Next manager

[boyyanno]

Hilarious. Genuinely creasing from that.



Well yeah, that's using it wrong. If someone is parroting single game sum of xG and providing no context then they don't understand it.

If you understand it and watch the matches then it can be really useful and a good indicator/predictor of performances. Hence how I was able to predict exactly what has happened over the last 7 games after the Leeds, to great derision by the "xG nerds don't know football like the man down the pub does" crowd (amongst others).

Man says "unless performances improve results won't get better"- thinks he's a genius.

Same man says "I predicted 7 games ago what would happen" looks at the league table, was 2nd, still 2nd.

You're fast becoming a bit of a loon on here.

[Rileybobs]

It is possible to be both happy with the results and position in the table and recognise that neither is sustainable if we continue to perform as we currently are.

Why is it unsustainable? Is xG a better indicator than points gained over 12 games?

[willsclarets]

This thread is a sad indictment of football and where we're at imo. It's beyond ludicrous.

[Tall Paul]

Why is it unsustainable? Is xG a better indicator than points gained over 12 games?

Yes

[Rileybobs]

Yes

How so?

[Mattster]

Why is it unsustainable? Is xG a better indicator than points gained over 12 games?

We scored 9 goals and picked up 6 points from the first 2 games. Since then we've scored 8 goals in 10 games and are bottom 3 in the league for creating goalscoring opportunities whilst we've just had the easiest (on paper) run of fixtures we'll have all season, dropping points and failing to score from open play against 4 of these.

Since the last time we faced a side in the top half of the table our points per game has dropped from 2 to 1.85. Goals per game has dropped from 2.2 to 0.85. We are now going to start coming up against more top half sides, there's an arguement that this will favour us and I hope that's the case but I'm not optimistic.

[Tall Paul]

How so?

http://web.archive.org/web/201811121758 ... ted-goals/

tl:dr

[Petersa]

Why is it unsustainable? Is xG a better indicator than points gained over 12 games?

Don't fully understand xG.....getting on a bit now but still beleive that if you have more points than teams below you, you are above them in the table and that could be relevant at the end of the season

[Rileybobs]

http://web.archive.org/web/201811121758 ... ted-goals/

tl:dr

I'm not a statistician so I'm finding it quite hard to understand at what point of a season actual points gained becomes a more accurate indicator of final league position (or total points accrued if you prefer) than xG? And, the chart you've shown takes into account xG against, which is obviously a significant factor for a team like ours which is conceding very few chances and as a result are conceding less than half a goal per game.

[Rileybobs]

We scored 9 goals and picked up 6 points from the first 2 games. Since then we've scored 8 goals in 10 games and are bottom 3 in the league for creating goalscoring opportunities whilst we've just had the easiest (on paper) run of fixtures we'll have all season, dropping points and failing to score from open play against 4 of these.

Since the last time we faced a side in the top half of the table our points per game has dropped from 2 to 1.85. Goals per game has dropped from 2.2 to 0.85. We are now going to start coming up against more top half sides, there's an arguement that this will favour us and I hope that's the case but I'm not optimistic.

I think everyone will agree that we need to create more chances, but in my opinion you are putting too much weight into what I would consider some fairly 'cherry-picked' statistics rather than the hard results on the board and subsequent league position.

[Mattster]

I think everyone will agree that we need to create more chances, but in my opinion you are putting too much weight into what I would consider some fairly 'cherry-picked' statistics rather than the hard results on the board and subsequent league position.

How can goals scored per game, xG, xG per shot, points per game and where our opponents are in the table be cherry picked? What's left to pick which doesn't show drop off?

[RVclaret]

While our chance creation is around 14th in the league using a median average (less sensitive to outliers such as Sunderland away in a small sample), our xG against is currently 3rd best in the league (level with 2nd pretty much), our xG difference is 5th best in the league and expected points is 4th in the league.

All the other metrics aside from purely chance creation (which is still not horrendous) are pretty solid, given context of starting a second pre-season 3 games in and the injury crisis. Combine that with actually sitting 2nd and the weight being placed by Mattster on purely chance creation (ignoring everything else, including more qualitative factors such as team togetherness, work rate, teamwork etc) seems unfair. We are pressing really well (2nd highest in the league for ball recoveries in the oppositions own third) and dominating the ball in good areas (2nd highest in the league for final third touches v opposition final third touches).

