Newcastle United 4-1 Manchester United - 13/04/2025

[Cf]

1 hour ago, Kaizero said:

 

xG is inherently flawed, as is all means of statistical analysis of football in general. If you, or anyone else, is interested in learning more about why that is, I've uploaded a Tifo video that explains it in much greater detail (and more understandably) than I could dream of doing:

 

 

 

That video doesn't make the conclusion you come up with at all.

 

Even with xG it says "it's far from perfect and hard to calculate compared to other sports".

 

The central premise that there are so many moving parts to football, and loads of permutations of how things will play out is a good one. But it doesn't follow that suddenly all stats are inherently flawed. Just that they're harder to usefully calculate and interpret. In other words you'll need some very skilled people to do so.

 

The thing I like about xG is it brings the concept of variance into the mainstream. People say it's a results based business but this is often taken too far.

 

"Eddie Howe got it spot on today and Newcastle move into the European spots"

 

"Eddie Howe got it totally wrong and Newcastle find themselves in the bottom half of the table"

 

Give us the ability to rewind time and replay a match and those are two perfectly normal headlines you'll see for the exact same setup. In one reality we converted that chance early on. In another we didn't and let the other team into the match.

 

Also apparently YouTube/The Algorithm/Me doesn't like Europe. Bit harsh on Salah too. In fact wtf is going on here 

 

[LionOfGosforth]

11 minutes ago, Nucasol said:

Me watching back how half arsed Eriksen was for the Barnes 2nd goal.

 

 

It was almost sad watching him. Been a great player in his career but trying to compete in there with our midfield 3 is a thankless task when you can barely run. There were times i'd see us up the pitch and idly think "where's their midfielders?" and then Eriksen would amble just into view, like a little old man trying to run with teenagers. 

[Big River]

30 minutes ago, joeyt said:

Seemed to be loads of slips yesterday, felt like the pitch had been watered too much

 

Murphy made a beeline for the groundsman at HT to tell him how wet it was

 

they were watering our right wing for almost all of half time. must have been something we had identified.

 

small margins. we operate on a totally different level these days. all down to the gaffer!

[Kaizero]

39 minutes ago, Hanshithispantz said:

Do you think they put xG on MotD for scouts like? 

 

xG is a vaguely accurate marker that people can use alongside possession, SoT, shot map's etc to get a general gist of how a game went. 

 

"Shots", "Shots on Goal", "Wide Shots" and "Saves" are much more relevant stats for a viewer to get a general gist of how a game went. "xG" is meaningless because it's based on nothing but general vibes and it is absolutely not "accurate", as the unimaginable number of unique variables involved that all would have to be calculated and compared with a non-existent "baseline" for an identical chance the xG is being calculated for that would serve as the basis for either a 0.0 xG or a 1.0 xG. Needless to say, those baselines do not exist – hence the number means absolutely nothing.

 

 

"This chance made me feel worried, 0.4?"

 

 

For the record, whilst I vehemently disapprove of "xG" as a metric symbolizing anything whatsoever – I am a massive supporter of analytical data in general. 

 

 

 

Edited Monday at 13:56 by Kaizero

[Kaizero]

16 minutes ago, Cf said:

Also apparently YouTube/The Algorithm/Me doesn't like Europe. Bit harsh on Salah too. In fact wtf is going on here 

 

 

   

 

...you know your suggested videos is based on your own youtube history...right? 

 

 

 

[Dr Venkman]

Well, this thread went to shit in record time after a great victory.

[Hanshithispantz]

7 minutes ago, Kaizero said:

 

"Shots", "Shots on Goal", "Wide Shots" and "Saves" are much more relevant stats for a viewer to get a general gist of how a game went. "xG" is meaningless because it's based on nothing but general vibes and it is absolutely not "accurate", as the unimaginable number of unique variables involved that all would have to be calculated and compared with a non-existent "baseline" for an identical chance the xG is being calculated for that would serve as the basis for either a 0.0 xG or a 1.0 xG. Needless to say, those baselines do not exist – hence the number means absolutely nothing.

 

Understat's xPTS is accurate to 1 win for 12 teams in the league, 16 are within 2 wins. The major outlier is Forrest (+14pts) which is what anyone who watches footballer would guess if you asked the question. 'Vaguely accurate' is more than fair, professional gamblers are obviously not using it but it's fine enough for fans.

 

 

Edited Monday at 14:02 by Hanshithispantz

[Froggy]

22 minutes ago, Dokko said:

 

Well sell him for more than that in the summer to Saudi, and no one can say shit, that's his market value. 

