Ole and xG

Discussion in 'Manchester United Forum' started by Pogue Mahone, Mar 7, 2019.

  1. Mar 7, 2019

    cyberman Full Member

    Joined:
    May 26, 2010
    Messages:
    16,128
    Nobody said we would win the league, I was being told we should have finished 5th last year based on this and how Spurs and Chelsea are miles better than us in every metric based on his stats this season. We are now ahead of Chelsea and 3 points behind Spurs by the way.
    Then I was told it could take multiple seasons for xG to even out when we finished a comfortable second..
    It just screams being too clever. A Jose side would be ahead of xG due to how efficient his sides are but a team that takes more risks would see a smaller number. That doesn't mean we should rehire Jose.
  2. Mar 7, 2019

    Rista Full Member

    Joined:
    Jun 17, 2006
    Messages:
    1,852
    Without going into whether xG is useful or accurate, my feeling is that we've been quite lucky since Ole took over. Not to be confused with undeserved, it's just that 95 times out of 100 we're not going through with last night's performance. Similar with away game vs Spurs and some others.
  3. Mar 7, 2019

    Guy Incognito Full Member

    Joined:
    May 21, 2011
    Messages:
    13,469
    Location:
    Somewhere
    I think some things worked to our advantage (no Pogba, relaxed happy camp going into the second leg) so it'll be interesting to see how Ole'll fare when expectations are risen next season.
  4. Mar 7, 2019

    EwanI Ted Full Member

    Joined:
    Feb 28, 2018
    Messages:
    604
    Given the injuries we’ve contended with going into last nights game, I’m not sure you could say luck has totally been on our side.
  5. Mar 7, 2019

    Dante Bang Average but can post Blindfolded for 15 secs

    Joined:
    Oct 22, 2010
    Messages:
    16,183
    Location:
    My wit's end
    Expected goals is about the position of the shots a team takes.

    It doesn't take into account our passing tempo and defensive shape:

    • A quick move is more dangerous than a slow move, even if they end with a shot from the same place on the pitch.

    • A compact defense will gladly give up a shot if they know a block can be easily made.

    It's not just about the raw numbers. We're an outlier in terms of the way we play football. The pace we have up front and the deliberatley lopsided nature of our formation (amongst other things) makes xG meaningless.
  6. Mar 7, 2019

    SqualorVictoria New Member

    Joined:
    Dec 25, 2013
    Messages:
    451
    Supports:
    City
    He's very lucky so far regarding chance conversion on both sides of the pitch, but then again he will probably have a summer of recruitment, and preparation (so not all of your matches will result in new injuries, not sure what was Mourinho doing all season). And being lucky or not, results and confidence isn't exactly the worst thing to build further upon. But your performances will obviously need to improve if he's there to stay.
  7. Mar 7, 2019

    Stocar Full Member

    Joined:
    Dec 30, 2011
    Messages:
    576
    That's a very bold claim, that perhaps indicates a slightly impoverished view on the world.

    But I agree that it is foolish to disregard xG. It is quite flawed, and the beauty of football is that it is hard to translate into stats. There are no algorithms for the most interesting things in life. However, xG is probably the best statistical model out there, and often corresponds well with what is seen on the pitch.
  8. Mar 7, 2019

    Revan Assumptionman

    Joined:
    Dec 19, 2011
    Messages:
    38,288
    Location:
    Munich
    There is a very high probability that what he says is true though ;)
  9. Mar 7, 2019

    bosnian_red Worst scout to ever exist

    Joined:
    Aug 13, 2011
    Messages:
    36,517
    Location:
    Canada
    Caley graphics is one of the main accounts on twitter for xG. There was loads of BS tweets floating around so I went through each of our games since Ole has come in and made an excel file summarizing the xG for all of them, adding the .76 for a penalty to every game and then at the end rounding to the nearest .5 of a goal, and from there giving points total (win/draw/loss) based on that for each individual game (because doing differently is dumb).

    Here is a summary.

