The GoatyGav Column
Prompted by the stats in recent games I decided to look at, what exactly, constitutes an XG (expected goals) stat in a match. In particular the Newcastle game surprised me in that the XG for Newcastle was 1.08 and for West Ham 2.01. When subsequently factoring in Newcastle’s 17 attempts at goal and 2 on target against West Ham’s 10 & 4, respectively, the stats, for me, were skewed.
Originally developed as a predictive tool for betting the way the XG figure is put together has to do with the number, and position, of shots at goal. Headed attempts, strikes from further distance and from wider areas score lower XG than shots, attempts from closer to the goal and more central efforts respectively. The pitch is split in to various zones. A shot from the zone directly in front of, and closest to, the goal will provide a high XG whereas an attempt from the zone covering the corner of the pitch yields a low XG.
Although expected goals are meant to be an indicator over a period of between five to ten games, and undoubtedly a relevant indicator of goals that can be expected by a team, many factors are not taken in to consideration. Some of the problem with the system is that it is now being used out of context by television and media organisations. XG is quoted for a single game as an analysis of what’s occurred over the ninety plus minutes of that match where, as detailed above, the system was designed to indicate goal expectation over much higher numbers of fixtures. Not the end of the world however that’s just part of the picture.
It would be understating it to say that association football is a dynamic sport. This is true to a greater degree, in general, in the moments before a shot at goal. With play building to such a crescendo before the ‘trigger is pulled’, and so many variables contributing to the final shot how can two goal attempts from the same zone be compared? For starters the striker’s body position will vary. Then you have the difference in the pace of the ball when it’s struck adding, or taking away, from the difficulty of the chance. Then there’s the consideration of whether the ball is hit off the ground or on the volley/half volley. The severity of a bounce and when the ball is hit during it’s arc in the air. Looking back at Romelu Lukaku’s second goal against Crystal Palace a couple of weeks ago was from a high XG scoring area however the finish was far from easy as the Belgian striker had to take the ball up high and close, with an acrobatic technique, to steer it home. The stats would suggest that was, what OPTA describe as, a ‘big chance’ however it was far from it. It should be noted that OPTA are not the only game in town and not all XG systems are the same with some more sophisticated than others but all, in my op
One of Pep Guardiola’s tenets is the creation of better scoring chances. Overall Manchester City’s XG reflects this despite, not always, living up to the expectation with finishing falling below the suggested level. Perhaps a more easy to follow stat, resulting from the former Barca and Bayern gaffer’s philosophy, is, now regular, season scoring tallies over 100 goals. Football from another planet? Maybe not but it’s pretty special and you wouldn’t complain if it was West Ham.
I’m not suggesting, for a moment, that expected goals is a useless statistic. Far from it however it’s a system that is open to vast improvement. I do feel that it’s being, slightly, misused when applied to isolated matches, as it is every Saturday night on Match Of The Day, but it certainly has it’s place. Frankly, with my history of regular weekend punts on the footie, perhaps I should pay more attention to it.
Back to the example that I gave in the opening paragraph of the recent home win over the team from Tyneside. Reflecting on the quality of the chances created, and the dominance which West Ham showed that day, perhaps the numbers were not so skewed. If anything I’d suggest that our boys were even better than the 1.08XG vs 2.01. More recently, in Saturday’s fixture in Cardiff, there was an XG of 4.10 vs 0.34. No prizes for guessing which one was which but, bearing in mind our dominance over Newcastle the previous weekend, was that an accurate description of the chances? Cardiff over twelve times more likely to score than our boys? I’ll leave that one with you.
Moving along, as I believe that we should now, this coming Saturday’s game offers a great chance to put another three on the board and place some pressure on Wolves and Watford, who play in the F.A. Cup, above us. We need to see a statement of intent from the team. A decent run in could still see us finish seventh and qualify for next season’s Europa League. Some European nights under the floodlights anyone?