[Albion] 'Enough with the data obsession, in the real world only goals for and against count'

[Pavilionaire]

Football is all about goals but ultimately it is also decided by money. The clubs with financial reach will - all things being equal - continue to dominate.

What gives the clubs outside of the elite some hope - in our case the other 14 teams in the Premier League - is that they can use data better than the big boys to close the gap.

Data - and intelligent use of it - is the ONLY hope the Albion have of ever challenging the Top 6.

What xG tells us is that the rest of the team are creating chances for our forwards, and it also tells us how xG compares across our forward group. This then helps with recruitment.

You could sign a prolific young forward in a minor European league that has a hatful of goals, but he might waste a hatful of chances in the process. Therefore, it might be better to sign a striker who scores 10 goals from 20 chances, rather than a striker that scores 25 goals from 75 chances.

 

[loz]

I have no idea about XG but have read this with interest....Now it might just be me but when a side out of the top six go a goal up or even two goals up, they do seem to sit back and let the other side attack and therefore the side with more chances and possession might well be the losing side ? and I think that tends to bear out looking at the stats...unless you are a top side confident in winning I would think the likes of Burnley, Palace etc are quite happy to go one up and try to defend.

I may well be wrong but I am not to sure we should have league tables of XG, although it may well be useful at looking at individual player performances and how they can improve their game.

 

[Wozza]

Data - and intelligent use of it - is the ONLY hope the Albion have of ever challenging the Top 6.

What xG tells us is that the rest of the team are creating chances for our forwards, and it also tells us how xG compares across our forward group. This then helps with recruitment.

I don't think anyone, anywhere would argue against data being used professionally/behind the scenes.

 

[CheeseRolls]

I don't think anyone, anywhere would argue against data being used professionally/behind the scenes.

Good thread this with a tasty opener.

 

[Dr Q]

"

Its very simple for those who think only results (and.. possession?) matters or is of any interest. They could just stare at the scoreboard for 90 minutes and then **** off. No point in watching someone kick the ball if you only care about the scoreline.

by that analogy though, wouldn't you spend 90 minutes waving to Potter?

 

[Stato]

For those that haven't had enough of the data obsession this Twitter thread is interesting: https://twitter.com/BetweenThePosts/status/1365967799064158208

Particularly this bit explaining that the huge discrepancy between our expected and actual goals scored and goals conceded is purely at the Amex. I told you, someone needs to wee in all four corners!

[tweet]1365976818294136834[/tweet]

By the way, has anyone got access to the Jim White article, that started this thread, without a subscription?

 

[Shooting Star]

As a fully paid up member of the data nerd society, I love stats. Its also the secret of Tony Bloom's success so should really be embraced by all Albion fans.

That said, our analytics team must be scratching their heads when trying to deciphier our numbers.

It seems our possession stats are all counter intuitive.

<50% possession in 10 games
4 wins
3 draws
3 losses
Average points per game 1.5
Average shots per game 11
Average Goals per game 0.8

>50% possession in 16 games
1 wins
8 draws
7 losses
Average points per game 0.69
Average shots per game 15
Average Goals per game 1.12

>60% possession in 4 games
0 wins
2 draws
2 losses
Average points per game 0.5
Average shots per game 20
Average Goals per game 0.75

So in games when we have the least possession and the fewest shots we earn the greatest number of points.

Many thanks for researching this as it makes for absolutely fascinating reading.

For me, this is the clearest indictment that the style of football we play, not just the perceived striker problem, is somewhat at issue. In light of these stats, I wish the most zealous Potter-Inners could concede that GP should consider tweaking (not overhauling) his style of football in order to get results, just as I’m sure they wish the most zealous Potter-outers would concede that the strikers are a significant part of the issue. If the free-flowing way Leeds, Villa and the top six-sides approach playing us works in our favour, surely GP needs to try and force the hands of other opponents to do similar. Otherwise we need another way to break them down.

