[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip / qa] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/sp/ - Sports

[Advertise on 4chan]


Thread archived.
You cannot reply anymore.


[Advertise on 4chan]


File: 83726.jpg (68 KB, 1080x498)
68 KB
68 KB JPG
Why is xG so fucking useless?
>>
>>107343641
Because it's not necessarily going to correlate well for any one particular game. But over a larger data set it should correlate very well. To help test this, since you brought it up, find a team that has a better xG (or xP) over the past two Champions League seasons.
>>
xG should probably be an interval instead of a single number but then you'd call it useless for giving Bayern somewhere between 1 and 6 expected goals.
>>
>>107343641
Because it makes brainlet seethe.
>>
>>107343641
Because it's easy to be a prophet of the past.
>>
The average /sp/ poster becomes more retarded every single year so I can't even tell if this is bait or not
>>
>brainlets still unable to grasp xG
>>
>>107343641
xG has little to do with the outcome of a game
xG has little to do with the outcome of a game
xG has little to do with the outcome of a game
xG has little to do with the outcome of a game
xG has little to do with the outcome of a game

Why is it so hard for you retarded brainlets to get that through your heads?
>>
i think someobdy watchd that movie with brad pitt and the fat fuck, he took it seriously while it's a comedy
>>
>>107343641
They should rename it to expected shot difference, make it a ratio of team A / team B and then people wouldn't seethe over it
>>
File: data-nonces.png (416 KB, 662x644)
416 KB
416 KB PNG
>>107343641
>>
>>107345820
>guys trust me team A will win because they got higher xG
>team A gets rekt
>actually xG has little to do with the outcome of a game ok? dont be a brainlet

this is every retard on /sp/ these days
>>
>>107345871
Except I don't say those first two lines at all, you fucking retard
>>
>>107343641
Because it clearly under-estimates and over-estimates certain play styles, and assumes a consistent performance pattern from every striker, when some players excel in certain areas and are below average in others.

There's a reason this stat isn't used by top level sports scientists, because it's a basic brainlet stat
>>
>>107345820
Why is xG a thing then?
>>
>>107345915
Yeah really from all exemple of it i've seen it's just say xPossession
>>
>>107343641
xCocksInOP'sMouth (3.0)
>>
It is not useless. Just because it doesn't tell the story one every ten games doesn't mean it is always wrong. You are just a victim of being biased to your own conviction that it is wrong, only remembering games that fits your view of xG. Look over 100 randomly picked games and see if xG is right more times than not.
>>
>>107345915
>Because it clearly under-estimates and over-estimates certain play styles
Which is why no one barring a few retards uses it as some definitive proof of which team should have won. This doesn't mean that it's useless. Surely you can see the difference between Big Sam's West Brom having 0.1 xG vs Liverpool having 0.1 xG?
>>
guys, check this out i invented new stat. it's called xGS (goals scored). teams on average score 1,5 goals per game while man city constantly overperforms this stat by scoring 2,0+ goals per game. see how retarded that is, that's exactly what xG is, it takes averages and doesn't account for quality of the shot. i could have said that bayern had more chances by just looking at shots yesterday by each team. i don't need xG for that. it's a stat for midwits.
>>
>>107345915
>There's a reason this stat isn't used by top level sports scientists, because it's a basic brainlet stat

I need to find that lecture from a Liverpool FC sports scientist from a couple years back, it was crazy the depth of the stats they actually use to analyse players and teams. Absolutely puts to shame the midwits who have Squawka on their favourites bar and spam shit about xG and xPoints
>>
>>107345918
you can use it to rate a performance. If someone is outperforming xG it means he's good. If you look at the op's picture it tells you how the game looked. Bayern had a lot of chances but they failed to convert it or/and PSG was good in defence.
>>
>>107345915
>There's a reason this stat isn't used by top level sports scientists
but it is. Literally every professional team uses it and different variations of it.
>>
File: 1255094341449.jpg (5 KB, 200x200)
5 KB
5 KB JPG
>>107343641
xg is to football analysis what mcdonalds is to cuisine
>>
>>107346286
They most likely also have dozens or hundreds of other stats you don't hear about because their name doesn't sell as well as "Expected Goals"
>>
>brainlets dickriding mbappe
>xG chads know he got shit on and will lose the next game
>>
>>107345918
Why is passes completed a thing and literally shown on screen during matches, or why is posession relevant to anyone but pep? Why do assists matter when often they were piss easy and the crucial pass was the one prior, or a la biscuits glorious messi assist, it was fucking nothing after which the scorer did everything?
Every statistic is stupid, but xG is actually slightly less so. It does a very good job of showing how well a team actually played, regardless of whether they won or lost.

When Sharteta was getting his early wins at Arseanal, his xGs showed that it was fluke wins and eventually that comes to catch up with you.

xG doesn't mean much in a single game, but if a team consistently creates fuck all chances, they will inevitably start to lose games and drop points. It's basically an early indicator of trouble brewing, when a team is winning it might not look like it, but if their xGs is constantly lower than oppositions, that will eventually show in them losing a lot too.
>>
>>107346344
don't move your goalposts, mate
>>
4.14 xg without lewa to convert.. bros.. maybe xg is right..
>>
File: 1612799890551.jpg (1.11 MB, 3000x2713)
1.11 MB
1.11 MB JPG
Results are random, perfomances aren't.
>>
>>107346371
High IQ
>>
>>107346233
Bumping out of interest
>>
>>107346371
>dribble past the keeper
>net is open, player is 4 meters from the line
>pass it into the net with ease, could score the goal 999 times in a row
>0.73xG because there was a defender within 10 meters of him
>>
why do thick retards hate xG?

