of chance creation. They're one of the best "buy low" opportunities in football analytics.

The Opposite Side of the Coin

It's worth briefly mentioning that xG overperformance — scoring significantly more than your xG — is generally even less sustainable than underperformance. While some teams can sustain mild overperformance through exceptional finishing talent (think of teams built around elite strikers), large overperformance almost always regresses.

This is because finishing is one of the least consistent skills in football. Even the best strikers have seasons where they convert above their expected rate and seasons where they convert below it. Building a prediction model that assumes overperformance will continue is a recipe for disappointment.

Conclusion

The unluckiest teams in Premier League history share a common trait: they did the hard work of creating chances but didn't get the rewards they deserved. Whether through wayward finishing, inspired opposition goalkeepers, or simple bad luck, these teams left points on the table that their performances merited.

For the analytics-minded, these teams represent something else: opportunity. xG underperformance is one of the most reliable regression signals in football. Teams that create chances don't stay unlucky forever. The data catches up. The goals come. And the teams that looked like they were struggling often turn out to be better than anyone gave them credit for.

The next time you see a team creating chance after chance and not converting, don't write them off. Feel sorry for them — but also take note. Because the data says their luck is about to change.