This article examines the biggest upsets and unluckiest teams from the 2018/19 Premier League campaign, by examining historical odds and expected goals (xG) data.
From reading the often extreme and divisive post-game reactions posted on social media, to being repeatedly told by hyped-up broadcasters they’re potentially witnessing a “giant killing”, soccer fans can be forgiven for thinking they’re witnessing a major upset or complete capitulation almost every weekend.
In addition, supporters will also traditionally believe that opposing teams are benefitting from huge slices of luck, while simultaneously believing their own club continually finds itself hard done by – either by mediocre officiating, moments of madness, or fluke goals.
Rather than being drawn in by the hype and hyperbole, bettors should remember to play close attention to what the actual data says – when determining whether they’ve actually witnessed David concur Goliath or whether Lady Luck was really all that separated two teams in any given game.
This article will look at the genuine big upsets from last year’s campaign – as well focusing on the actual “unluckiest” teams in England’s top division, by examining Pinnacle’s odds and expected goal (xG) figures for individual matches.
All stats in this article are provided by Understat. For reference, in each table, DEEP refers to passes completed within an estimated 20 yards of goals, excluding crosses. While expected points (xPTS), is the number of points a team would have been expected to have won based on their expected goals (xG) total.
The three biggest upsets in the Premier League 2018/19
Manchester City 2-3 Crystal Palace
- - Gundogan 1-0 27’ (xG 0.29)
- - Schlupp 1-1 33’ (xG 0.5)
- - Townsend 1-2 35’ (xG 0.1)
- - Milivojevic (pen) 1-3 51’ (xG 0.76)
- - De Bruyne 2-3 85’ (xG 0.01)
According to Pinnacle’s odds, the biggest upset of last season’s Premier League came courtesy of Crystal Palace’s Christmas trip to eventual champions Manchester City on Saturday, December 22.
With opening odds of 22.93, the Eagles had an implied win probability of just 4.4% of taking all three points from the Etihad.
City came into the game with the only 100% home league record in the top six tiers of English soccer. While Palace, who had not won away at City for 28 years – lost their previous two visits 5-0.
With odds of 19.77 at kick-off, which granted them a 5.1% chance of victory, Palace found themselves 1-0 down on 27 minutes when Ilkay Gundogan ghosted into the box to head home Fabian Delph’s cross.
For City, all seemed to be going according to plan – but what happened next left fans, bettors and bookmakers, alike, stunned.
In the space of just eight minutes, Palace managed to completely reverse the score line.
First Jeffrey Schlupp equalised with a fizzed shot from the edge of the area, before Andros Townsend produced arguably the goal of the season with a superb volley from 25 yards out – after City failed to clear a free-kick.
As result, Palace’s price eventually moved to 4.55 (22%) – dropping even further down to 1.32 (75.80%) after a 51st minute Luka Milivojevic penalty had put them 3-1 up.
The underdogs eventually won the game 3-2, after an 85th minute goal from Kevin De Bruyne gave City some hope.
A remarkable game in many ways, given Palace’s three goals came from two strikes from outside the area and a penalty – the Eagles actually had little in the way of any possession near City’s goal.
Newcastle 2-1 Manchester City
- - Aguero 0-1 1’ (xG 0.44)
- - Rondon 1-1 65’ (xG 0.09)
- - Ritchie (pen) 2-1 79’ (xG 0.76)
Newcastle vs. Manchester City
Traveling to St. James’ Park on Tuesday January 29, Manchester City were also involved in the season’s second biggest upset – falling 2-1 to a Newcastle United side who at the time were just two points above the relegation zone.
With just one win in their previous six league games, few expected the Magpies to put any sort of a dent in City’s title hopes.
The market reflected this with Pinnacle giving opening odds for a Newcastle victory at 11.75 – which gave them an 8.5% chance of victory.
Looking at historical data, City had not lost to Newcastle in 22 league games – with only three draws between the teams in that period.
get the best odds on epl betting
Bet on the 19/20 Premier League season at PinnacleSign up hereLog in here
With odds of 15.11 (6.6%) at kick-off, Newcastle suffered the worst possible start when after just 24 seconds Sergio Aguero pounced inside the box to score from a David Silva headed pass.