    Premier league
    12 games played
    29 goals for, 9 goals against, 32 points
    25.1 xG for, 12.5 xG against, 27 points

    Dropped points to Spurs(loss), Leicester (draw), Liverpool (draw), Palace (draw).

    FA Cup
    3 games played.
    United 2.46 - 1.0 Reading
    Arsenal 1.8 - 1.2 United (this one is very questionable how the shots got modeled anyway, but not a big deal)
    Chelsea 0.5 - 0.9 United

    Champions League
    United 0.3 - 1.9 PSG
    PSG 1.3 - 2.36
    Total: United 2.66 - 3.2


    As you can see, even the analysts on twitter are full of shit and arent being objective with their assessment of their own data, trying to push some bullshit agenda. Of course you need some luck when you go on a huge run. These are the raw numbers though. How people interpret them is up to themselves. Pretty clearly not a huge variance in the xG and the real values, certainly not as big as they try to claim.

    So in summary, OLES AT THE feckING WHEEL
  10. Mar 7, 2019

    Leftback99 Full Member

    Joined:
    Jan 11, 2015
    Messages:
    4,551
    If you flipped a coin the 'expected heads' (xh) would be 0.5. If you got 20 heads in a row it wouldn't make the 'xh' a nonsense stat despite expecting only 10. Xg is obviously far more complicated to put an exact figure like 0.5 on but people are dismissing it here in the same way.
  11. Mar 7, 2019

    Classical Mechanic Full Member

    Joined:
    Aug 25, 2014
    Messages:
    24,855
    Well, that or the fact that top teams have better players at the crucial moments.
  12. Mar 7, 2019

    shabadu84 Mint? Berry?

    Joined:
    Jan 12, 2010
    Messages:
    4,589
    Location:
    Muppet Treasure Island
    Ultimately, the basic xG/xA conclusion is that we are overperforming. Given our record under Ole, I don't think any of us could realistically expect us to carry that form over an entire season. That's not really that controversial. We probably can't expect Lukaku to score a brace every game, he doesn't have a track record to suggest he can sustain that level of over-performance. But we can probably also expect DDG to continue to outperform the xA because he does have the track record. These are obvious conclusions, it's just you apply numbers to it that people get in a tizzy.
  13. Mar 7, 2019

    bucky Full Member

    Joined:
    Apr 5, 2011
    Messages:
    8,080
    That's definitely true. That person from Pogue's OP is a bitter Arsenal fan for example.
  14. Mar 7, 2019

    FlawlessThaw most 'know it all' poster

    Joined:
    Oct 26, 2005
    Messages:
    27,393
    YeahI mean that when I say the top teams are ruthless.
  15. Mar 7, 2019

    Rish Sawhney New Member

    Joined:
    Mar 15, 2015
    Messages:
    378
    Location:
    State College
    That is such a bad analogy its hilarious.

    I wasn't aware that "good coin tossers" can consistently toss a head. The fact is that scoring goals is not a random event where each chance is taken in isolation. Good players will outperform their xG's where you really can't outperform an xH (for a coin toss) for long without being extremely lucky.

    However when skill becomes involved these probabilistic metrics lose their predictive capabilities. Its one thing to analyze a game in terms of xG and say "one side did or didn't take their chances, or one side didn't create enough" and a different thing completely to say "they're outperforming the xG, so we can't expect this to hold". In contexts where skill is involved, statistics aren't good metrics for predictions. Of course thats not to say that sometimes factors don't change. For example, you lose your star goalscorer (Costa for Chelsea) and now suddenly you're not outperforming your xG nearly as much and end up having a poorer season. That doesn't mean that "they regressed to the mean", it just means that they lost their star striker. But people will argue that they saw it coming because they outperformed their xG the season before. Thats straight up BS and cherry picking of data points.