why is it so difficult to understand for them?
>>
File: 3_t2xcer46[1].jpg (43 KB, 615x409)
43 KB
43 KB JPG
It could be 8 xG but with this Stoke City legend there I wouldn't expect anything
>>
>>107346272
>or/and PSG was good in defence
polish intellectuals everyone
>>
fraudulent statistic
>>
>>107346371
Except strikers aren't random number generators. Some strikers are just better than others, but xG is an objective stat and does not account for the individual player. A forward (and their team) could consistently overperform vs xG because they're better at finishing than the average player.
>>
>>107345915
yeah the style of creating good goal scoring chances. launching it at every opportunity wont net you 4 xg
>>107346351
pretty much this
>>107348074
0.73xG is because from what ive seen it uses mean of two: position of the shot and defender pressure. once there is more data so instead we could use three dimensions: x and y for position and z as defender pressure to acces specific data value.
>>
>>107348153
You can tell from his comment he didn't watch the game and bases his analysis, of which half is correct, on a single stat.
Although you could get the same analysis just looking at the shots made per team
>>
Because footballs dependent on betting retards who can't into stats

Teaching plebs actual data analytics would endanger the models the betting sites use

xG is an improvement on "muh possession" that people used to use with Barca a few years back.

I'm sure at some point in time people will start to understand xG a bit deeper and we'll start seeing adjusted xG being reported

Examples of adjusted xG would take in to account something like;
two footedness of player
finishing ability of player

The other problem with xG is it assumes that the score is 0-0 and each team has equal incentive to score throughout the game. Let's consider the following match situation;

team A scores a penalty in minute 1
team A proceeds to play with a low block for 90 minutes, limiting their opponent to lots of low % but perhaps a high frequency of shots

In this way, they generate a high xG, and a high number of shots.

Arguably in this instance, the model is overestimating low probability shots...

but the point still stands that team A will only ever show an xG of 0.95 (or whatever the penalty conversion ratio) - but this is to do with their gameplan and strategy being based around being in the lead.

This is why comparing xG across the season doesn't make sense - because of the variation of the gamestate within a football match
>>
xG is a slightly more precise stats than shots and possesion, nothing more.
>>
>>107343641
Bayern should have scored more desu
>>
Datanonces ruining more sports per usual, many such cases!
>>
>>107348437
You can complain about literally any stat in football using this logic, posession, shots on target or whatever else.
>>
because navas
>>
>>107346371
Unironically this.

You can have 9 players intercepting every ball, winning every tackle and creating chances from every forward pass - but if you have proverbial Werner and Havertz finishing the goalscoring chances, you'll still draw 0-0 at best.
>>
xG gods punished bayern, too often they overperformed their xG, it was time for them to get punished to balance things out.
>>
>>107348074
>0.73xG
bullshit
>>
Unexpected goals are innately superior to expected goals. Many of those expected goals are blocked/saved in reality because, as the name says, they are EXPECTED.
Sun Tzu was right: the greatest asset to an army or team is the element of surprise.
The thinking man’s metric is not expected goals, but UNEXPECTED goals.
>>
>>107348540
If chelsea wins the cl, will havertz and werner be considered chelsea legends?
>>
>>107348525
this also, great goalkeeper have the ability to suck in balls, they enter in the players head and they try to overshoot or undershoot due to the higher pressure, they lose their cool.
>>
>>107348581
no, the guy mixing up and picking up the balls during the CL drawing match ups will be the chelsea legend
>>
>>107348604
They did get atletico first. That was not an easy matchup in paper at all.
>>
>>107348118
It's meaningless and we unironically know better
>>
When they introduce this to FIFA stats at the end of games tthe seething will be even more insane
>>
>>107348118
It's the other way around bud. There are plenty of anons ITT who know their stuff pointing out genuine problems with xG. Meanwhile, the proponents of xG are going full Dunning-Kruger and think they're smart for saying "xG is a long-term stat, it is irrelevant for individual matches".
>>
It's literally just a better measure of chances created than shots on target
that's all it's supposed to do
>>
>>107348770
then why do anoraks pretend like it's solved football?
>>
Predictive statistics in soccer have the lowest correlation with game outcomes out of any major sport.
>>
>>107348817
source
>>
>>107346371
based ZEMANLandia poster
>>
>>107348786
Nobody claims that xg does this except from the people who want to moan about it
>>
>>107348634
atletico and porto can't even play in their home stadium, yikes
>>
>>107343641
I don't think they took in consideration Lewy's injury when calculating xG.
Brainlets who can't understand xG always are trying to diminish it's effectiveness when predicting the outcome of a game.
>>
>>107345915
>>107345932
this

xG retardedly underrates counter attacks even it's if it's more likely to score from that than a slow tikakaka attack
>>
Stats that matter:
Wins
Losses
Draws/Ties (in gay sports that have them)
Simple as
>>
>>107348770
tell that to the brainlets on this board
>>
stats are for losers
>>
>>107349343
>>107349363
The duality of frog
>>
Anti-xG brainlets are hilarious
>>
Brighton have single-handedly deboonked xG this season
>>
>>107346233
i wanna watch too
>>
xG is irrelevant when you have choupo moting instead of an actual striker and the opposition keeper is navas
>>
your example shows its usefulness though?
>>
Letting nerds in to sports has been a disaster.
>>
>>107348144
But he scored
>>
>>107349394
The Egyptian Football Understander has logged on
>>
>>107343641
Buyern dominated the game, but PSG finished all their chances. It's literally simple as that.
>>
>>107350747
bayern 6 big chances, psg 4 big chances



Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.