You’ve most likely heard of xG. The modern game is full of such stats, but often, you’ll only hear about a team’s xG on a game-by-game basis.
Expected goals, or xG, is a figure given to a team based on their chances in a game. For instance, a penalty gives you an xG of around 0.75 depending on which metric you use – 75% of the time, you expect a penalty to be scored. Every shot your team has will generate an xG number, and at the end of the game, your xG is added up. If you have an xG of 1.75 but scored three goals, then you’ve shown better finishing than the chances suggested. If your xG is 1.75 and you haven’t scored, you could (and perhaps should) have had at least one goal to your name.
Pundits spot trends within these numbers – a team who continually outperforms their xG is in danger of eventually not being as clinical – in 2017/18, Plymouth finished seventh in League One but scored far more goals than their xG suggested they should. That means they have a problem creating good chances, something that would eventually catch up with them; they were relegated a year later.
Brentford used xG heavily in their rise through to the Premier League. If they lost games, but xG data suggested they should have won, the pressure was not on their manager. If they won matches, but xG data suggested they weren’t creating chances, then questions were asked. It is even reported that they routinely looked at a striker’s xG rather than his goals return when making a purchase.
Using data from Football xG, every game played in the Championship has been reimagined this season. Instead of wins being allocated to goals, we’ve got the league table as it would be if every bit of xG led to a goal. It’s hypothetical, of course, but every result has now been allocated win, loss or draw based on xG data, not on goals. We’ve used that to create an alternative table, and it might give us an indication of who could climb the table or tumble down it later in the season.
These stats are relevant to December 8th and do not take into account this weekend’s fixtures. Do be aware that different outlets do have slightly different xG outcomes – Fotmob, for instance, has an xG published straight after a game but is often less accurate than Wyscout, published up to 24 hours later.
If games were won on xG, then Leeds United and Sunderland would be tied at the top of the Championship. This suggests that perhaps Sunderland were hasty in dismissing Tony Mowbray and that Leeds are in a perfect position to keep up the pressure on Ipswich and Leicester, predictably third and fourth in our alternative table. Coventry City, currently 14th in the table, would be fifth if their results had reflected their xG. That suggests there’s more to come from Mark Robins’ side and that they might benefit from a finisher in the transfer window. In sixth, Hull City are the only team to be in the same position in both tables.
At the bottom, there are some notable numbers. Sheff Weds were 24th going into this weekend, but their xG should have seen them out of the bottom three. That suggests that as things stand, they are in a position to avoid relegation despite being adrift of safety. Millwall, currently keeping their heads above water, could be in danger of being dragged into the relegation battle. It might also be worth keeping an eye on West Brom and Preston – both are in the top ten, but their xG outcomes are not good – they’re going to need to be more creative if they’re to sustain their current play-off ambitions.
As for Rotherham, a swift return to League One looks likely. Based on xG in their games, the Millers would only have four points, which suggests they’re constantly being outperformed by their rivals. Unless they can find a way to stop that, they’re likely to go down, with Huddersfield, another struggler, currently in danger of the drop. As for Swansea and Stoke, both have pulled the trigger on their managers but have created more chances than their goal haul suggests.
Here’s the full table as of December 8th – keep your eyes peeled tomorrow for League One.
Editor In Chief
