Babu:
Everything is fine.
Their avg. xG is 1.5 higher than of their opponent.
I wonder if this xG thingy factors in shit defenders?
xG Goal Expectancy Rate (a mentalist RAWK statistical spinoff thread) ft Babu of RAWK fame
- Thread starter Harold_Giles
- Start date
Their avg. xG is 1.5 higher than of their opponent.
I wonder if this xG thingy factors in shit defenders?
My word. He'd melt my fecking head if he was on here spouting all that bollocks.
https://understat.com/team/Liverpool/2017
Not that it needed clarifying, but they've been closer to matching the opposition than being that far ahead of them for most of the season anyway. See the xG and xGa at the bottom for the average over the season.
What the feck is xG? Sounds something like Klopp would drop before heading out to the touchline.
Another clinker from Baba Yega:
Edit: He doesn't actually compare him to Xavi but he does mention Lallana in the same breath as Isco, De Bruyne, Messi and Silva.
What a load of fecking drivel. fecking mentalist.Welcome to Liverpool Football Club Alex. You should really have done this back in 2011 and saved yourself a rough 6 years at Arsenal, but all is now forgiven
So what should we expect from our newest addition?
Part 1 - Why Alex Oxlade-Chamberlain
Firstly, it is said he lacks production - i.e. goals and assists. However, if your only measurement of attacking performance is goals ( or even goals and assists) then you will run into trouble pretty quickly when trying to understand the value and usefulness in important players. Let me give you an example.
Lukaku is brilliant at scoring goals. He is powerful and hard to stop. So what is the easiest way to stop him?
- Control the supply to him. Force him to receive in non-threatening areas only.
Okay so how easy is that? At Everton that pretty much meant stopping Barkley. How easy is that?
- Very easy. He is tactically one of the stupidest players I have seen. You can easily separate him from Lukaku and push him into non-effective areas and force him into taking risks in areas of little to no rewards for doing so.
So therefore if your best link player is Barkley, you can completely nullify an attack by simply nullifying Barkley which is remarkably easy to do. This is what happens when Liverpool play Everton. We simply cut the tether between those players and then shut down all their secondary supply lines to Lukaku. He starts drifting just to get touches on the ball and you have him trying to bully Lovren inside his own half on a 50/50 ball. Completely nullified.
In contrast, Liverpool. Mane is good at scoring goals. As is Firmino. And Salah. And Coutinho. How do you stop them all? At the moment our supply is out in Lallana & Coutinho, and yet we still make very dangerous things happen every single game with them both missing at the moment.
Another question - If Everton kept Lukaku, signed Seri to replace Barkley. How easy now would it be to cut the supply to Lukaku? Seri's production at Everton would likely be equal at best to Barkley. However, instead of someone who wants to be a star of the team worrying about his own numbers, Seri's entire focus is on the team. That Xavi-like game intelligence, movement and passing (albeit on a lower level ability wise).
Look at players in terms of a skillset rather than production. For a team to be effective, it needs a healthy mix of a variety of skillsets. For me, the most crucial one is players that can operate in tight spaces, offer penetration and an ability to destabilise a low block. We have lots of players who can score goals. Mane, Salah, Firmino, Sturridge, Origi, Solanke, Ings, Gini, Lallana, Coutinho - they all look good for 7+ goals per season. However, in terms of players who can link the front 3 to the rest of our team in a low block, we have Lallana, Coutinho then who? Woodburn maybe? Dhanda? Thus we end up needing to go around the block everytime. This worked against Palace, but the second your attacks have a distinctive pattern, you are easy to stop as we demonstrated with Lukaku above.