Newcastle were behind for over an hour before striker Salomon Rondon fired them level on 65 minutes.
Growing in belief and encouraged, both by the home support and an unusually toothless Manchester City, Rafael Benítez’s men then found themselves 2-1 up in the 78th minute.
Overplaying at the back, Fernandinho fouled Sean Longstaff inside the area after being robbed of possession.
Waiting more than two minutes to take the penalty, as goalkeeper Ederson was receiving treatment, Matt Ritchie duly dispatched the spot kick to send the Newcastle faithful into ecstasy.
Despite their promising start, City were flat and failed to adequately trouble Newcastle goalkeeper keeper Martin Dubravka – with their last shot on target coming in the 58th minute.
Chelsea 0-1 Leicester City
- - Vardy 0-1 50’ (xG 0.13)
Chelsea vs. Leicester City
In an odd coincidence, two of the three biggest upsets of the 2018/19 season came on the same day – Saturday 22nd December. For last season’s third biggest upset – mid table Leicester City travelled to Stamford Bridge to face fourth-placed Chelsea.
Unbeaten at home so far that season, Maurizio Sarri’s Chelsea were heavily favoured – with Leicester City, who had not won at Stamford Bridge since 2000, being given opening odds of 13.77 (7.3%) – and odds of 10.58 (9.5%) just moments before kick-off.
Failing to make the most of a good start, Chelsea found themselves behind on the 51st minute– when Jamie Vardy latched onto James Maddison’s pass to fire past Kepa Arrizabalaga from inside the area.
The result of which saw Leicester’s odds of winning slashed from 9.82 (10.2%) to 2.8 (35.7%).
Resisting heavy pressure in the final moments of the game, Leicester somehow managed to hold out – with Antonio Rudiger seeing his header go just past the post on 89 minutes, before Marcos Alonso struck the post two minutes later.
The three “unluckiest” teams in the Premier League 2018/19
We can use the expected goals (xG) statistic and find which games contained the biggest swing between the xG and actual score line.
Given we’re concentrating on “unluckiest teams”, only results in which teams who had a higher xG than their opponents, but failed to win, will be considered.
However, an honourable mention should go to West Ham’s home game against Huddersfield on March 16, which finished 4-3 to the Hammers, despite both teams recording an xG of 1.88 and 1.67, respectively.
It’s important to state however, that there are many different factors that contribute to a team underperforming or over preforming its expected goals total rather than pure ‘luck’.
Examples include finishing skill (Kane), goalkeeping ability (de Gea) and tactics (Atletico Madrid).
Fulham (xG 1.32) 1-5 (xG 1.14) Arsenal
- - Lacazette 0-1 28’ (xG 0.43)
- - Schurrle 1-1 43’ (xG 0.33)
- - Lacazette 1-2 48’ (xG 0.02)
- - Ramsey 1-3 66’ (xG 0.15)
- - Aubameyang 1-4 78’ (xG 0.14)
- - Aubameyang 1-5 90’ (xG 0.14)
Fulham vs. Arsenal
Finishing second-bottom of the Premier League with just 26 points from 38 games, Fulham have no one but themselves to blame for their own misfortunes.
However, after their October home game against Arsenal, they could have considered themselves more than just a little unlucky not to walk away with a respectable draw.
Hosting Unai Emery’s Arsenal – the game is perhaps best remembered for the collective exclamation of “We’ve got our Arsenal back!” by the delirious away fans.
In what would become the ninth game of their 22 game unbeaten run, Arsenal fans appeared to celebrating a little too early – as their early season promise and good form would soon regress.
Eventually, Arsenal would finish fifth, one point outside of a coveted Champions League spot, after winning just two of their last six league games.
Recording an expected goals (xG) tally of just 1.14 – compared to the five goals they actually scored, Arsenal’s forwards where in superb form, albeit with unsustainable levels of over-performance. An outrageous Aaron Ramsey back heel was the pick of the goals following a silky counterattacking move.
Arsenal were also perhaps aided, and Fulham compounded, by the performance of goalkeeper Marcus Bettinelli would play just one more league game all season – a 4-2 away loss to Cardiff City, who would also be relegated to the Championship.