    I think a lot of the criticism of xG isn't really a criticism of xG in general. Its a criticism of people who use xG to make stupid arguments like "a certain manager is over performing the xStats so they're not likely to hold it in the future". There's nothing about xG stats that says that, but its proponents keep using an analytic stat to make future predictions.
    Last edited: Mar 7, 2019
  16. Mar 7, 2019

    Paul_Scholes18 Full Member

    Joined:
    Sep 13, 2014
    Messages:
    2,938
    Since football is not just a lot of random chances though and you can't treat them all the same. Lots of mental things going on and context is very important. Like a big chance to win a game in the last minute is very different from being able to score a late 5-0 goal. Form and confidence will see player do much better than expected for spells of time. Being a few goals down you will chase the game and might create chances although not having the confidence to put them away.
  17. Mar 7, 2019

    Paul_Scholes18 Full Member

    Joined:
    Sep 13, 2014
    Messages:
    2,938
    I think we might struggle to keep it up in the league. Although it is mainly due to our injuries and having some hard games to come and not based on xG or xA. Still think we can raise our games in the cup in which we can play a bit more on the counter.
  18. Mar 7, 2019

    Treble Full Member

    Joined:
    Sep 11, 2008
    Messages:
    10,476
    You seem to think that the better the players, the bigger the difference between xG and reality. Which is wrong as City (are supposed to) have the best players but their xG stats are prety close to their actual results whereas Spurs have worse players than them but outperform xG to a much higher degree than City.
  19. Mar 7, 2019

    bosnian_red Worst scout to ever exist

    Joined:
    Aug 13, 2011
    Messages:
    36,517
    Location:
    Canada
    And hes trying to pass off a blatant lie as the truth because he thinks people wont look into it. He replied to other people saying penalties were included in the total expected goals when they actually weren't (and it's a pretty big difference when you included 7 penalties on top of the open play xG).
  20. Mar 7, 2019

    Treble Full Member

    Joined:
    Sep 11, 2008
    Messages:
    10,476
    It's uinrealistic to sustain that level unless we sign Messi and Neymar in the summer. Btw, Barca won with them "just" 94 pts. Only Juve have won 102 pts in a big league.
  21. Mar 7, 2019

    anant Full Member

    Joined:
    Feb 28, 2015
    Messages:
    5,085
    Personally, I love the xG metric. It gives a good indication of how sustainable some run is, like we saw with our xA numbers last season and Mou getting the boot this season.

    However, stats should be used to back up a point which has been formed based on what you see in games rather than opinions being formed due to a stat. This is the primary step, and the step where xG argument pretty much fails

    In addition to that, if we think of comparing these numbers with other teams in the league since Ole took over, we've scored 4 (2nd behind Spurs) more than expected and conceded 5 less than expected (2nd behind Newcastle). While in terms of GD, this is the highest, one needs to factor in that apart from Burnley game, we've not exactly trailed an opposition team for a long period, which means our xG is going to be lower per game as we're not taking shots for the sake of it and only a few of their shots against us are going to be speculative efforts. Add to that, once we include the Spurs game, a difference of 1.8 xGA was carved in our favour in the game vs Spurs alone only.
  22. Mar 7, 2019

    Rish Sawhney New Member

    Joined:
    Mar 15, 2015
    Messages:
    378
    Location:
    State College
    Its because they don't try to show their quality while shooting, their entire play is setup to create easy openings for their players. So the opportunities they do create are very clear cut with high xG's and their forwards usually take them. As opposed to other clubs like outs where the the chances created are more a matter of speed of play and taking split second chances to score.

    It emphasizes a difference in style of play certainly, but I never said "the better the players, the bigger the difference between xG and reality". The better the players, the more they have the ability to outperform their xG, thats true. But if your xG is fairly high, you don't really need to outperform it to win do you? The best example here is Aguero's xG, which is very similar to his actual goals. But you seem to think this is the common scenario for good players, whereas I think its the rare case. Most good players don't play at City under a very strict rotational postition system where they rarely take speculative shots and mostly wait for clear cut openings.