Oxlade Chamberlain is something we don't have. A Lallana/Coutinho type player with that penetration in midfield coupled with pace. It is that little burst of pace that makes Coutinho more effective than Lallana in terms of playing through the lines of a team. Lallana has the skills to beat people but not really to then burn away from them. Sure, Lallana & Coutinho are both better, more complete players than Chamberlain now. But he still has that skillset we are sorely lacking in, and a unique approach to it that we lack. Keita offers that same skillset (and a feckload of others) also. We need to add as many players as we can like this into our squad rather than finding reasons to just dismiss them, talk about signing DM's, more goalscorers, etc. Those things will not dictate how effective we are in the league. Our ability to break down a low block will. As Pep showed last season, and throughout his career, there are certain types of players who will help you in this pursuit. They look a lot like Chamberlain.
Part 2 - Can he unlock sides with creative passing?
I want to show the statistical output of Arsenal & Liverpool players in terms of their creative passing.
Key Passes(KP)/90 = This shows the number of goalscoring chances created per 90 minutes played. A goalscoring chance ranges from laying the ball off for a Coutinho shot 35 yards out or Firmino's cross to Can against Hoffenheim. Both = 1 Key Pass
Expected Assists(xA)/90 = This is the value of all goalscoring chances created per 90 minutes played. Effectively this is the likelihood of creating a goal every 90 minutes played. Laying the ball off to Coutinho to score a 35 yard screaming = xA 0,01. Firmino's cross to Can = xA 0,91.
xA/KP = This is the average value of the chances a player creates. A high score can mean the chances a player creates a typically good goalscoring chances. E.g. Welbeck created the least chances on this list, however the very few he created were good goalscoring chances. This should be looked at in conjunction with KP/90.
Assists/90 = This is actual output. How many assists were created on average per game played.
Anybody scoring orange/red in most if not all categories is a player who is highly involved in the teams creative output. Anybody who is red in multiple categories is showing an elite output. Anybody scoring very high in one category and low not in any others is likely an outlier and can be dismissed.
Player Key Passes(KP)/90 Expected Assists(xA)/90 xA/KP Assists/90
Mesut Özil 2,88 0,26 0,09 0,28
Philippe Coutinho 2,33 0,25 0,11 0,28
Roberto Firmino 2,05 0,23 0,11 0,21
Alexis Sánchez 1,9 0,25 0,13 0,28
Sadio Mané 1,57 0,18 0,11 0,2
Daniel Sturridge 1,52 0,19 0,13 0,12
Jack Wilshere 1,48 0,13 0,09 0,09
Alex Oxlade-Chamberlain 1,44 0,26 0,18 0,4
Alex Iwobi 1,42 0,14 0,1 0,18
Aaron Ramsey 1,39 0,32 0,23 0,29
Adam Lallana 1,34 0,16 0,12 0,27
Granit Xhaka 1,19 0,08 0,06 0,07
Jordan Henderson 1,19 0,11 0,09 0,17
Emre Can 1,18 0,08 0,07 0,08
Santiago Cazorla 1,16 0,12 0,1 0,29
Olivier Giroud 1,06 0,15 0,14 0,23
Georginio Wijnaldum 1,03 0,16 0,16 0,27
Divock Origi 0,8 0,1 0,12 0,19
Francis Coquelin 0,76 0,03 0,04 0
Theo Walcott 0,51 0,12 0,23 0,09
Mohamed Elneny 0,39 0,02 0,04 0,13
Danny Welbeck 0,36 0,14 0,39 0,24
Also wanted to show Dribbling on it's own because it's usually the most important element to breaking down a low block.
Player Dribbles/90
Alex Oxlade-Chamberlain 3,51
Alexis Sánchez 3,04
Philippe Coutinho 2,81
Sadio Mané 2,73
Jack Wilshere 2,72
Roberto Firmino 2,02
Daniel Sturridge 1,99
Theo Walcott 1,92
Alex Iwobi 1,72
Aaron Ramsey 1,46
Santiago Cazorla 1,31
Adam Lallana 1,27
Divock Origi 1,17
Mesut Özil 1,01
Georginio Wijnaldum 0,91
Francis Coquelin 0,91
Emre Can 0,8
Granit Xhaka 0,69
Mohamed Elneny 0,39
Danny Welbeck 0,36
Jordan Henderson 0,21
Olivier Giroud 0,15
So, statistically speaking, he is perfect for our system in that Lallana/Coutinho role. Or in a wide forward role as a creative hub rather than a goalscorer. Which is fine as we have fecking tonnes of players who can finish.
