Other Games 2024-25

[Kevin Carr's Gloves 4424]

14 minutes ago, Sonatine said:

 

 

Not only does he look like a cunt, he looks like a wrong cunt as no one with half a brain still thinks Edison invented the lightbulb.

[PaddockLad 19784]

5 minutes ago, The Fish said:

 

Not understanding it is and not caring is fine. Not understanding it and thinking it's useless is patently ridiculous behaviour.

Xg data for the European cup semi final last night brought to you in conjunction with not only the likes of Saka who missed an open goal in one of the most highly pressurised environments in the sport itself but seemingly 1000s of lads who had a career like this lad…

 

Then we get to the individuals who actually do the data gathering; those stout fellows who judge how difficult any given shot is at any given game throughout football, all issued with a set of criteria that they interpret having gained their relatively brief football knowledge almost exclusively through a computer simulation of the game itself. Am not suggesting I could do better btw, am suggesting most people would be unable to do it accurately and keep up with everything else that’s going on in the game that they may or may not be assessing.  
 

Those are my issues with xG, all context at the top level appears to be removed and at its very basic level data collection is very likely to be arbitrary and partial. I understand the football world uses it and good luck to them but it appears it’s most vocal adherents don’t ask any questions about it at all. Has there been any studies on how accurate it actually is? 
 

 

[Monkeys Fist 46579]

1 hour ago, Kevin Carr's Gloves said:

Not only does he look like a cunt, he looks like a wrong cunt as no one with half a brain still thinks Edison invented the lightbulb.

Joseph Swan- top lad. 
 

 

[The Fish 11464]

15 minutes ago, PaddockLad said:

Xg data for the European cup semi final last night brought to you in conjunction with not only the likes of Saka who missed an open goal in one of the most highly pressurised environments in the sport itself but seemingly 1000s of lads who had a career like this lad…

 

Then we get to the individuals who actually do the data gathering; those stout fellows who judge how difficult any given shot is at any given game throughout football, all issued with a set of criteria that they interpret having gained their relatively brief football knowledge almost exclusively through a computer simulation of the game itself. Am not suggesting I could do better btw, am suggesting most people would be unable to do it accurately and keep up with everything else that’s going on in the game that they may or may not be assessing.  
 

Those are my issues with xG, all context at the top level appears to be removed and at its very basic level data collection is very likely to be arbitrary and partial. I understand the football world uses it and good luck to them but it appears it’s most vocal adherents don’t ask any questions about it at all. Has there been any studies on how accurate it actually is? 
 

 

That's not what happens though. 

 

All they do is look at the chance (not the shot) and compare it to all the other chances that are sufficiently similar. e.g. a penalty. Of the thousands and thousands of penalties taken, how many end up in the back of the net? Works out to be around 76% of them. For less distinct instances automated video capture is used, plus some human input. They don't say "Ooooh I reckon we should add .002 because there was a crisp packet rustle in the second row. They just say what happened in that chance and the model says this is how often that chance ends up in a goal. Player 35yds out, on the volley, weaker foot, while under pressure is going to be fuck all xG. Player on the goal line, with control of the ball, under no pressure is going to be very high xG. Not because of what that chance sounds like or how it's interpreted, but because looking back at what has happened for those types of chances, this is how rarely or how often the ball ends up in the goal. 

 

 

[PaddockLad 19784]

1 minute ago, The Fish said:

That's not what happens though. 

 

All they do is look at the chance (not the shot) and compare it to all the other chances that are sufficiently similar. e.g. a penalty. Of the thousands and thousands of penalties taken, how many end up in the back of the net? Works out to be around 76% of them. For less distinct instances automated video capture is used, plus some human input. They don't say "Ooooh I reckon we should add .002 because there was a crisp packet rustle in the second row. They just say what happened in that chance and the model says this is how often that chance ends up in a goal. Player 35yds out, on the volley, weaker foot, while under pressure is going to be fuck all xG. Player on the goal line, with control of the ball, under no pressure is going to be very high xG. Not because of what that chance sounds like or how it's interpreted, but because looking back at what has happened for those types of chances, this is how rarely or how often the ball ends up in the goal. 

 

 

Well I’ve understood all that since you’ve mansplained it dozens of times because you think I don’t understand it. 
 