If the two-third’s of games we’ve played with over 50% possession have only materialised into ONE win, is it not obvious that we should be looking for pointers from the third of games we had less position than the opposition, but racked up 4 out of our 5 wins from?

Inviting pressure clearly gives our strikers a better chance of scoring - or at least the team a better chance of winning, according to those stats. I think back to the Villa game where we were under the cosh for large portions, but scored two excellent free-flowing goals. For all the xG we have, how often do we find ourselves with chances like those, where there are few defenders in the box?

I concede I’m not 100% sure how xG works, but my instant reaction to the fawning over it is that if we took the lead (and maintained it), surely our xG would be lower in those games as GP’s style sees us take the foot off the gas when we take the lead (consciously or subconsciously). Isn’t it natural that the games where we have very high xG are the ones where we’re chasing the game, because opponents are happy to sit back and stifle us? So it’s probably a fallacy when people argue that we could have won such-and-such a game 4-0 based off of chances. If we took the lead, we’d probably have fewer shots per game.

 

[andy1980]

Many thanks for researching this as it makes for absolutely fascinating reading.

For me, this is the clearest indictment that the style of football we play, not just the perceived striker problem, is somewhat at issue. In light of these stats, I wish the most zealous Potter-Inners could concede that GP should consider tweaking (not overhauling) his style of football in order to get results, just as I’m sure they wish the most zealous Potter-outers would concede that the strikers are a significant part of the issue. If the free-flowing way Leeds, Villa and the top six-sides approach playing us works in our favour, surely GP needs to try and force the hands of other opponents to do similar. Otherwise we need another way to break them down.

If the two-third’s of games we’ve played with over 50% possession have only materialised into ONE win, is it not obvious that we should be looking for pointers from the third of games we had less position than the opposition, but racked up 4 out of our 5 wins from?

Inviting pressure clearly gives our strikers a better chance of scoring - or at least the team a better chance of winning, according to those stats. I think back to the Villa game where we were under the cosh for large portions, but scored two excellent free-flowing goals. For all the xG we have, how often do we find ourselves with chances like those, where there are few defenders in the box?

I concede I’m not 100% sure how xG works, but my instant reaction to the fawning over it is that if we took the lead (and maintained it), surely our xG would be lower in those games as GP’s style sees us take the foot off the gas when we take the lead (consciously or subconsciously). Isn’t it natural that the games where we have very high xG are the ones where we’re chasing the game, because opponents are happy to sit back and stifle us? So it’s probably a fallacy when people argue that we could have won such-and-such a game 4-0 based off of chances. If we took the lead, we’d probably have fewer shots per game.

These are the stats we should be looking at alongside xg. There could be any number of reasons why this is happening and instead of saying our strikers are rubbish we should be seeing if we can find out why. What is for sure is that this season when we have less possesion we are more likley to win. Therefore until we are safe then that is what we should do.

Those stats also indicate that when we have more possesion we concede more goals. Therefore it suggest a lack of concentration from our defence. If we want to play possesion Football then our defence has to be better. Until then less possesion works better.

People often use the cliche goals wins you games. In fact scoring more goals than the opposition wins you games. As a possesion based team our goals against must go up more than our goals for.

 

[Johnny RoastBeef]

For those that haven't had enough of the data obsession this Twitter thread is interesting: https://twitter.com/BetweenThePosts/status/1365967799064158208

Particularly this bit explaining that the huge discrepancy between our expected and actual goals scored and goals conceded is purely at the Amex. I told you, someone needs to wee in all four corners!

[tweet]1365976818294136834[/tweet]

By the way, has anyone got access to the Jim White article, that started this thread, without a subscription?

Interesting descrepancy. Probably just a coincidence but the Amex pitch is a few yards longer than most other Premier League pitches. A tenuous stat, but only Molineux, The Etihad and Old Trafford are as long and we've never won under Potter at any of those stadiums..

 

[London Irish]

"Brighton have become unwitting poster boys for a way of looking at the game which does not focus on goals scored or conceded, nor possession."

