As well as VAR another newer thing in football is the xG (expected goals) statistic, and unsurprisingly Man City fans are whinging that this too is evidence that Liverpool are undeserving of their success. In a complete misunderstanding of probability they have claimed Liverpool don't deserve to be winning the league since Man City are top of the xG league. xG tells you the averaged probability of scoring a goal in different situations, like a header inside the 6 yard box or a volley from 30 yards. The xG for a penalty is 0.76. This means on average a penalty is scored 76% of the time. xG is averaged over many players in many teams in many leagues over many years. xG is calculated based on professional players but doesn't take the specific player into account. James Milner has scored 17 penalties in the Premier League and missed 2. So he has a penalty conversion rate of 89%. Is Milner luckier than the average penalty taker? Or is he just better? If you're deluded and think he's lucky you would take Milner off penalties because you'd expect him to miss the next 4 penalties so he returns to the average of 76% conversion. If you think he's better then you tell him to keep taking penalties because he's probably going to score the next 4 penalties. Elite sportspeople do not revert to the mean. Things evening out over a long enough time is in the context of random events or 50/50 decisions. The point of having skill is to outperform random results. And the point of being #worldclass is to outperform other top professionals. So it shouldn't be surprising that the best team in the best league is frequently doing better than xG suggests. xG also has nothing to do with "fair". It isn't fair for a game to end in a draw just because both teams have the same xG. If teams have the same xG, the one with better players most likely wins. This is like the idea of saying "we'll make it an end to end game and back our players to score more than yours". Teams don't "deserve" to score 76% of penalties and you aren't unlucky if you miss 50%, you're just crap at them. xG isn't perfect, it's just information. It's better than shots on target which was previously used as the main way to measure scoring chances. You can win a game with zero shots on target by putting a cross into the 6 yard box and a defender scores an own goal trying to stop a tap in. That doesn't mean shots on target are meaningless at how you do in a game, just that they don't tell the whole story. xG doesn't tall the whole story either but at least it takes the quality of the chance into account, not just counting all hoofs towards the goal as equally valuable. A large xG is a good sign that you're making lots of chances, but also that you're wasteful if the actual goals scored isn't also high. Some recent results: Spurs 0-1 Liverpool (1.10-1.78) Liverpool 2-0 Man Utd (2.01-1.42) Wolves 1-2 Liverpool (1.84-1.26) But we deserved to beat Wolves even though they "won the xG". Alisson deserved to make his save against Jiménez because his positioning was perfect, not average. Firmino deserved to score the winner because his finishing was perfect, not average. If we had average players we would have scored about 6 fewer goals this season and conceded about 7 more goals. It's not actually that simple though, because different players would play in different ways and not make the same chances in the first place. xG is like imaging what would happen if just as an opponent was about to take a shot you teleport out Alisson and replace him with someone like Pickford. Of course we would concede more goals and be lower in the table. Man City have scored more goals than us but their actual goals and xG is almost identical. That just means they make lots of chances (De Bruyne) but their strikers aren't doing anything special. Sterling loves a tap in but how many screamers can you remember him scoring?
You lost me at 'Man City fans'. I have to turn off 5Live when they belly-ache now, especially when Uniturd fans, the statistically biggest beneficiaries of VAR so far, jump on to support them.
And this highlights using quantitative information on its own without qualitative information. The blight of modern, neo-conservative economics - the price of everything and the value of nothing.
I will swim against the tide once again and say thanks for the explanation as I really didn't know what or how xG was working out
Useful guide but could still do with knowing exactly how it is compiled as it all sounds a bit too subjective to me e.g. Is it an observer who after every attack decides that it is worth .34 of an expected goal and adds it to the total?
I think every opportunity will be given an assigned value. So a shot outside the box will have a 5% chance of going in but a header in the 6 yard area will have a 50% chance of going in. Our something along those lines.
The chances are categorised based on factors like position on the pitch (distance and angle to goal), type of shot (header/foot), and state of play (set-piece/regular play). There's a little bit of subjectivity in defining the chances but then it's just based on numbers of how many of that type of shot actually go in. Here's the xG data for Salah's screamer against Chelsea which was 2.5% chance of going in for an average player. please log in to view this image
I remember reading an article not so long ago about the stat analysis that goes into the Liverpool management team. In short, they have people who upload games onto computer software where everything is coded as numbers and coordinates. They can, frame by frame, determine the probability of situations and the likely outcomes. Sounds useless. But to give an example: they see that Mane wide on the left on a counter attack has two choices, a risky pass to open up Salah at a probability of 20% or continuing his run at a 60% rate given defender positioning, etc. They ran it by a few more frames and now the pass to Salah was up to around 50% due to player movement and the angles changing. If Mane took the risky, poorer probability chance earlier in the move, they chat to him after the game with video analysis and go through what he should do next time given similar situations. The whole objective is teaching players (and coaches) about the better 'odds' when making choices and improving the decision making process. It's something that, on the face of it, many would scoff at and laugh at; football's gone too far, over complicating it, players instinctively know what to do and when, etc. But all these little things Liverpool are doing, similar with throw in coaches (46.8% average ball retenation under pressure in the PL compared to liverpool's 68.4%), is clearly adding up. Our players are improving in every aspect - decision making through this approach is one of them. The xG stat and others are useful to our club because they use them correctly. City fans saying we're lucky because of xG stats are just idiots using them wrongly.
I said the other day you can't write it all off as useless and you've just articulated very well here a great example of how it's improving us
Obviously, I'm a curmudgeonly old get and routinely dismiss anything invented after 1971. However, I said days ago that I can accept that these things might be useful to coaching staff, but as a consumer of the game I find it doesn't enhance the experience for me at all. I'm no more interested in what people think should happen than I am about having a detailed run-down of players' eating regime or toilet routine.