    "Which is wrong as City (are supposed to) have the best players but their xG stats are prety close to their actual results whereas Spurs have worse players than them but outperform xG to a much higher degree than City"

    No its not the good players that cause them to match xG with goals, its their system. Their system means thir players don't have to score difficult chances because their goals are usually tap from within the 6 yard box.
  23. Mar 7, 2019

    sunama Baghdad Bob

    Joined:
    Apr 26, 2014
    Messages:
    14,254
    I beg to differ. Liverpool fans care about the xG numbers. From what I understand, LFC fans only care more about xG, than actually winning matches and scoring goals.
    For me, I don't care about the xG statistic. I think it's pretty useless.
  24. Mar 7, 2019

    Treble Full Member

    Joined:
    Sep 11, 2008
    Messages:
    10,476
    Yet, there was a big gap between City's xG in 16/17 and their actual results. No need to pretend that you understand how something works when you don't have much of a clue.
  25. Mar 7, 2019

    Rish Sawhney New Member

    Joined:
    Mar 15, 2015
    Messages:
    378
    Location:
    State College
    I'm sorry but this is just a straight misinterpretation of what the stat actually measures. Its a fallacy of projecting past outcomes into the future when the context changes.
    Mourinho self imploded, and tactically left our defense both extremely exposed and while also not allowing us to attack properly. Its not about what can or cannot be sustained, but its a measure of how many clear cut shooting opportunities you create or concede. Three's no law saying that you can't be clinical with a lower xG and still win everything in sight (in fact teams playing a more pragmatic counter attacking style would).

    As a counter, see how Ole coming in has put us back in the position we were in last year of outperforming our xGA. Why is our defense not conceding all these goals then? De Gea has hardly been in God mode apart from the one Spurs game.
  26. Mar 7, 2019

    Rish Sawhney New Member

    Joined:
    Mar 15, 2015
    Messages:
    378
    Location:
    State College
    Yes because unlike you I can hold a bigger picture in my head than just raw numbers. What changed is that they got a better keeper and better fullbacks. This allowed them to not concede as many goals as they did (even though they had a fairly good xGA stat, they were underperforming because they didn't have a top class keeper). Had they kept Bravo, there would still be a big gap between City's xG in 17/18 and their actual results.

    Its not just a matter of time that your results and xG will match up. There's no guarantee of that.

    For reference: https://understat.com/team/Manchester_City/2016

    The biggest difference was between the xGA and the GA from open play.

    To put it back to you: No need to pretend that you understand how something works when you don't have much of a clue.

    I don't have anything against xG stats, they have their use. Its the over application into predictions without understanding the context that does my head in and reeks of lazy thinking.
    Last edited: Mar 7, 2019
  27. Mar 7, 2019

    Classical Mechanic Full Member

    Joined:
    Aug 25, 2014
    Messages:
    24,855
    Their xG in Pep's first season really said that they needed to sign a new goalkeeper, which they did. Although you didn't need xG to tell you that.
  28. Mar 7, 2019

    Schneckerl Full Member

    Joined:
    Jun 17, 2016
    Messages:
    2,518
    Wasn't City's goalkeeping extremly terrible for a while? Not sure if it's enough to have a noticeable effect like that, but might be a factor.
  29. Mar 7, 2019

    Treble Full Member

    Joined:
    Sep 11, 2008
    Messages:
    10,476
    That's way too simplisitic. The difference re xGA between 16/17 and 17/18 is 6 goals, not, say, 16. The bigger difference came from the attack, not from the defence albeit Ederson was a marked improvement indeed.
  30. Mar 7, 2019

    Leftback99 Full Member

    Joined:
    Jan 11, 2015
    Messages:
    4,551
    Yes like i said its far more complicated but for example:

    10 identical chances with an xg of 0.4 each:
    Average PL striker scores 10/10.
    Messi scores 0/10.

    Expected goals 4:4. Actual score 10:0.

    Messi's individual xg for those chances might be 0.7 but he could still miss them all.

    Expected goals 4:7. Actual 10:0.