Its funny he mentions possession as it kind of reveals him: its not about "results are the important thing", its about "these stats did not exist when I was young and therefore they do not matter".

There's always been ways of looking at the game which does not focus on results / goals in or out. Read some old shite from the 1960s and they would talk about George Best dribbling, go forward a few years and they would talk about Leeds playing style, or the Netherlands total football in the 1970s World Cups. Its part of what makes it interesting.

Its very simple for those who think only results (and.. possession?) matters or is of any interest. They could just stare at the scoreboard for 90 minutes and then **** off. No point in watching someone kick the ball if you only care about the scoreline.

George Best won the European Cup, Leeds won the FL title and Holland reached two World Cup finals. They were talked about because they won football games and lots of them. It’s never been unimportant, ever

 

[Danny Wilson Said]

George Best won the European Cup, Leeds won the FL title and Holland reached two World Cup finals. They were talked about because they won football games and lots of them. It’s never been unimportant, ever

And Leeds were Champions of Europe, don't forget. Which is the title-winning equivalent of our xG achievements.

 

[BadFish]

As a fully paid up member of the data nerd society, I love stats. Its also the secret of Tony Bloom's success so should really be embraced by all Albion fans.

That said, our analytics team must be scratching their heads when trying to deciphier our numbers.

It seems our possession stats are all counter intuitive.

<50% possession in 10 games
4 wins
3 draws
3 losses
Average points per game 1.5
Average shots per game 11
Average Goals per game 0.8

>50% possession in 16 games
1 wins
8 draws
7 losses
Average points per game 0.69
Average shots per game 15
Average Goals per game 1.12

>60% possession in 4 games
0 wins
2 draws
2 losses
Average points per game 0.5
Average shots per game 20
Average Goals per game 0.75

So in games when we have the least possession and the fewest shots we earn the greatest number of points.

But the one factor that makes sense of the numbers is that all the other teams, outside the tradition top six, have a Goals per shot average of 0.09 or roughly a goal every 11 shots.

Brightons Goals per shot average is 0.06, which equates to a goal every 16 shots.

The really worrying thing about that stat is that we have only manged 16 shots in 8 matches out of 26 this season.

Thankfully, for all of Graham Potter's talk of not wanting new signings and wanting to develop the players he has, (presumably bullshit aimed at not damaging our strikers already low confidence) you can be sure that behind the scenes Tony Bloom and his army of data nerds will be scouring the planet for our next undervalued star striker.

There was a time (and it coincided with some good results) where I thought we looked like we had morphed into a somewhat counter-attacking side. My reading of your stats here suggest that this in fact a better way to score goals. i.e. when there are less defenders in the box we have more time and space to carve out a decent shooting opportunity.

My frustration with some of our play under Potter is we seem to hold on to possession so much that it gives the opposition time to get back in formation and get players behind the ball. Our players then do not have the technical ability to breach a well set up defence. It has been mentioned over and over that we do not have the players to play 'potterball' and I think that this is borne out by both your stats and our results (I guess this is obvious).

This comes back to what I said in the first paragraph, there seemed to be a time where we moved the ball through the thirds much quicker and it looked like it was going to pay dividends.

Potterball in its purest form will not succeed with our current group of players because:

- Those teams that put 10 behind the ball cannot be broken down because we are not technically proficient enough to calve out good enough chances or finish the half chances.
- We are more suited to teams that play possession football and have enjoyed a few good results. However on the whole we are playing teams with better players than ours who will more often create and convert chances. So we will not win the lion's share of these.

In short, we struggle to beat the teams around us because we can't break them down and we struggle to beat the teams that are better than us . . . because they are better than us.

I would like to see us explore the counter attacking game that we seem to have somewhat abandoned - is this because we have lost both Solly and Lamptey on the flanks and their replacements are not blessed with the greatest of pace?

 

[andy1980]

There was a time (and it coincided with some good results) where I thought we looked like we had morphed into a somewhat counter-attacking side. My reading of your stats here suggest that this in fact a better way to score goals. i.e. when there are less defenders in the box we have more time and space to carve out a decent shooting opportunity.

