Expected Goals (or ‘xGʼ): The Perfect Stat.

“Men lie, Women lie, Numbers donʼt -Jay-Z

-Wayne Gretzky

-Michael Scott”

-Logan Smith.


I like stats, the ability to quantify how good a player is just by numbers is a lot easier than the “eye test,” which is a real thing i believe in, it just can get very subjective. Stats arenʼt, theyʼre as objective as possible, you canʼt interpret someone scoring 5 goals in 2 games as anything short of amazing, right?

Well, not entirely.

Basic stats like goals or clean sheets donʼt tell the entire story. If someone scores a hat trick from 3 shots all from outside the area, chances are theyʼre probably not going to consistently do that. As well, if a keeper keeps a clean sheet, but only faces 2 shots on target, then thatʼs not really because of the keeper, rather the defense in front of him. You can also argue these stats with one another, for instance someone can say X player is good because he scored 15 goals, and I could double back by seeing that X player takes 7 shots a game, mathematically speaking heʼs going to score a lot because of his volume of shots compared to other players. So how do you find the perfect stat, the stat that does tell the full story? Finding it is not easy, but it exists: Expected Goals.


Expected Goals (or xG) is a stat that determines how many goals a team/player is expected to score/concede based on the number of quality chances they take/ give up. It basically determines if a player or team is lucky or not. It can also delve even deeper, with expected assists (xA), xG chain, xG buildup, and expected goals against (xGA). Now that all sounds complicated, because it is. Itʼs very complicated. But, iʼm prepared to activate maximum virgin mode and try and explain them all. Now since xG, xA and xGA are the easiest because of their similarities, letʼs get started there.


A great case of showing how xG works is looking at Borussia Dortmund this season. They started out the hottest team in world football, but should they have been? xG tells us no, as they massively over preformed in xG all season. Their first game they scored 4 goals, xG had them at less than 2. Then they beat Nuremberg 7-0, xG only had them at 2.19. Then they won 4-0 against Stuttgart, yet they actually lost by xG, 0.94 to Stuttgartʼs 1.04. The team finished with +14 xG, and +11, in xA which basically which means the offense contributed to 25 more goals then expected. Thatʼs a fucking TON of goals. To put that into context, Man City actually UNDERPERFORMED in xG by 4 and in xA by 2. The stat Expected Points (xPts) tells you how many points a team wouldʼve scored if they perfectly met their xG a game, and off of xPts Dortmund finished almost 13 points ahead of their xPts, which actually wouldʼve made them 4th as RB Leipzig and Hoffenheim wouldʼve finished ahead of them. An example of xG being proven right is Freiburg. They finished 13 in the Bundesliga, while only over preforming xG and xGA (expected goals against) by 0.6 each. They finished only 1 point off of what they shouldʼve finished, meaning a team that almost perfectly met their xG & xGA, also almost perfect met their xPts. Thatʼs the beauty behind this stat, it always gets it right.


Now for the even more complicated ones, xGChain & xGBuild-up, or xGC and xGB as Iʼm going to refer to them to spare my thumbs. Simply put, xGC is the total xG of every possession a player is involved in. xGB is the same, minus shots and key passes. Simple, right? WRONG. NOTHING IS EVER THAT EASY. This comes straight from Statsbomb; you have to first find all the possessions each player is involved in, then find all the shots within those possessions, then sum together their xG, and then assign that sum to to each player involved no matter how involved they were, whether you made the first pass or hoofed it 40 yards downfield, you get credit. Got it? Not really? Good, because itʼs xGBʼs turn. This is where it gets fun. As i said before, xGB is exactly the same and xGC, but no shots or key passes are involved. So essentially this stat shines a light on the guys at the back, who donʼt get a lot of shine in attack, and gives them credit for what happens in the build up before the attack. This makes players like Thiago Alcántara & Jorginho (ew) get a lot of credit since they usually start the build up play for the attacks. This system has its flaws however, as a player that does small build ups at the back, like Jorginho, gets the same amount of credit as someone who plays a lot more deep passes to get attacks going, like Thiago. I do still refer to this stat as it more often than not still relevant, just keep in mind it does have its flaws. However these are still very relevant stats, as the players with the highest xGC & xGB are your likely candidates; Messi, Salah, Thiago, Firmino, and many other great attackers. xGB is less attacker heavy, with the likes of the aforementioned Jorginho, as well as guys like Alaba and Kimmich getting a lot of love. These stats definitely give you a more deep dive into how certain teams score and who facilitates and gets the scoring going.


xG and all of its contemporaries might revolutionize how we look at footballing metrics, as more and more teams are investing in analytics departments. It makes me think I should go back to school for this, because itʼs beginning to be essential in the world of football, and it hopefully will help us all out in better understanding the mathematics behind football.


Also massive credits to Statsbomb and Understand, where i pulled all of this information from. Check out their work in these links: https://statsbomb.com/ 2018/08/introducingxgchain-and-xgbuildup/ & http://understat.com

Take care guys.

Author: Logan Smith - 22, from Florida. Liverpool fan for 4 years now, also just a huge fan of the sport of football. You can find me on twitter @smittyboy8833