Sheffield United not that bad - warning: stats...

[Hague Blade]

From The Athletic

In real terms, Chris Wilder’s Sheffield United aren’t just struggling. They’ve endured the worst start of any side in Premier League history. When combined with their poor finish to last season, this has prompted suggestions that they’ve been “found out” by opponents.

The xG numbers, though, make the case that Sheffield United haven’t been as disastrous as the league table would suggest. They’ve created around 12 xG of chances, from which they’ve somehow only managed to score five times.

There’s also been an underperformance at the other end. Aaron Ramsdale has been beaten 18 times from 16.6 xGA — barely worth mentioning, were it not for the fact that, last season with Dean Henderson in goal, Sheffield United conceded only 39 goals from 51 xGA, indicating the change in goalkeeper has been significant.

 

[raul]

I want to kill xG

 

[Hague Blade]

Bit “made up” isn’t it? Love cricket stats, don’t like them in football. Possession percentage? Yeah, whatever.

 

[Carlton Blade]

It's lazy reporting.

How many times have we actually won the "xG" match convincingly? According to infogol its once, against West Brom.

Villa was a close xG defeat (0.07). Arsenal basically nothing happened, under 1 xG between the teams. Leeds (by 0.39), Fulham (by 0.01) were close "victories"

Every other game has been a rout in terms of xG, with a United deficit of at least 0.84 xG which is the equivalent of two gilt edged chances.

So, 11 games.

1 convincing win
4 coin toss games
6 convincing defeats

So a best case scenario of 15 points from 33, probably fair to say an actual expected out come of 4-9 points.

But none of that takes into account the continuing poor goal scoring figures since January 2020 or the fact we play in Toblerone shaped boots.

The numbers back up what I seeing on the pitch on a game by game basis.

 

[Webding]

I don’t care about expected goals, I just expect our midfield to fucking do something (xFDS)

 

[Le Cynique]

From The Athletic
There’s also been an underperformance at the other end. Aaron Ramsdale has been beaten 18 times from 16.6 xGA — barely worth mentioning, were it not for the fact that, last season with Dean Henderson in goal, Sheffield United conceded only 39 goals from 51 xGA, indicating the change in goalkeeper has been significant.

I'm surprised that this is from The Athletic, as it completely ignores the defence and lays the blame squarely at the GK's feet (or gloves, I suppose). This is patently utter bullshit, and shows why stats should never be used without relevant context.

 

[Working Class Hero]

It's lazy reporting.

How many times have we actually won the "xG" match convincingly? According to infogol its once, against West Brom.

Villa was a close xG defeat (0.07). Arsenal basically nothing happened, under 1 xG between the teams. Leeds (by 0.39), Fulham (by 0.01) were close "victories"

Every other game has been a rout in terms of xG, with a United deficit of at least 0.84 xG which is the equivalent of two gilt edged chances.

So, 11 games.

1 convincing win
4 coin toss games
6 convincing defeats

So a best case scenario of 15 points from 33, probably fair to say an actual expected out come of 4-9 points.

But none of that takes into account the continuing poor goal scoring figures since January 2020 or the fact we play in Toblerone shaped boots.

The numbers back up what I seeing on the pitch on a game by game basis.

Correct. What's the point in looking at XG as a whole over 11 games, when 4 out of the 12 XG came against West Brom?

 

[cooperblade]

I'm surprised that this is from The Athletic, as it completely ignores the defence and lays the blame squarely at the GK's feet (or gloves, I suppose). This is patently utter bullshit, and shows why stats should never be used without relevant context.

To be fair, by saying "barely worth mentioning" they're trying to go back on starting with "an underperformance in goal". What they are saying is Ramsdale has performed pretty much as you'd expect the average goalkeeper to perform but last season we had a goalkeeper who was anything but average and that downgrade has hurt us.

It's more that Henderson was brilliant rather than Ramsdale's been shit.

 

[Le Cynique]

To be fair, by saying "barely worth mentioning" they're trying to go back on starting with "an underperformance in goal". What they are saying is Ramsdale has performed pretty much as you'd expect the average goalkeeper to perform but last season we had a goalkeeper who was anything but average and that downgrade has hurt us.

It's more that Henderson was brilliant rather than Ramsdale's been shit.

It's impossible to say without seeing Ramsdale behind the same solid defence that Henderson had. A goalkeeper relies on the players in front of him and should literally be the last line of defence. If they're not doing their job, it's unfair to lay the blame purely at the goalkeeper's door. No keeper could possibly keep out every single shot, even Henderson.

 

[Foody]

To be fair, by saying "barely worth mentioning" they're trying to go back on starting with "an underperformance in goal". What they are saying is Ramsdale has performed pretty much as you'd expect the average goalkeeper to perform but last season we had a goalkeeper who was anything but average and that downgrade has hurt us.

It's more that Henderson was brilliant rather than Ramsdale's been shit.