My frustration with some of our play under Potter is we seem to hold on to possession so much that it gives the opposition time to get back in formation and get players behind the ball. Our players then do not have the technical ability to breach a well set up defence. It has been mentioned over and over that we do not have the players to play 'potterball' and I think that this is borne out by both your stats and our results (I guess this is obvious).

This comes back to what I said in the first paragraph, there seemed to be a time where we moved the ball through the thirds much quicker and it looked like it was going to pay dividends.

Potterball in its purest form will not succeed with our current group of players because:

- Those teams that put 10 behind the ball cannot be broken down because we are not technically proficient enough to calve out good enough chances or finish the half chances.
- We are more suited to teams that play possession football and have enjoyed a few good results. However on the whole we are playing teams with better players than ours who will more often create and convert chances. So we will not win the lion's share of these.

In short, we struggle to beat the teams around us because we can't break them down and we struggle to beat the teams that are better than us . . . because they are better than us.

I would like to see us explore the counter attacking game that we seem to have somewhat abandoned - is this because we have lost both Solly and Lamptey on the flanks and their replacements are not blessed with the greatest of pace?

It Will be intetesting to see goals against in these contexts. I suspect the answer might pie in that.

 

[Guinness Boy]

Interesting. I wonder if the reason is teams are coming to the Amex to park the bus more than before?

Interesting descrepancy. Probably just a coincidence but the Amex pitch is a few yards longer than most other Premier League pitches. A tenuous stat, but only Molineux, The Etihad and Old Trafford are as long and we've never won under Potter at any of those stadiums..

Absolute classic examples of how statistics can tell you WHAT but not WHY.

 

[Audax]

Many thanks for researching this as it makes for absolutely fascinating reading.

For me, this is the clearest indictment that the style of football we play, not just the perceived striker problem, is somewhat at issue. In light of these stats, I wish the most zealous Potter-Inners could concede that GP should consider tweaking (not overhauling) his style of football in order to get results, just as I’m sure they wish the most zealous Potter-outers would concede that the strikers are a significant part of the issue. If the free-flowing way Leeds, Villa and the top six-sides approach playing us works in our favour, surely GP needs to try and force the hands of other opponents to do similar. Otherwise we need another way to break them down.

If the two-third’s of games we’ve played with over 50% possession have only materialised into ONE win, is it not obvious that we should be looking for pointers from the third of games we had less position than the opposition, but racked up 4 out of our 5 wins from?

Inviting pressure clearly gives our strikers a better chance of scoring - or at least the team a better chance of winning, according to those stats. I think back to the Villa game where we were under the cosh for large portions, but scored two excellent free-flowing goals. For all the xG we have, how often do we find ourselves with chances like those, where there are few defenders in the box?

I concede I’m not 100% sure how xG works, but my instant reaction to the fawning over it is that if we took the lead (and maintained it), surely our xG would be lower in those games as GP’s style sees us take the foot off the gas when we take the lead (consciously or subconsciously). Isn’t it natural that the games where we have very high xG are the ones where we’re chasing the game, because opponents are happy to sit back and stifle us? So it’s probably a fallacy when people argue that we could have won such-and-such a game 4-0 based off of chances. If we took the lead, we’d probably have fewer shots per game.

So ... if you're a data nerd, you don't just stop at picking out the possession stats and whether or not we won those games. You also start looking deeper: how did the opposition play that day? Did we have high possession because the other team wasn't interested in having possession except on the counter? Or did we have high possession because that was the way we set up to play? Or a combo of both? Who were we playing when we had high possession, and who when we had low? Without looking at the numbers, I'd suggest our low possession matches that we won were against teams higher than us in the league who didn't feel the need to park the bus (Liverpool and Leeds stand out) and tried to outplay us. The low-possession wins against Liverpool and Leeds games were both away fixtures, in a period where we hit a purple patch. Neither opponent renowned for parking the bus at home. We also had a relatively (vs the high possession games where we've not taken our chances) lower possession home win vs Tottenham, another team not really known for parking the bus against lower opposition.

