Analyzing the Premier League table based on Expected Points (xP)

The use of probability models in football to calculate Expected Goals, Expected Goals Conceded and Expected Points, have taken off to new heights in recent years – with more fans, coaches, managers and even players paying attention to these hypothetical numbers. Between statistical probability, the actual result, and the expectations for all that has been produced, we analyze how many points each team should have won in each match, based on Expected Points.

GOING DEEPER INTO THE DATA…

Football is a sport where many factors come together, but advanced statistics can argue that a team is more likely to win if they continue to generate more quality chances than their opponents.

The Expected Points (xP) model is based around the Expected Goals (xG) of a match. Once we have the xG of each team, we intricate through a distribution model, what would happen if we simulated the same match thousands of times, getting to know the probability each team would have to win, draw and lose with the same expected goals generated. In this way, we obtain a probability percentage for each team and a value of Expected Points.

HOW xP IS CALCULATED

Expected Points (xP) takes into consideration a team’s underlying numbers in xG and xG against. Teams higher in xP typically have a high xG, and a low xG against. So you may be asking – what is xG?

xG (Expected Goals) takes into consideration the quality of chances in a match, based on several factors – including distance from goal, pressure applied, and the type of shot taken. An absolute worldie that results in a goal, such as Andros’ Townsend’s goal against Manchester City a few years back, may even have an xG of 0.01 – meaning that you would expect that kind of goal to be scored one out of every one hundred times.

After simulating with the distribution model and multiplying each percentage obtained by the value of each sign (win, draw, loss), we obtain the expected points of that match if it happened thousands of times. And therefore, we would obtain a form of deservedness that becomes useful to study at any point in the Premier League season.

Let’s take as an example, the Premier League table until today – February 21, 2022.

RankTeamGPGoalsConcededPts.
1Manchester City26631763
2 Liverpool25642057
3 Chelsea25491850
4 Manchester United26443446
5 West Ham26453442
6 Arsenal23362642
7 Wolverhampton Wanderers24231840
8 Tottenham Hotspur23313139
9 Brighton & Hove Albion25252833
10 Southampton25323732
11 Leicester City23374327
12 Aston Villa24313727
13Crystal Palace25323626
14 Brentford26274224
15 Leeds United24295023
16 Everton23284022
17 Newcastle United24264522
18 Watford24244318
19 Burnley22202917
20 Norwich City25155317

If we take into consideration, the xG, xG Conceded, and xPoints, the picture would be different in many cases. As you can see in table 2, the first 4 teams are in a fair position and until now, they have accumulated the points they statistically deserve. The bottom feeders also remain in the same places, with only Everton achieving a significant change. Here is how the table would look based on Expected Points rather than actual points.

RankTeamsPointsxPDifference
1Manchester City6360.50
2Liverpool FC5754.20
3Chelsea5045.50
4Manchester United4637.30
5Arsenal FC4236.30
6West Ham United4233.70
7Brentford2432.4+7
8Crystal Palace2631.1+5
9Brighton & Hove Albion3330.40
10Tottenham Hotspur3929.6-2
11Southampton3229.1-1
12Everton FC2228.4+4
13Wolverhampton Wanderers4028.1-6
14Aston Villa2723.7-3
15Leicester City2722.7-3
16Newcastle United2222.1+1
17Leeds United2321.3-2
18Watford1818.60
19Burnley1718.10
20Norwich City1712.20

As you can see from the table, Brentford, Crystal Palace and Everton are the clubs who have been the most “unlucky”, otherwise known as “underperforming”. On Expected Points, Thomas Frank’s Brentford would be 7th place, above the likes of Spurs, Wolves and Brighton. The example of Wolverhampton Wanderers also provides an interesting case study. Bruno Lage’s Wolves sit in 7th place at this time, but based on Expected Points, they wouldn’t even make the top ten, falling all the way to 13th. Interestingly, Antonio Conte’s Spurs would be tenth, despite their positive performances under the Italian to make up for Nuno’s horrific start.

In addition to xP, a team’s control of the possession, or lack thereof, is another intriguing data source to study. While possession isn’t the be all end all, it typically allows a team the possibility to create more chances, and more chances of a higher quality. By re-organizing the table based on possession, we can attempt to discover whether or not this thought process is in fact true, or a myth.

RankTeamPossession %ChancesGoals ScoredConversion Rate
1Manchester City67%2276328%
2Liverpool FC61%2276428%
3Brighton & Hove Albion57%1342519%
4Chelsea56%1804927%
5Leeds United54%1122926%
6Manchester United52%1914423%
7Arsenal FC52%1413626%
8Crystal Palace51%1313224%
9Tottenham Hotspur49%1403122%
10Southampton49%1243226%
11West Ham United48%1684527%
12Leicester City48%1293729%
13Wolverhampton Wanderers47%962324%
14Aston Villa46%1063129%
15Brentford45%1272721%
16Norwich City44%911516%
17Everton FC43%1212823%
18Watford43%1172421%
19Newcastle United41%1002626%
20Burnley39%1102018%

Based on the above possession data, the Premier League table would not vary greatly. Wolverhampton and West Ham drop undeservedly into the bottom half, while Vieira’s Crystal Palace moves into 8th. The biggest leap of all sees Leeds United move into fifth, with Brighton in third. Beyond those outrageous moves, the Premier League table based on possession doesn’t look drastically different from the actual Premier League table. This suggests that possession could be a potentially underrated value, even if it’s not the be all end all – as demonstrated by Leeds and Brighton.

Conclusion

Data analysis in football continues to grow, with xP and xG providing a different way of examining the game, and examining a team’s performance. While these underlying numbers only tell one side of the story, it has to be said that actual point and actual goal data may also only tell one side to the story, and that both metrics can be useful to understanding the context of the game.


Processing…
Success! You're on the list.

Explaining the Wide Warrior – Player Role Analysis

A ‘Wide Warrior’ is a full-back who hasn’t quite kept up with the modern trends associated with their position. Rather than relying on attacking threat and potency to make their name, the ‘Wide Warrior’ continues to be an ever-present at the back, doing their best work closer to goal. They excel at the defensive side of the game above all else, even if they may offer certain advantages going forward (like a wing-back), or in half-spaces (like an inverted fullback). Further, not only do they excel at the defensive side of the game, their manager has made clear intentions for that to be the most important facet of their role within the team, restricting their attacking height.

Canadian Premier League – Player Ratings

After the first week of action in the 2022 Canadian Premier League season, we have officially launched our brand new Player Evaluation System. The details of this can be found in this article – Evaluating players based on role continuity, and we have already begun the process of identifying roles for specific players that may have stood out on the day. Here are how the players of the Canadian Premier League currently rank based on our brand new model.

Explaining the Sweeper Keeper – Player Role Analysis

The ‘Sweeper Keeper’ came to deserved acclaim at the 2014 World Cup, when Manuel Neuer showcased his ability to rush out of his goal, and “sweep” in behind Germany’s high-line, almost playing like another centre-back out of possession. Since then, the role of the sweeper keeper has only grown and evolved, becoming a player who is capable in build-up phases, ventures out of their goal to involve themselves in passing sequences higher up the pitch, and completely commands their penalty area by coming off their line and taking control of any situation.

2 thoughts on “Analyzing the Premier League table based on Expected Points (xP)

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s