Burnley (xG 1.89) 1-5 (xG 1.89) Everton
- - Mina 0-1 1’ (xG 0.36)
- - Digne 0-2 12’ (xG 0.07)
- - Sigurdsson 0-3 21’ (xG 0.76)
- - Gibson 1-3 36’ (xG 0.91)
- - Digne 1-4 70’ (xG 0.02)
- - Richarlison 1-5 92’ (xG 0.14)
Burnley vs. Everton
Just two points off the bottom of the Premier League table, struggling Burnley hosted mid table Everton on Boxing Day – hoping for some serious Christmas cheer to help ease their relegation worries.
Eventually finishing 15th, Burnley had to every right to have a post-match ponder about how exactly they managed to lose 5-1 to an Everton side they generally matched for 90 minutes.
With an equal expected goals (xG) tally of 1.89 each, combined with an expected points (xPTS) total of 1.37 and 1.33 respectively, the ‘fair’ result would have been a much needed point for Burnley rather than the underserved thrashing they received.
Unfortunately for them, they found ex-Barcelona and current Everton left-back Lucas Digne in inspired form.
Eventually named Everton’s Player of the Season, the French international whipped home a left-footed free kick on 12 minutes, and, despite a strong hand from Joe Hart, scored with an expected goal (xG) total of just 0.07.
Twenty minutes from time, Digne doubled his tally with an even better goal – when he scored with a drilled effort from way outside the box, with an expected goals (xG) total of just 0.02.
Deep into injury time, Richarlison compounded Burnely’s misery – with an exquisite finish with the outside of his left boot. An opportunity with an expected goals (xG) total of just 0.14.
Bournemouth (xG 2.85) 3-3 (xG 0.98) Watford
- - Deeney 13’ 0-1 (xG 0.04)
- - Deeney 26’ 0-2 (xG 0.56)
- - Ake 33’ 1-2 (xG 0.55)
- - Wilson 36’ 2-2 (xG 0.04)
- - Sema 37’ 2-3 (xG 0.04)
- - Fraser 39’ 3-3 (xG 0.21)
Bournemouth vs. Watford
Separated by just two points and three places on January 2nd, Watford’s New Year trip to Bournemouth might not have surprised most casual observers when it ended a draw.
However, the reality was somewhat different, with the home side unlucky not to take all three points – despite actually being 2-0 down inside the opening 26 minutes.
In a game that saw all six goals come within the first 39 minutes of play, Troy Deeney opened proceedings with a perfectly directed header from outside the six yard box – with an expected goals (xG) total of just 0.04.
Doubling Watford’s advantage with a conventional side footed strike 13 minutes later, Nathan Ake then pulled one back for Bournemouth on 33 minutes, heading home a rebound after Ben Foster produced a stunning save to deny Dan Gosling.
Just three minutes later, Callum Wilson’s looping header from Ryan Fraser’s free kick drew Bournemouth level – with the England international’s effort having an expected goal (xG) tally of just 0.04.
In one of the most entertaining games of the season, Watford immediately regained the leader through a Ken Sema half volley from outside the area – again, with an expected goal (xG) total of just 0.04.
With the final goal of the game, just two minutes later, Ryan Fraser levelled for the home side.
Despite never leading, Bournemouth’s 25 shots, 12 of which were on target, gave them an expected goals tally of 2.85 – which compared to Watford’s total of just 0.98 – should have seen them win the game comfortably.
Why this doesn’t tell us the whole story
While it’s interesting to look at the biggest upsets or the “unluckiest” teams on an individual basis, it would be foolish to take any insights garnered from this analysis and apply it to your assessment of a team moving forward.
This single game sample will tell us very little about how a team will perform on a regular basis and contains far too much noise to be useful in any form of predictive modelling. If however, we begin to see pattern develops over a longer period of time (and a bigger sample of games), it might be worth investigating further.
For example, over the coming season, bettors may choose to use expected goals (xG) to judge whether a team is actually over/under performing – and whether or not they’ll eventually regress to the mean. This can be particularly helpful when making early-mid season predictions.
Similarly, bettors would also be best advised to closely track the movement of odds in the lead up of an event.
To know the “fair” market price on a game, one can simply check Pinnacle’s odds one hour before the start of the event.