I don’t think I’ve seen anyone say we don’t need to be better in the final third in fact most posters (who get accused of being ‘defensive’ or ‘hyping Parker’) have said as much and tried to discuss ways we might improve.

[bumba]

We scored 9 goals and picked up 6 points from the first 2 games. Since then we've scored 8 goals in 10 games and are bottom 3 in the league for creating goalscoring opportunities whilst we've just had the easiest (on paper) run of fixtures we'll have all season, dropping points and failing to score from open play against 4 of these.

Since the last time we faced a side in the top half of the table our points per game has dropped from 2 to 1.85. Goals per game has dropped from 2.2 to 0.85. We are now going to start coming up against more top half sides, there's an arguement that this will favour us and I hope that's the case but I'm not optimistic.

Does XG take in to account the quality of player? Quality of opposing keeper? Weather? Quality of teams? Match circumstances?

XG doesn't take everything in to account to make a fair statistic so it's absolutely irrelevant to an arguement around football.

[Big Vinny K]

The statistics show a definite correlation between points and expected goals…..of that there is no doubt.
Of course there will always be exceptions but the point is to look at it over a longer period of time and across all clubs.

Given what the expected goals statistic is actually about this is hardly breaking news. The more and easier chances you make the more likely it is you will score - especially over a decent enough amount of time. Dyche has mentioned this many times whilst he has been Everton manager and he’s been proven right that if the good chances keep on being created then at some point the form turns (as long as of course your team does not concede a lot of great chances).

Very few statistics should be viewed in isolation and expected goals is no different. But it is a key statistic and I think the only reason a lot of people question it is because it’s called xG and the fact it is a statistic !! Take it back to basics and have a discussion with the same people who are against it and I am sure they would agree that you are a lot more likely to win games the more great chances you make in a game.

[Spijed]

http://web.archive.org/web/201811121758 ... ted-goals/

tl:dr

If things follow that pattern where do you expect us to finish?

[boyyanno]

How can goals scored per game, xG, xG per shot, points per game and where our opponents are in the table be cherry picked? What's left to pick which doesn't show drop off?

You've just made Rileys point and I suspect you didn't even realise it.

Picking and looking for stats that only show what you want them too is pretty much how not to use statistics 101.

[Vegas Claret]

I hope someone lets Parker know about this thread, it's amazing that they don't know the issues the team is having despite telling us in every single interview. Jeeeeeeeeeeeeeeeeeeeeesus, give the guy some time

[Tall Paul]

If things follow that pattern where do you expect us to finish?

Sorry I'm not sure what you mean. If what things follow which pattern?

All I was saying is that xG is a better indicator of future performance than league points and the article in the link shows where someone has done the work to show that is the case.

[Elizabeth]

I know, I know, you all hate me because I want the tedious one gone.

But bear with, bear with...

If I did get my wish and we got rid of the tedious one, who would you want to replace him?

Carlos Corberan?
Bo Henriksen?
Henrik Rydstrom?
Thierry Henry?
Frank Lampard?
Ole Gunnar Solskjaer?
Kasper Hjulmand?


Igor Tudor was my 1st choice last time, still available.

What are you?

Attention seeker?
Rovers fan?
Over sensitive?
Scared?
Just wanting a chat?

[Mattster]

You've just made Rileys point and I suspect you didn't even realise it.

Picking and looking for stats that only show what you want them too is pretty much how not to use statistics 101.

I'm asking what's left to pick that shows the opposite of what I've posted.

The clue is in the word "doesn't". The stats I've picked are ones that "do" show drop off. Read.

[Mattster]

While our chance creation is around 14th in the league using a median average (less sensitive to outliers such as Sunderland away in a small sample), our xG against is currently 3rd best in the league (level with 2nd pretty much), our xG difference is 5th best in the league and expected points is 4th in the league.