That the  degree of difficulty of the chance is the same in the parc des princes in the last 10 of the European cup semi final as it is at the most basic level of local semi pro football doesn’t make sense. Context isn’t taken into consideration so how can the data be accurate? The human factor has to be allows for, which of course is utterly impossible. 
 

Why did Saka miss the open goal? What was the xG of the chance? Why is the xG of the chance the same as it is at West Allotment Celtic? It is indisputably a more difficult chance at the parc des Prince last night… 
 

 

[Howmanheyman 38140]

1 hour ago, Sonatine said:

 

 

 

 

What a fucking knacker. No surprise, but what a bull's Clem that bloke is.

[The Fish 11464]

4 minutes ago, PaddockLad said:

Well I’ve understood all that since you’ve mansplained it dozens of times because you think I don’t understand it. 
 

That the  degree of difficulty of the chance is the same in the parc des princes in the last 10 of the European cup semi final as it is at the most basic level of local semi pro football doesn’t make sense. Context isn’t taken into consideration so how can the data be accurate? The human factor has to be allows for, which of course is utterly impossible. 
 

Why did Saka miss the open goal? What was the xG of the chance? Why is the xG of the chance the same as it is at West Allotment Celtic? It is indisputably a more difficult chance at the parc des Prince last night… 
 

 

 

Well, because it averages out. The ability of Saka, the coaching he's had, the support he's had to manage pressure situations, the experience he's got of playing in high stakes games, versus the fatigue, the speed of the game, the weight of the pressure, the height of this particular stake. And the law of big numbers. That chance he missed was valued at 0.78xG, which means if that chance happened 1000 times, you'd expect the ball to hit the back of the net 780 times. This was just one of those 220 times in a thousand that it didn't. 

 

Let that chance fall to some clogger from Morecambe he's more likely to shank it that Saka is, but you'd still be in the stands saying "My nan could have finished that", because you know that if you watched 1,000 chances that were pretty much the same, the vast majority of them would be buried, regardless of who it fell to. 

 

We've all seen world class stars fuck up a relatively easy chance, and we've seen fucking clogger spaff one in against all odds. But because xG looks at thousands and thousands of data points it can ignore the ability of the player and look instead at the chance itself. It doesn't need to worry about the pressure of the stakes, or the BMI of the the player. It's the chance that matters. Not the competition, not the stadium, not the player... just the chance.

[Gemmill 52889]

Feels good to be officially retired from having to explain this shit AGAIN. 

[Gemmill 52889]

PL not yet retired from asking questions about xG and responding with an indignant "I'LL THANK YOU NOT TO MANSPLAIN, KIND SIR!" 

Edited 2 hours ago by Gemmill

[PaddockLad 19784]

8 minutes ago, The Fish said:

 

Well, because it averages out. The ability of Saka, the coaching he's had, the support he's had to manage pressure situations, the experience he's got of playing in high stakes games, versus the fatigue, the speed of the game, the weight of the pressure, the height of this particular stake. And the law of big numbers. That chance he missed was valued at 0.78xG, which means if that chance happened 1000 times, you'd expect the ball to hit the back of the net 780 times. This was just one of those 220 times in a thousand that it didn't. 

 

Let that chance fall to some clogger from Morecambe he's more likely to shank it that Saka is, but you'd still be in the stands saying "My nan could have finished that", because you know that if you watched 1,000 chances that were pretty much the same, the vast majority of them would be buried, regardless of who it fell to. 

 

We've all seen world class stars fuck up a relatively easy chance, and we've seen fucking clogger spaff one in against all odds. But because xG looks at thousands and thousands of data points it can ignore the ability of the player and look instead at the chance itself. It doesn't need to worry about the pressure of the stakes, or the BMI of the the player. It's the chance that matters. Not the competition, not the stadium, not the player... just the chance.

It’s a harder chance for Tubby Brewster last night for precisely the same reasons though. So why is his data being included in the xG? I get the large sample size explanation but Tubby Brewster is never going to be in the European cup semi final.  Context absolutely counts . His data is irrelevant in the context of last night. So that to me means xG as a reference point is fundamentally flawed. 

[PaddockLad 19784]

29 minutes ago, Gemmill said:

PL not yet retired from asking questions about xG and responding with an indignant "I'LL THANK YOU NOT TO MANSPLAIN, KIND SIR!" 