 

Romano just reporting Saudi interested in Onana. 

 

Please. 

[Nucasol]

Just now, Froggy said:

 

Romano just reporting Saudi interested in Onana. 

 

Please. 

Wish our overlords would stop bailing out these profligate fuckwit competitors of ours. First Chelsea and now your lot.

[Kaizero]

18 minutes ago, Cf said:

That video doesn't make the conclusion you come up with at all.

 

Even with xG it says "it's far from perfect and hard to calculate compared to other sports".

 

The central premise that there are so many moving parts to football, and loads of permutations of how things will play out is a good one. But it doesn't follow that suddenly all stats are inherently flawed. Just that they're harder to usefully calculate and interpret. In other words you'll need some very skilled people to do so.

 

The thing I like about xG is it brings the concept of variance into the mainstream. People say it's a results based business but this is often taken too far.

 

"Eddie Howe got it spot on today and Newcastle move into the European spots"

 

"Eddie Howe got it totally wrong and Newcastle find themselves in the bottom half of the table"

 

Give us the ability to rewind time and replay a match and those are two perfectly normal headlines you'll see for the exact same setup. In one reality we converted that chance early on. In another we didn't and let the other team into the match.

 

FWIW, I used to not mind xG before I entered the world of true footy data rooms after coughing up an insane amount of money to give the agency I started access  Can't look at xG the same way ever again. Did you know there's actual analytical stats that actually utilize relevant datasets to inform you whether a player fluffed a chance or not, as well as how badly they fluffed it if they did? Yet, because they need to have something to make people cough up 3k+ annually for access to, they throw xG at the general public instead of actually useful and informative stats?  

 

 

[andycap]

If they're after onana it must be for entertainment purposes, try and get the punters through the door with his slapstick silent movie routines. 

[Kaizero]

3 minutes ago, Hanshithispantz said:

Understat's xPTS is accurate to 1 win for 12 teams in the league, 16 are within 2 wins. The major outlier is Forrest (+14pts) which is what anyone who watches footballer would guess if you asked the question. 'Vaguely accurate' is more than fair, professional gamblers are obviously not using it but it's fine enough for fans.

 

 

 

 

Happy to hear you're using Understat, it's the only free-to-use site I've seen that actually offers up a couple of useful stats (though rather simplified variants, still).

[PauloGeordio]

 

[Dokko]

12 minutes ago, Froggy said:

 

Romano just reporting Saudi interested in Onana. 

 

Please. 

 

Biggest save of his career if he pulls that off.

 

Sure he said casemiro was off there last summer, probably more hope than anything else. Same goes for us, Saudi bailout just hasn't happened. 

[Hanshithispantz]

21 minutes ago, Kaizero said:

 

Happy to hear you're using Understat, it's the only free-to-use site I've seen that actually offers up a couple of useful stats (though rather simplified variants, still).

Someone mentioned 'momentum' a little while back as being pointless, but it's sort of the same thing. They're simplified data points that broadcasters and score apps can use to give the viewer/user a vaguely accurate reflection of a game.

 

If Man City get beat off Forrest 0-1, and they have an xG of 0.4, 70% possession with 15 shots at goal, I can assume they were forced to pass across the box all game taking pot shots. If the xG was 2.4, I could assume Forrest were probably lucky. You see the highlights on MotD, and can even check the shot and heat maps if you're feeling like a true James Lawrence Alcott, but it's obviously nowhere near what companies like StatsBomb sell.

 

 

Edited Monday at 14:28 by Hanshithispantz

[Solitude20]

40 minutes ago, Kaizero said:

 

"Shots", "Shots on Goal", "Wide Shots" and "Saves" are much more relevant stats for a viewer to get a general gist of how a game went. "xG" is meaningless because it's based on nothing but general vibes and it is absolutely not "accurate", as the unimaginable number of unique variables involved that all would have to be calculated and compared with a non-existent "baseline" for an identical chance the xG is being calculated for that would serve as the basis for either a 0.0 xG or a 1.0 xG. Needless to say, those baselines do not exist – hence the number means absolutely nothing.

 

 

"This chance made me feel worried, 0.4?"

 

 

For the record, whilst I vehemently disapprove of "xG" as a metric symbolizing anything whatsoever – I am a massive supporter of analytical data in general. 

 

 

 

 

 

xG is definitely the best single metric. It tells you if your team is creating real scoring opportunities and not just shooting for the sake of it. It also helps evaluate performance beyond just the final score.
 

The final score is just one possible outcome in a game full of probabilities. Sometimes, the result reflects the performance; other times, it doesn’t. That’s where xG (expected goals) steps in — to give context.