That xG Home vs Away stat stands out to me. I'm wondering if that's a stat that's held true throughout the season, or one that's emerged more recently as a result of the Palace and Villa games where we were dominant but just couldn't make it count.

 

[CheeseRolls]

Absolute classic examples of how statistics can tell you WHAT but not WHY.

Even then there are people looking for hidden meaning.

The poorer teams in the division, have worked out we struggle to break down well organised defences. They have been happy to let us have possession and try and fashion scoring opportunities, knowing we will likely struggle and they stand a far better chance of getting something from the game, than if they try and play an open game, where we can match some of the best sides in the league.

One overly long sentence, no stats, no magic insight.

Guess what? A focal point for our attacks, who can score more from half chances, bully defenders, force teams to double up, will also create space and better opportunities for others in the team and take the pressure off.

There you have it folks no stats we need a quality centre forward, which is stating the bleeding obvious.

 

[blue-shifted]

There was a time (and it coincided with some good results) where I thought we looked like we had morphed into a somewhat counter-attacking side. My reading of your stats here suggest that this in fact a better way to score goals. i.e. when there are less defenders in the box we have more time and space to carve out a decent shooting opportunity.

My frustration with some of our play under Potter is we seem to hold on to possession so much that it gives the opposition time to get back in formation and get players behind the ball. Our players then do not have the technical ability to breach a well set up defence. It has been mentioned over and over that we do not have the players to play 'potterball' and I think that this is borne out by both your stats and our results (I guess this is obvious).

This comes back to what I said in the first paragraph, there seemed to be a time where we moved the ball through the thirds much quicker and it looked like it was going to pay dividends.

Potterball in its purest form will not succeed with our current group of players because:

- Those teams that put 10 behind the ball cannot be broken down because we are not technically proficient enough to calve out good enough chances or finish the half chances.
- We are more suited to teams that play possession football and have enjoyed a few good results. However on the whole we are playing teams with better players than ours who will more often create and convert chances. So we will not win the lion's share of these.

In short, we struggle to beat the teams around us because we can't break them down and we struggle to beat the teams that are better than us . . . because they are better than us.

I would like to see us explore the counter attacking game that we seem to have somewhat abandoned - is this because we have lost both Solly and Lamptey on the flanks and their replacements are not blessed with the greatest of pace?

I agree

The answer involves making the final pass or cross earlier. This means being prepared to risk the loss of possession and having our attack countered by the opposition.

I actually think we have the players for a decent 4-4-2 team. With the midfielders, told to get it wide early, to Lamptey and Trossard, who are told to slide the ball in fast and low first time, or very early, to two strikers, who are told to scream into the box every time the midfield get it.

Though I don't see this being GP's style. We'd certainly concede more.

 

[Johnny RoastBeef]

Understat.com have a table comparing the xG stats for all the formations Potter has used this season. The two most regularly used are 3-4-2-1 (used 916 minutes) and 3-4-1-2 (used 434 minutes). It shows that although there is only a slight difference between the formations, 3-4-1-2 is so far producing much better numbers than 3-4-2-1, both expected and actual.

3-4-2-1
Goals for 10 (0.98 per 90)
Goals against 12 (1.18 per 90)
xG90 1.2
xGA90 1.27

3-4-1-2
Goals for 6 (1.27 per 90)
Goals against 4 (0.84 per 90)
xG90 1.91
xGA90 0.89

 

[vagabond]

Don’t know if this is the best thread to post this amazing snippet in, but someone asked Maupay on one of his game streams about xG. And he said in a pissed off matter of fact way: “xG?? xG is bullshit”.

 

[Bakero]

Don’t know if this is the best thread to post this amazing snippet in, but someone asked Maupay on one of his game streams about xG. And he said in a pissed off matter of fact way: “xG?? xG is bullshit”.

No surprise there. Why would he be a fan of a stat that suggests he's not clinical enough?