All the other metrics aside from purely chance creation (which is still not horrendous) are pretty solid, given context of starting a second pre-season 3 games in and the injury crisis. Combine that with actually sitting 2nd and the weight being placed by Mattster on purely chance creation (ignoring everything else, including more qualitative factors such as team togetherness, work rate, teamwork etc) seems unfair. We are pressing really well (2nd highest in the league for ball recoveries in the oppositions own third) and dominating the ball in good areas (2nd highest in the league for final third touches v opposition final third touches).

I don’t think I’ve seen anyone say we don’t need to be better in the final third in fact most posters (who get accused of being ‘defensive’ or ‘hyping Parker’) have said as much and tried to discuss ways we might improve.

You're choosing median for your stats because it shows a slightly better picture. We're 9th for xGD on a mean average and 21st for xG on a mean average. For me, 12 games is not a small sample size, being over a quarter of a season.

And you're doing that whilst ignoring the context of the stats you're using (xGA, final 3rd ball recoveries, final third touches vs opposition) will be artificially boosted by the fact we've faced 9 bottom half teams and only 3 top half*

*also think you're taking these stats from a site that only gives the last 5 games average too (and thus wouldn't include the Sunderland game anyway). Open to being corrected on that.

[boyyanno]

I'm asking what's left to pick that shows the opposite of what I've posted.

The clue is in the word "doesn't". The stats I've picked are ones that "do" show drop off. Read.

You've just made it again.

[kentonclaret]

We scored 9 goals and picked up 6 points from the first 2 games. Since then we've scored 8 goals in 10 games and are bottom 3 in the league for creating goalscoring opportunities whilst we've just had the easiest (on paper) run of fixtures we'll have all season, dropping points and failing to score from open play against 4 of these.

Since the last time we faced a side in the top half of the table our points per game has dropped from 2 to 1.85. Goals per game has dropped from 2.2 to 0.85. We are now going to start coming up against more top half sides, there's an arguement that this will favour us and I hope that's the case but I'm not optimistic.

Following the sale of 6/7 players that featured in the opening 2 victories surely nobody expected the goal fest to continue, I certainly didn’t. Following the Leeds game I posted that 1-0 scorelines would more likely now become the norm. We lack a striker of any real quality to convert any chances created, hence the clamour from posters on the January transfer thread to bring in 1 or even 2 forwards of quality. Not many clubs sell their top striker (s) in the January window since many clubs are involved in a fight for either a play-off spot or to avoid relegation so am far from optimistic on that front. A loan would be more likely than a permanent signing.

[Mattster]

You've just made it again.

Then please point me to the stats that don't show drop off. Because otherwise you're suggesting I post every single stat in existence relating to Burnley FC 24/25 to avoid "picking" stats which is just ridiculous.

[RVclaret]

You're choosing median for your stats because it shows a slightly better picture. We're 9th for xGD on a mean average and 21st for xG on a mean average. For me, 12 games is not a small sample size, being over a quarter of a season.

And you're doing that whilst ignoring the context of the stats you're using (xGA, final 3rd ball recoveries, final third touches vs opposition) will be artificially boosted by the fact we've faced 9 bottom half teams and only 3 top half*

*also think you're taking these stats from a site that only gives the last 5 games average too (and thus wouldn't include the Sunderland game anyway). Open to being corrected on that.

I’d just argue the Sunderland game was that much of an anomaly with everything going on that the data from that game should be relied on less, which is what the median would help with at this stage. For a more ‘true’ picture of performance that game’s effect should be reduced. In a larger sample it would obv have less weight. And yeah you are factually right about the position of teams faced albeit Sheff Weds in 13th (bottom half) are level on points with 12th (top half), while Oxford (14th) are just 1 point behind, so it’s marginal and that number (9 v 3) seems large but in reality it isn’t. Then we could consider Oxford actually were top half when we played them and have a top 6 home record as things stand (additional context). But overall, sure, the teams faced could tilt some of that - on the flip side, it’s also possible we create more against teams who open up a bit more (of course could also not happen). And no the averages are from the full season.

[Mattster]

Following the sale of 6/7 players that featured in the opening 2 victories

Including Gudmundsson, Amdouni, Zaroury, McNally and Weghorst in this figure is ridiculous given they played a combined 77 (out of possible 450) minutes in those 2 victories.