I’ve told you before not to butt in when the men are talking  

Edited 1 hour ago by PaddockLad

[Holden McGroin 8091]

How have they managed to analyse the exact same scenario of shot to get the result though (0.78 xg in this example)? Do you they also clock the speed of the pass and what foot (stronger or weaker) of the player it falls to? If that cross was mirrored from the right and it fell to his left side then I'd expect him to have more of a chance of scoring.

 

I appreciate its a stats tool used but its clearly not accurate.

[LondonBlue 1072]

when Saka missed that chance i had horrific flashbacks of Sterling doing the same for us

[LondonBlue 1072]

32 minutes ago, Holden McGroin said:

How have they managed to analyse the exact same scenario of shot to get the result though (0.78 xg in this example)? Do you they also clock the speed of the pass and what foot (stronger or weaker) of the player it falls to? If that cross was mirrored from the right and it fell to his left side then I'd expect him to have more of a chance of scoring.

 

I appreciate its a stats tool used but its clearly not accurate.

 

 

this is pretty much why i hate modelling data. i can be a little obsessive with perfection. let me just tweak the model and run it through again.

 

 

 

[Craig 7520]

36 minutes ago, Holden McGroin said:

How have they managed to analyse the exact same scenario of shot to get the result though (0.78 xg in this example)? Do you they also clock the speed of the pass and what foot (stronger or weaker) of the player it falls to? If that cross was mirrored from the right and it fell to his left side then I'd expect him to have more of a chance of scoring.

 

I appreciate its a stats tool used but its clearly not accurate.

As with all stats you can manipulate the presentation to suit whatever narrative you like. Which is why I won't give it any airtime. 

[Gemmill 52889]

The Top Gear crew thoroughly dismantling the whole idea of "electric vehicles" itt. 

[Howmanheyman 38140]

Just now, Gemmill said:

The Top Gear crew thoroughly dismantling the whole idea of "electric vehicles" itt. 

 

I've an EV and it's great and cheap to run so Clarkson and co can swivel. 

[Gemmill 52889]

I couldn't agree more.

 

[wykikitoon 23898]

[The Fish 11464]

1 hour ago, PaddockLad said:

It’s a harder chance for Tubby Brewster last night for precisely the same reasons though. So why is his data being included in the xG? I get the large sample size explanation but Tubby Brewster is never going to be in the European cup semi final.  Context absolutely counts . His data is irrelevant in the context of last night. So that to me means xG as a reference point is fundamentally flawed. 

The chance isn't harder. The chance is the same. It's the executor of the chance that's different. 

You can tac on explanations why that chance wasn't taken and include pressure or lack of ability, but the chance is what the xG model is reporting.

 

The model isn't there to explain why the chance was or wasn't taken. It's there to say that team A created more good opportunities to score than team B, but they didn't spend money on a striker so they're out of the Champions League.

Player A has fuck all assists but has been creating really good chances for his team to score from, it's just that his teammates are useless. 

Player B is scoring more goals than you'd expect given the chances that come to him, if he's good he'll keep doing it, if he's just been lucky it'll regress to the mean.

Team C has conceded the fewest goals so far, but that's due to their 'keeper pulling off world class saves, and profligacy from their opponents

 

 

 

1 hour ago, Holden McGroin said:

How have they managed to analyse the exact same scenario of shot to get the result though (0.78 xg in this example)? Do you they also clock the speed of the pass and what foot (stronger or weaker) of the player it falls to? If that cross was mirrored from the right and it fell to his left side then I'd expect him to have more of a chance of scoring.

 

I appreciate its a stats tool used but its clearly not accurate.

 

Not sure if they clock the speed of the pass (it's definitely doable and I'd argue that that's valuable information). They definitely record which foot (stronger or weaker) it comes to.

 

Quick google and yes, they do consider the speed of the pass.

Edited 40 minutes ago by The Fish

[Dazzler 12708]

[Dazzler 12708]

[Kevin Carr's Gloves 4424]

Xg is a pointless statistic for individual games, it’s meant for long term analysis of a team but is supposed to be used alongside other stats. There are various Xg models and they can all result in different outcomes but only by small margins. It also has inherent limitations like different skill levels, luck (good and bad) and the dynamic nature of the game. 
 

In short both sides of the argument should really stop going on about it being fucking useless/the best thing in the world.