 

A team might win 1–0 with one low-quality chance (xG = 0.1) while the other team had 3 big chances (xG = 2.3) but failed to convert. Every shot is a probability of scoring a goal, and there’s randomness to it. Barnes scored a wonderful third goal yesterday, but the exact same chance could repeat next week and he might hit the post instead. 
 

It helps answer the key coaching question:

“Did we play well enough to win, regardless of whether we did?”

 

Coaches and analysts love xG for exactly this reason. it separates process from outcome. It’s certainly not the only metric we should look at, as we should use it along with xGA, progressive passes, possession, shots, etc.  

 

 

 

Edited Monday at 14:36 by Solitude20

[Cf]

26 minutes ago, Kaizero said:

 

   

 

...you know your suggested videos is based on your own youtube history...right? 

 

 

 

 

The fact I shared it in the first place suggests otherwise 

 

Looks nothing like my YouTube home. I played it embedded on the forum. So is it taking the forum's stats into account?

[Kaizero]

9 minutes ago, Solitude20 said:

xG is definitely the best single metric. It tells you if your team is creating real scoring opportunities and not just shooting for the sake of it. It also helps evaluate performance beyond just the final score.
 

The final score is just one possible outcome in a game full of probabilities. Sometimes, the result reflects the performance; other times, it doesn’t. That’s where xG (expected goals) steps in — to give context.

 

A team might win 1–0 with one low-quality chance (xG = 0.1) while the other team had 3 big chances (xG = 2.3) but failed to convert. Every shot is a probability of scoring a goal, and there’s randomness to it. Barnes scored a wonderful third goal yesterday, but the exact same chance could repeat next week and he might hit the post instead. 
 

It helps answer the key coaching question:

“Did we play well enough to win, regardless of whether we did?”

 

Coaches and analysts love xG for exactly this reason. it separates process from outcome. It’s certainly not the only metric we’ve should look at, as we should use it along with xGA, progressive passes, possession , shots, etc.  

 

 

In an ocean of wrong, the statement I've bolded is the worst offender as that's genuinely just not the case. At least not when you get to a semi-professional level. NPxG, xGC and xGBU are much more valuable stats for those positions if we have to play with xG stats, and even then we'd just be scratching the surface as you will find genuinely insightful xG metrics in the data pool – but with endless more layers and nuance than the average xG stat you'll see in the post-match stats section of the BBC (or similar).

 

 

Edited Monday at 14:38 by Kaizero

[Kaizero]

11 minutes ago, Cf said:

 

The fact I shared it in the first place suggests otherwise 

 

Looks nothing like my YouTube home. I played it embedded on the forum. So is it taking the forum's stats into account?

 

Are you logged into your Youtube account on your browser? (and is that browser Chromium based?) If so, then it'd still be your algorithm throwing up the suggestions. If not, then it's probably random based on the video context pulled by Youtube.

[Cf]

Just now, Kaizero said:

 

Are you logged into your Youtube account on your browser? (and is that browser Chromium based?) If so, then it'd still be your algorithm throwing up the suggestions. If not, then it's probably random based on the video context pulled by Youtube.

 

Nah not logged in. And if I load YouTube up the video doesn't show on my history. 

[Kaizero]

Regardless, this isn't the thread for an xG debate. If people want to continue, let's head to the not worthy of a thread thread or the football pet hates thread  

[Kaizero]

1 minute ago, Cf said:

 

Nah not logged in. And if I load YouTube up the video doesn't show on my history. 

 

Will be Youtube doing it on their own then, most likely based on what they perceive the context of the video to be.

 

Either way, that suggested videos line-up you got served was top tier  

[RodneyCisse]

40 minutes ago, Nucasol said:

Wish our overlords would stop bailing out these profligate fuckwit competitors of ours. First Chelsea and now your lot.

My only hope is that it’s all

part of a plan, for us to flog stuff off down the line.

[Ginola]

36 minutes ago, PauloGeordio said:

 

The weight of pass from Isak and finish from Tonali are both sublime. Could, and probably will, watch this goal over and over 

[Elma]

1 hour ago, Big River said:

 

they were watering our right wing for almost all of half time. must have been something we had identified.

 

small margins. we operate on a totally different level these days. all down to the gaffer!

 

Keegan used to do the same. Soon as the players went in after the warm up he'd have them pile water onto the pitch before home games, and then we'd come out expecting it and the opposition would be slipping all over for 10-15 mins.

 

Interesting if we're targeting specific areas of the pitch these days.