[Mattster]

I’d just argue the Sunderland game was that much of an anomaly with everything going on that the data from that game should be relied on less, which is what the median would help with at this stage. For a more ‘true’ picture of performance that game’s effect should be reduced. In a larger sample it would obv have less weight. And yeah you are factually right about the position of teams faced albeit Sheff Weds in 13th (bottom half) are level on points with 12th (top half), while Oxford (14th) are just 1 point behind, so it’s marginal and that number (9 v 3) seems large but in reality it isn’t. Then we could consider Oxford actually were top half when we played them and have a top 6 home record as things stand (additional context). But overall, sure, the teams faced could tilt some of that - on the flip side, it’s also possible we create more against teams who open up a bit more (of course could also not happen). And no the averages are from the full season.

Link please.

[Row x]

Including Gudmundsson, Amdouni, Zaroury, McNally and Weghorst in this figure is ridiculous given they played a combined 77 (out of possible 450) minutes in those 2 victories.

But 2 goals.

[boyyanno]

Then please point me to the stats that don't show drop off. Because otherwise you're suggesting I post every single stat in existence relating to Burnley FC 24/25 to avoid "picking" stats which is just ridiculous.

R.V has posted some, you've made it quite clear that you'll ignore and discredit anything that doesn't "prove" you to be the "genius" you apparently are.

Apparently expected goals against isn't a relevant metric for you. And beside the stats which I think are reasonably irrelevant after 12 games I still think what you are missing is the points on the board column.

You predicted us falling down the table after the Leeds game, we haven't, we've obviously done somthing right. Are we the finished article- no. Most people can recognise that things changed following the sale of some very important and key first team personal and a massive upheaval in the summer, that occurred after the first few games. Your stats could easily show a team that is struggling to adapt to that challenge but have still found ways to pick up points anyway.

One thing I know about football is it isn't linear. Teams have good form and bad form, statistics may account for it in an average over 46 games, but it does not do so over 12.

Nevertheless I'll give you a chance to show all of us chumps up- it's simple, use the stats to tell me what the score and result will be at the weekend? Should be simple for you

[Mattster]

Including Gudmundsson, Amdouni, Zaroury, McNally and Weghorst in this figure is ridiculous given they played a combined 77 (out of possible 450) minutes in those 2 victories.

Possible 900 minutes, actually. My mistake

[Mattster]

Apparently expected goals against isn't a relevant metric for you.

xGA averaged 0.8 after the Leeds game. 0.78 for the 7 games (all against bottom half sides) since. So a slight improvement rather than a decline. I'm not hiding from stats, I just figure it's so small a change as to be irrelevant.

[Dark Cloud]

This current team/squad simply does not have enough goals in it. Well strong enough in other areas, but sadly lacking in goal threat. Even Pep wouldn't be able to rectify that issue without the help of a transfer window. Changing manager would be totally futile.

[RVclaret]

Link please.

There’s no link that states it specifically but the ‘Total’ for each / total amount of games = ‘Average’.

Burnley Total Possession = 692.5

692.5 / 12 = 57.71 (the number presented in ‘average’, although I used median for reasons stated)

[TsarBomba]

This current team/squad simply does not have enough goals in it. Well strong enough in other areas, but sadly lacking in goal threat. Even Pep wouldn't be able to rectify that issue without the help of a transfer window. Changing manager would be totally futile.

Absolutely. The level headed recognise that.

[Mattster]

There’s no link that states it specifically but the ‘Total’ for each / total amount of games = ‘Average’.

Burnley Total Possession = 692.5

692.5 / 12 = 57.71 (the number presented in ‘average’, although I used median for reasons stated)

How do you work out a median xGD /xGA etc from a total figure and without the individual match figures?

[RVclaret]

How do you work out a median xGD /xGA etc from a total figure?

It’s already been done on the site, obviously it doesn’t use the total as that’s not how you find a median, I’m basing my assumption that ‘the (mean) Average is presented using 12 games so surely the Median is the same’ as its within the same drop down toggle on the site. I haven’t done it myself to check.

[RVclaret]

It’s already been done on the site, obviously it doesn’t use the total as that’s not how you find a median, I’m basing my assumption that ‘the (mean) Average is presented using 12 games so surely the Median is the same’ as its within the same drop down toggle on the site. I haven’t done it myself to check.

