For those old school football fans who write off xG, as it is new and confusing, I apologise, this one is not for you. However, for those who are curious to discuss or even learn about xG, welcome…Come along!
In the 2020-21 season, Brighton were hailed as the xG Kings, which in short, is a scientific way for saying “they play great, but are unlucky!”
xG is loved by many and probably dismissed by a many-and-one. Despite the dismissal of pundits and dinosaurs alike, it really is not that difficult to understand. There is an algorithm in which the likelihood of scoring is quantified and put into a number, with 1.0 being a definite goal and 0.01 being the half-est of half-chances. The algorithm takes into account distance, angle, defenders, goalkeepers etc. and so that we simpletons don’t have to worry our little brains about it, puts it into a number. As one of my lecturers often told us on my Sports Coaching degree, “don’t worry…you don’t need to do the sum, just know what the algorithm is doing for you!!” So if a team or a player has an xG of 10 but has scored 2, they should have scored 10 but are missing 80% of their chances, with me? I hope so.
The reason xG is important in a modern age is important.
Football has always been very late compared to other sports to adopt many improvements and fact-driven assets. For example, Rugby, NFL, Basketball and many other sports were way ahead of Football in getting sports scientists. The fact we even call these people “Sports scientists” is behind in itself. Considering Sports Science is holistic. A coach, a physio, a psychologist, an analyst, a bio-mechanist and all the other “ists” are all sport scientists, but they don’t dabble in other sciences when they become specialists. No more so than the car salesman doesn’t sell you the car and fix its brake pads or does the pilot serve you drinks mid-flight. There are some cross-overs, but on the whole, sports scientists stay in their lane.
Other examples of Football’s reluctance to change would be goal-line technology, referees not being heard by audiences, head injury protocol or video refereeing…or even looking at successful video refereeing and doing what that sport is doing. Even changing rules for the better of the game, seems to take forever, as if the only thing religions across the World could all agree on is that all the varying God’s took the time to write down the rules of football themselves and thou shalt not ever change them. But that is probably a whole piece for me to write in the future unless someone more important covers it first.
The big difference with football in comparison to other sports is coverage. By being the most popular sport around the world, naturally, it gets more coverage than the others. There are many plus points to this coverage, however, one of the massive negatives (in my opinion) is many unlearned people, having very loud voices, on very large stages, to voice their unlearned opinion. In other words, just because you can kick a ball, does not mean you understand the myriad of progressions and how they work. So, because you don’t get it and it wasn’t around in your day…doesn’t mean it is rubbish. Otherwise, we would all be using pigeons and smoke signals and I don’t know how to even start smoking a blog.
In the pre-xG days, a teams form was solely based on results. There may be a feeling that a club is doing well, usually because they play the “right way”. But even with the right way, there are a number of teams that played that way and went down to the Championship, Norwich, Bournemouth, Swansea etc. Likewise, there have been a number of teams who play a perceived wrong way and stay up. This usually leads to managers either being given longer or sacked, based on “feeling”.
But post xG and for the managers who embrace the future and are “forward-thinking”, it has been possible to quantify the qualitative evidence. In other words, factually prove you’re doing well, without breaking out the “feels”. For modern clubs, focussing on performance more than results.
Brighton 2020-21 xG Kings!
When considering Brighton in 2020-21, this was vital. Graham Potter seems to be linked with every job in England, including the England job itself. I have to admit…I’m a believer!
10 games into the campaign, to the non-believer’s, they would have thought you were mad. Brighton had won just 2 games out of those 10. Indeed, by the end of the season, they had only managed to win another 7 games to make it 9 overall. Sitting in 16th, you would be forgiven for believing they had a bad season…but…
With 15 losses, Brighton had lost fewer games than Southampton, Crystal Palace, Wolverhampton Wanderers and Newcastle United, all above them. They only conceded 46 goals and that is less than 8 of the teams that finished above them. So, what on earth was happening in the attack?
This is where xG can help. Brighton scored 40 goals last season, 1 of which was an own goal, the highest goal scorer being Neal Maupay with 8. Consider Southampton, one place above them got 47 and you see that, fair enough, that is about right…right? Wrong! Brighton had an xG of 51.6, which as we know, means they should have scored an additional 11 goals. Figures that with their good defensive record, should have had them sitting around 10th or 11th.
The explanation is simple. Brighton got into the right positions and created the right chances, they just couldn’t score! Top scorer Maupay himself should have scored an additional 5 goals, making him the worst in the team for G-xG (Goals – expected goals). When you consider Southampton again, the worst G-xG was Nathan Redmond with -1.3, in other words, he only missed a goal and a third of the goals he should have scored. Meanwhile, 8 players all outscored their xG.
Consider the games that would have been won by Brighton rather than the 14 draws they ended up with and it is easy to see why they are indeed xG Kings of 2020-21.
Wolves, the xG Jokers!
Now consider Wolverhampton Wanderers this season, 2021-22. Currently, Wolves sit in 8th place, between Manchester United in 7th and Brighton in 9th.
Undoubtedly, results-wise this does indicate that Wolves are having a good season. Indeed the only teams that have won more games sit above them and the only teams to have won fewer games are all below them. There is one outlying variable that will not appear in the league table though…Jose Sa!
As he arrived from Olympiacos in summer 2021, annoyingly for stat nerds like me, there aren’t many. This season though, he has been brilliant!
In simplistic Match of the Day level statistics, he has conceded 15, which is less than Manchester United’s David De Gea with 29 and also less than Brighton’s Robert Sanchez with 19. But that does not tell the half of it.
Firstly, Sa concedes 0.75 goals per-90, compared to De Gea’s 1.45 per-90 and Sanchez’s 1.00 per-90. Look deeper and it gets even better. Save percentage Sa = 85.9%, De Gea + 71.7% and Sanchez = 76.1%. These statistics can even outdo the Premier League’s leading goalkeeper Ederson and his 73.9%. However, on another important metric PSxG+/- (post-shot expected goals minus goals) Sa with +6.7, loses out only to De Gea (+8.4) but is performing better than Sanchez (+1.3) and significantly outperforms Ederson (-1.8). This statistic basically means how many goals they should have conceded minus the ones they did concede, with a + being how many better off they are. Therefore, this takes into account the quality of the chances they are facing.
Finally, there is the goal scoring statistics. Wolves have scored 16 of their 17 goals, with 1 own goal to help. This actually underperforms, much like Brighton did in 2020-21.
Predictably, Liverpool, the Premier Leagues top scorers, also have the highest xG (52.4) and are outscoring it (55). Equally, Burnley, the league’s lowest-scoring team have the lowest xG (16.9) and are underscoring if you were to not count own goals (15). Manchester United sitting one place above Wolves in the table have the 6th highest xG (27.7) and outscore it (31) and Brighton sitting one place below Wolves in the table, have improved on last season to be 13th in the xG statistics (xG=23.7) underperforming by (-3.)7 with (20) goals. Wolves sit 3rd from bottom on xG (20.7), which is even more concerning when you consider Wolves have scored just 17 goals so far, this puts them above only 20th place Burnley and 18th place Norwich.
So to summarise, Wolves are conceding high probability chances, but thanks to Jose Sa, they do not concede as many of them as they should and they do not create many clear chances and they take even less of them. But ultimately the only statistic that decides the league table are points on the board. There is no doubt that the 31 points on the board for Wolves are impressive, especially when all the evidence points to the fact they are not even playing well.
Making them the Premier League’s xG Joker in the pack!
Andrew Hopper-Davies Football blog 