Part 3 - Who is he similar to? Ibe? Barkley? Coutinho?
Chamberlain vs Ibe
Similar or nothing at all alike other than running fast past people with the ball?
Player Key Passes(KP)/90 Expected Assists(xA)/90 xA/KP Assists/90
Alex Oxlade-Chamberlain 1,44 0,26 0,18 0,4
Jordan Ibe 1,53 0,09 0,06 0,00
My eye test tells me Ibe is ineffective as he only knows how to run fast in straight lines rather than attack dangerous areas. Therefore his use is limited other than a counter attacking player in behind teams. Against a low block he will simply run down it's sides and across it's front. Nothing to destablize the shape of it and no other tools to use to break it down to fall back on. All he has is dribbling + pace. Stats back that up. He scouts as someone who creates a high volume of shite chances which nobody turns into goals.
Chamberlain vs Barkley
Similar or nothing at all alike other than their ability to complete passes in the final 3rd being similar and how often they touch the ball in the box?
Player Key Passes(KP)/90 Expected Assists(xA)/90 xA/KP Assists/90
Alex Oxlade-Chamberlain 1,44 0,26 0,18 0,4
Ross Barkley 2,35 0,22 0,1 0,25
My eye test tells me Barkley is technically excellent but tactically crap. He takes up poor positions. Can easily be removed from the game by encouraging him to play in lower risk areas. Allows himself to become detached from Lukaku too easily (see Kane-Alli for an example how that COULD work). Barkley needs a coach like Rafa to educate him tactically and mould him into a more complete player the way Rafa did with Gerrard. This isn't to compare their relative abilities or skills as players, mere just to show how Gerrard's weakness was addressed early on by Rafa in his career turning a very talented average output player into a world class high output player. The stats above tend to back that up.
He scouts as someone who creates a high volume of average quality chances which an elite finisher turns into goals. Does that sound accurate? Dribbles about half as much as Chamberlain too.
Chamberlain vs Coutinho
Similar or nothing at all alike? I would say Chamberlain is similar in terms of penetration & creativity but not the goalscoring. Which we don't actually need him for at all. We have people for that already.
What all this shows for AoC is that he is at elite or near elite levels for dribbling, attacking passing, passes into the box and expected assists per 90. Also above average in throughballs, open play chances created and touches in the box. He also rarely loses the ball despite an exceptionally high amount of dribbling. This all points to a lockpick type player. He is doing things in tight areas to an extremely high success rate. What this doesn't show is his pace, which makes all of that stuff far more effective. But we can see things like the pace of a player with a quick scouting job on youtube.
So if you are imagining a player who would want to move between the lines of a side, or to destabilize a low block, and try to imagine what things you would want them to do at very high volumes successfully - your are basically listening the things AoC is at elite levels at statistically speaking in order.
1. Dribbling. (especially his high success rate at high volume)
2. Attacking Passing (indicates a high volume of forward passing high up the pitch with a high success rate)
3. Expected Assists (indicates his ability to get the ball into goalscoring positions consistently to the right players)
4. OP(open play) Passes into the box (How many passes he completes in the opponents box. Our two elite players in the squad in this metric are Coutinho & Lallana).
5. Throughballs - self explanatory
You give Klopp that player, add pace on top, that is going to feck teams up. Doesn't matter if he scores goals or not himself, that will ruin a low block. That is the 5 key numbers you would see if you looked at Silva, De Bruyne, Messi, Coutinho, Lallana, Isco... basically any player who would thrive against a compact defence. They offer other additional things too. But that is the platform that any lockpick needs to base the rest of their game around. Of course you need to scout the player too, just make sure those numbers align with what he does on the pitch. Chamberlain scouts very well also, as the video I will post below show also.