Having just checked, the median is using all 12 games.

[Mattster]

Having just checked, the median is using all 12 games.

Are you talking about this site?
https://football-match-reports.streamlit.app/

[RVclaret]

Are you talking about this site?
https://football-match-reports.streamlit.app/

Yep

[Mattster]

Yep

That has our median xGD as 0.3.

Our median xGD for the season is 0.15

All our xGD (sorted)
-0.6 , -0.6 , -0.4 , -0.3 , -0.2 , 0.1 , 0.2 , 0.2 , 0.5 , 0.8 , 1.1 , 1.4

[Spijed]

How did our XG compare under SD

[RVclaret]

That has our median xGD as 0.3.

Our median xGD for the season is 0.15

All our xGD (sorted)
-0.6 , -0.6 , -0.4 , -0.3 , -0.2 , 0.1 , 0.2 , 0.2 , 0.5 , 0.8 , 1.1 , 1.4

Hmm interesting, 0.3 wouldn’t be right for the last 5 either. The median xG is 1.41, and the ‘middle two’ of the sort are 1.2 and 1.62 (this is the one I checked).

Edit - checked your numbers and not sure where you are looking but that sort isn’t correct, 0.3 is correct.

[Mattster]

Hmm interesting, 0.3 wouldn’t be right for the last 5 either. The median xG is 1.41, and the ‘middle two’ of the sort are 1.2 and 1.62 (this is the one I checked).

Edit - checked your numbers and not sure where you are looking but that sort isn’t correct, 0.3 is correct.

FBref, which also uses Opta xG model same as that site states.

Screenshot_20241029_200258_Samsung Internet.jpg (162.91 KiB) Viewed 1576 times

[RVclaret]

FBref, which also uses Opta xG model same as that site states.

Screenshot_20241029_200258_Samsung Internet.jpg

Ah, that site uses Opta data as inputs but its own xG model (while on the topic one of my criticisms of the xG family is the wide variety in final numbers from FotMob to FbRef to this site to StatsBomb, 3 of those had a different final number on Saturday ranging from 1.2 - 1.97). The individual match data can be found in the same manner as you are doing with FbRef though.

[RVclaret]

How did our XG compare under SD

Publicly available data doesn’t go back to his Championship days to do a like for like comparison. The season we finished 10th in the PL were pretty impressive numbers though looking at it simply.

2019/2020 xG 47.4 xGA 49.8 Diff -2.4
2020/2021 39.3, 54.7, -15.4
2021/2022 39.7, 57.1, -17.4

[Mattster]

Ah, that site uses Opta data as inputs but its own xG model (while on the topic one of my criticisms of the xG family is the wide variety in final numbers from FotMob to FbRef to this site to StatsBomb, 3 of those had a different final number on Saturday ranging from 1.2 - 1.97). The individual match data can be found in the same manner as you are doing with FbRef though.

Fair enough, was thrown by it saying it was from Opta data and it looked way off what I'd been looking at.

An xG of 1.97 seems very inflated for the game against QPR based on what I saw (I personally thought 1.2 was a bit high but put it purely down to quantity of shots).

[Row x]

"Football is a simple game based on the giving and taking of passes, of controlling the ball and of making yourself available to receive a pass. It is terribly simple." ~ Bill Shankly

[Casper2]

Absolute ******** he knew!, the rug was well and truly pulled from under him. But don't let that spoil your agenda fella!

What agenda would that be ?
Anyone with an ounce of sense knew what was going to happen and Parker certainly isn’t stupid

[bumba]

FBref, which also uses Opta xG model same as that site states.

Screenshot_20241029_200258_Samsung Internet.jpg

How is Blackburn 0.8 Foster scores a free header from 10 yards out central of the goal and it doesn't even class that as an expected goal??

It's total garbage

[Mattster]

How is Blackburn 0.8 Foster scores a free header from 10 yards out central of the goal and it doesn't even class that as an expected goal??

It's total garbage

The point I'm making is literally that it doesn't end up a goal everytime (/frequently) and as such it's not a significant chance.

If you can't understand something as simple as that then I'm done engaging with you.