    People think average PL striker is brilliant, xg is nonsense, Messi has lost it.
  31. Mar 7, 2019

    Treble Full Member

    Joined:
    Sep 11, 2008
    Messages:
    10,476
    No wonder, you seem bigheaded.
  32. Mar 7, 2019

    Rish Sawhney New Member

    Joined:
    Mar 15, 2015
    Messages:
    378
    Location:
    State College
    Jesus how old are you.. 12? You started with the digs but can't handle them when you're proven to be incorrect. Maybe take some basic probability classes and learn that just because city's stats look a certain way doesn't mean those stats are what a good team should be aiming for or that those are the holy grail. Stats are tools made by people usually with their own blindspots and biases with certain models in their mind. And then interpreted by completely different people who often have no idea of how stats and probabilities work.
  33. Mar 7, 2019

    Stocar Full Member

    Joined:
    Dec 30, 2011
    Messages:
    576
    Teams that consistently create the largest number of goal chances are those that do best in the long run (at least in league competitions). Although there is something to the fact that reactive teams are usually more efficient with their chances and are often overachieving in comparison to possession based ones.

    But it also seems that this feature of possession based approach to produce quality chances is more sustainable than the ability of reactive teams to "overachieve" due to maintaining a high level of efficiency. (This applies mostly to top teams, that maintain a level of quality during longer periods.)

    At the end of the day, xG actually does a decent enough job in estimating quality of chances. How much a team is good in making them count is mostly due to psychological factors. That's where reactive teams have an advantage, because with dominant football usually comes a sense of complacency. But in the long run, it's impossible to maintain a level of mental edge to surpass a team that will consistently create more and concede less chances.
  34. Mar 7, 2019

    Tiber Full Member

    Joined:
    Apr 22, 2014
    Messages:
    4,773
    Location:
    England
    This XG twattery annoys the feck out of me.

    For years I have rolled my eyes at NFL fans who spend more time looking at PFF spreadsheets than actually watching games. Now this same bullshit has crawled into football.


    This tweet (probably already posted) says it perfectly:

    Last edited: Mar 7, 2019
  35. Mar 7, 2019

    Treble Full Member

    Joined:
    Sep 11, 2008
    Messages:
    10,476
    Agree about that and pointed out earlier that they have important limitations. Disagree about your take though and what you wrote did little to convince me.
  36. Mar 7, 2019

    Rish Sawhney New Member

    Joined:
    Mar 15, 2015
    Messages:
    378
    Location:
    State College
    Thats debatable. Playing good football isn't about sustainability, its about consistently playing good football. If a "reactive" teams outperforms their xG due to having good forwards, whereas a possession based team creates better clear cut chances, the "high quality football" still has to take place somewhere.

    In a possession based team, that usually happens around elsewhere on the pitch where they open teams up with their quick movement and passing. However we've seen times where they (possession based teams) aren't quick enough in their passing, they end up not creating as much clear cut chances.

    Same for the "reactive" teams. The good plays have to happen somewhere on the pitch to open teams up and score goals. If its your forwards mostly doing that, you'll outperform your xG. If its your midfielders doing that with their movement, you'll have a higher xG and might not need to outperform it. But the dangers for "non-sustainability" are the same IMO in both approaches.
  37. Mar 7, 2019

    worldinmotion66 Full Member

    Joined:
    Jun 12, 2015
    Messages:
    1,538
    I've just been singing United songs all day at work, when I clearly should have been more concerned about xG.
  38. Mar 7, 2019

    Rish Sawhney New Member

    Joined:
    Mar 15, 2015
    Messages:
    378
    Location:
    State College
    My only opposition is to the "regression to the mean", "its unsustainable to outperform the xG" idea. If you hold on to that (I dunno as I responded to someone else and you responded to me), then you're wrong. But you have a right to be as wrong as you want in life. And most people are wrong most of the time about most things. So its not something to be pissed about that much.
  39. Mar 7, 2019

    Crashoutcassius Full Member

    Joined:
    Oct 10, 2013
    Messages:
    6,570
    Location:
    playa del carmen
    they might get expected goals to where it needs to be in 10 more years but its a bit of a joke right now
  40. Mar 7, 2019

    Treble Full Member

    Joined:
    Sep 11, 2008
    Messages:
    10,476
    OK, how often do teams consistently outperform the xG over 2-3 seasons? Examples in England (only top teams)?

    What is this xG tool good for, according to you? Does it show something interesting about United since Ole?