When you have someone putting up those numbers, and their every touch and compilation videos shows you the areas, consistency, pace and technique in which he is doing them, that is a player you can put to work in Klopp's system for sure in multiple roles.
Part 4 - I've seen the stats, but I don't know how that fits into our midfield
As mentioned, he has played a handful of times in central midfield, only two of which have every touch compilations. On of the problems he often encountered was he was usually played in a double #6 behind Ozil, or as an #8 beside Ramsey. Both players who have no discipline whatsoever and thus the midfield easily gets overrun. Plus he is a pressing player in a side who focus on shape for containment - as show against Bayern with his tantrum.
This is him playing as a #6 against Hull in a low block.
So compared to the Liverpool v Palace match you just watched some days ago, how would you describe the penetration of his passes from deeper areas? Doesn't do much dribbling here as he is a #6, the Henderson/Alonso role. Therefore it's more about passing through to the creative players than being one himself. However in that Palace game we failed to complete passes, or even touch the ball really, in zone 14 - the space central right in front of their goal. Instead we had to go round them in the absence of a player to go through them. While it worked that game, it does make our attack predictable and therefore unsustainable.
This is him playing as an #8 against Southampton. Some things to look at here in terms of facing a low block.
1) 0:56 plays through the midfield line centrally into zone 14. Something we did a grand total of 0 times against Palace.
2) Does it again at 1:17
3) 1:30 Long through ball resulting in goal.
4) 2:45 another chipped through ball
5) 3:00 Tactical dribbling. Losses the ball but the intent is to get around the midfield press then attack zone 14 to commit defenders. If you succeed at this 2 or 3 times per game, you are pretty much guaranteed 1 or 2 goals. This is a trademark Coutinho run.
6) 3:10 A diagonal forward ball from half space into zone 14. If the Arsenal player can escape the press here the entire Southampton defensive line collapses.
7) 3:22 - another long through ball this time splitting CB and FB. A toenail away from pulling it off too.
8.) 3:37 - another chipped through ball getting a player behind the defence on a counter.
9) 4:30 - dribble into the box, short through ball gets player behind the defence in the box ultimately resulting in a goal.
10) 5:20 - another chipped through ball over a low block in behind the defence in the box.
11) 6:07 - another through ball on the counter results in a 1-v-1
That is a long list of very penetrating touches. Even though it finished 5-0 in the end, he was doing those things at 0-0, set up the first goal, then continued doing them at 1-0 and onwards.
Watching that, can you see Klopp having a use for him in midfield? Do you see why many of us see him fitting in our midfield in Lallana's place during his injury and even long term given his age? And while yes, videos can make any player great, this is why you align it to stats to see if this output is something he produces consistently or a one off. The stats say he has consistent output in terms of creative passing, through ball, passes into the box, dribbles, etc.
This isn't cherry picking. He's literally played a handful of times in a central midfield last season and these are the only videos every touch that exist for him doing so. I cannot find any videos of his other 2 or 3 central midfield performances.
Summary
I like him.
I want to be clear though, I am not saying he is like, or a replacement for, Coutinho. I was demonstrating how he is a FAR more well rounded player than people give him credit for and closer to the level of Coutinho in that regard than the likes of Ibe or Barkley that have been trotted out as comparisons of "other shite English players".
They look nothing alike in terms of how they play when you watch them, however, the net result in terms of what they provide on the pitch in terms of breaking through the lines, playing in tight spaces, committing players in a low block, and creating good goal scoring chances for team mates are very similar.
He is obviously not without his weaknesses but then that is true of all players. We aren't trying to sign a 24 year old Steven Gerrard here. It's just a player with a skillset we need that is sorely lacking so far this season who will be useful for us to have going forward. He could grow into much more for Klopp, in the way Kevin Großkreutz did at Dortmund as a versatile player who was used more often than not. Who knows. It should be fun finding out with more meep meep than ever before in this side.
Edit: He doesn't actually compare him to Xavi but he does mention Lallana in the same breath as Isco, De Bruyne, Messi and Silva.
Last edited:
What in the name of Christ are those crop circle stat things?
If I posted on RAWK, it would be the first time I'd ever use "tl;dr"
I've no idea. It's way over my head anyway. One thing I know, though - OX is fecking shite!
I got halfway through that before realising wtf I was doing with my life.
Expected Goals is computed to turn each goal scoring attempt into a number between 0 and 1, representing the odds of that attempt producing a goal.
The easiest attempt to explain is a penalty.
Typically, penalties are awarded around 0.75 ExpG, based on historic conversion rate. A penalty is the easiest attempt to classify, since it’s a situation isolated from play, with a standard spot for taking it.
You can expand the number of situations and run separate regression models for other situations such as open play shots, headers, direct free-kicks, corners, throw-ins, fast breaks, rebounds from a keeper save, rebounds from the woodwork etc.
Essentially you can use a database (Opta) that will allow you hundreds of thousands of simulations to come up with an Expected Goal value for each situation.
And in a low scoring game like football, you can have a much xG than the opponent and not win. That's why in turn statistical models are not necessarily accurate predictors of match outcomes.
Last edited:
Don't confuse poor language skills with being erratic. There's not a lot of wrong with what he says.
What? Football is still a game, right?
Been considering explaining this for a long while but I only ever see it mentioned in the rawk thread so people will ignore it and still use it to laugh at Liverpool, which after all is what the thread is for.
I've gone right off the xG, the game's moved on
Last edited:
All of that xG stuff is complete nonsense.
I think they do want to start doing that,
<ponders throw-in distribution map & xTI data>
Yes, that's true, a chance is a chance.
In your system when you take into account the player itself, then every goal scoring chance presented to Welbeck is a zero, even if he's facing an open net. And then every shot on goal towards let's say Karius is a sitter because well he is a bad keeper in your view. It doesn't work that way. You evaluate the chance itself using parameters such as distance to goal, the angle, the pressure on the ball, the movement of the ball, the crowd in front of the ball etc, regardless of the players involved.
So essentially a great chance is a great chance regardless if it falls to Ronaldo or Crouch. Then, world class players perform well above the xG, the average ones perform exactly at the xG, and the poor ones perform below xG.
Last edited:
Anyways, there's value to statistical models and regressions in soccer, but not as much as in baseball or basketball, because in a low scoring sport like footie you can end up with huge variances which can distort data, and an unsophisticated observer will misrepresent or misinterpret the output.
Erm, yes it is, that's why Lukaku is better than Smalling at putting away chances.
Or if you create a chance against Burnley and the same exact chance against Bayern Munich, you apply the same finish, the Burnley one goes in, the Bayern one is saved by Neuer... would you blame the finishing? Or would you say the keeper made a great save against a shot that normally goes in.
Thread made by moving posts from RAWK.
Though, on a separate note, if any other posts derail the rawk thread, I might just lump them in here too. I'd like to understand this xG thing because I don't have a clue at the moment. Seems like an over-complicated way of making yourself feel better for not scoring because statistically you should have scored. It's creating a new reality almost. I could have it all wrong, of course.
Though, on a separate note, if any other posts derail the rawk thread, I might just lump them in here too. I'd like to understand this xG thing because I don't have a clue at the moment. Seems like an over-complicated way of making yourself feel better for not scoring because statistically you should have scored. It's creating a new reality almost. I could have it all wrong, of course.
.I love how Liverpool fans and their manager are arguing about "creating" loads of chances and then go on to talk about that one chance that Sturridge should have apparently put away. That chance was entirely created by an shocking defensive mistake and had nothing to do with Liverpool's buildup play.
I'm not sure you can add them together to try & say the team deserved 3 goals because of a load of imagined / measured fractions of goals that haven't been scored.
I think it does take into account the player now. As of this season.
Though doing so would make the prediction less reliable as they'd have tiny data sets to work from.
Lots of Opta Analysis is snake oil though. They preach in absolute but record in vagaries.
So is this Babu Yagu actually PhaseOfPlay in disguise? Very similar style, and RAWK and their mods are well-known for having multiple profiles.
He made some brave predictions about us finishing 6. again, Ox setting the league alight and of course Lukaku flopping.
Looks like he will be hounded by Christmas, just like the fellow PoP
He made some brave predictions about us finishing 6. again, Ox setting the league alight and of course Lukaku flopping.
Looks like he will be hounded by Christmas, just like the fellow PoP
No one says this surely?
It's pretty foolish to use on a single match basis IMO. It's a statistic that's best over a larger sample.
In Liverpool's case, if they keep underperforming xG in attack as they have this season that indicates a desperate need for a clinical finisher to be added. This was an issue for United last year and one that adding a player clinical like Lukaku has helped, without even accounting for the impact he's had on our overall play.
Wow a thread about Baba yoghurt’s erudition.
xG is football analysis gone bonkers.
I think essentially it's an attempt at quantifying and visualizing performance beyond counting statistics by applying quality as a factor. Sometimes that can reveal things you didn't know, which can help address weaknesses or identify talent. We already do some of this ourselves when we say things like "half-chances" or "missing a sitter". This is a good read on it.
http://www.telegraph.co.uk/football...does-show-man-city-should-win-premier-league/
Often just tells us what we already know:
http://www.telegraph.co.uk/football...does-show-man-city-should-win-premier-league/
Often just tells us what we already know:
But I think there's something potentially useful to it, even if that is just highlighting to Liverpool that they're even more hopeless than they think.
Aren't the BBC MOTD using that xG metric as part of their analysis as well? This maybe originated from there?
https://www.google.com/amp/www.bbc.co.uk/sport/amp/football/40699431
called the 80s?
It actually isn't as bad as people make out. It's a detailed way of the old saying managers 'I'll start to worry if we're not making chances'
xG attempts to quantify the number of chances and quality of chances one creates against the other team.
But, I do agree with you that context is important.
eg. we're leading 3-0, we drop a gear in intensity in the last 10 minutes, the opposition creates 4 chances without scoring.
xG would be useless in this context, because the game is practically over and racking up chances in the end without scoring is kinda meaningless. Maybe if the opposition scored, the manager would have made some tactical changes, subs, etc to tighten up. As it is, he's just letting the game peter out.
on the other hand, if this was a hard fought 0-0 with both teams going at it until the end, yes xG would be a useful illustrative tool.
Oh yes. International break in full swing.
We drew a lot of games that we should have won last year. xG quantifies that notion.
The craziness continues to astound me, there seems to be no level they cannot reach.
Seems pretty straightforward to me, as long as you don't invest it with more meaning than it claims to provide.
Then again, some posters still seem to have a problem with those alternative PL tables so xG probably has a way to go.
Then again, some posters still seem to have a problem with those alternative PL tables so xG probably has a way to go.
From what I know xG works off an average given a goal scoring opportunity. This means that if over the long term you are scoring less than your xG then you are below average in one or more aspects of attack (finishing, creating tap-ins etc.) Similarly, if you are conceding more than the opponents xG over the long term then your team has defensive problems.
Regression to the mean cannot in most circumstances be applied to a single team; because there is a natural variance between each team. There is also natural variance between players; its perfectly normal for a team to outperform their xG in one season and do the opposite the next season. Only the best teams can regularly outperform their xG.
Regression to the mean cannot in most circumstances be applied to a single team; because there is a natural variance between each team. There is also natural variance between players; its perfectly normal for a team to outperform their xG in one season and do the opposite the next season. Only the best teams can regularly outperform their xG.