It isn't even as simple as this imho. I like xg, but you have to understand it's weaknesses, there are a lot of variables that it ignores. The positional sense, ability to close down players, structure etc meant that it was rare last season that a striker had much time to compose themselves, this season that hasn't been the case. Not only are we conceding more high quality chances, but those attackers have more time to choose when they want to strike. Ramsdale is being left exposed by shoddy defending in pretty much every single game.
Even when we have a set defence to face attacks, we still look fragile. Last season we almost never conceded when we had time to set the defence.

 

[CarlMcBlade]

Gerrit int back ot net

 

[cooperblade]

It's impossible to say without seeing Ramsdale behind the same solid defence that Henderson had. A goalkeeper relies on the players in front of him and should literally be the last line of defence. If they're not doing their job, it's unfair to lay the blame purely at the goalkeeper's door. No keeper could possibly keep out every single shot, even Henderson.

But xG takes into account the quality of a chance (and therefore to some extent, the quality of the defence). The difference between expected xG for the opposition and actual goals against is relevant whether you're playing behind Maldini and Baresi or Nosworthy and Stewart.

If we'd defended better (like last season), Ramsdale might've conceded 6 goals from 5.5xGA and his performance (in this metric) would've been just the same.

Like I said, I think the Athletic article is a bit clumsily worded. They aren't blaming Ramsdale and they aren't saying he's a poor goalkeeper. They're saying he's not as good as Henderson. Would anyone dispute that?

 

[cooperblade]

It isn't even as simple as this imho. I like xg, but you have to understand it's weaknesses, there are a lot of variables that it ignores. The positional sense, ability to close down players, structure etc meant that it was rare last season that a striker had much time to compose themselves, this season that hasn't been the case. Not only are we conceding more high quality chances, but those attackers have more time to choose when they want to strike. Ramsdale is being left exposed by shoddy defending in pretty much every single game.
Even when we have a set defence to face attacks, we still look fragile. Last season we almost never conceded when we had time to set the defence.

The best xG models (used by the likes of Benham and Bloom) would take all of that into account.

 

[cooperblade]

Ramsdale should, in theory, be one of the safest signings we ever make. It's very rare that we'll know as much about a player and his ability before he comes to the club. He'll have worked with Darren Ward day in, day out. Ward and Wilder obviously rate him very highly and they do so from a position of having lots of information about him.

I just wonder if he's a bit tense at the moment, trying a bit too hard and not as relaxed as he would've been if he'd gone elsewhere? And I fully agree that he's playing behind a defence that's much less solid than last season so he hasn't had the luxury of a gentle settling in period with not much to do.

 

[Foody]

The best xG models (used by the likes of Benham and Bloom) would take all of that into account.

Not necessarily though, it only accounts for the basics, for example it will look at where the shot was taken from, rather than where it could have been. If that makes sense. It's impossible to have a completely accurate system, no one really claims it is.
We were exceptionally good at being structured in the box, now we seem pretty poor at it.

 

[Gimli]

Ramsdale should, in theory, be one of the safest signings we ever make. It's very rare that we'll know as much about a player and his ability before he comes to the club. He'll have worked with Darren Ward day in, day out. Ward and Wilder obviously rate him very highly and they do so from a position of having lots of information about him.

I just wonder if he's a bit tense at the moment, trying a bit too hard and not as relaxed as he would've been if he'd gone elsewhere? And I fully agree that he's playing behind a defence that's much less solid than last season so he hasn't had the luxury of a gentle settling in period with not much to do.

When we signed him, many people on here said ‘that’s the keeper position sorted for the next 5+ years’ - I was on of them.

Now we’re losing every week and suddenly some people think he’s a waste of money.

I still think Ramsdale is a great signing, he’s not performing terribly, maybe a bit short of confidence but behind a defence that is a shadow of last season.

 

[Balham]

xGA in its simple form doesn't tell the whole story if you're critiquing the goalkeepers. That's where PSxG (post-shot expected goals) comes in: if a shot is off target then its post-shot xG is zero.

Ramsdale, on that measure, has stopped 1.4 expected goals. We are fourth bottom in terms of PSxG, our 19.4 is just behind Burnley (18.3 despite them playing one fewer game) and better than Leeds (21.2), Fulham (21.8) and West Brom (24.8).

Last season, over 38 games our record was +3.5 but obviously Henderson didn't play against Newton Heath. Henderson's record was +5.2* which was 6th best behind Dubravka (+10.0), Lloris (+7.3 from 20 games!), Guaita (+6.8), McCarthy (+6.1 from 28 games) and Leno (+6.1 from 28 games). On that measure, for what it's worth, people's favourite Nick Pope (-1.8) was just behind Aaron Ramsdale (-0.3).

*meaning that Simon Moore's was -1.7 from two games...over a full season that's more than 30 additional goals conceded :-|

2020/21 data: https://fbref.com/en/comps/9/keepersadv/Premier-League-Stats
2019/20 data: https://fbref.com/en/comps/9/3232/keepersadv/2019-2020-Premier-League-Stats

 

[tincan]

The manager has known that we cannot convert the XG chances we create all last season -so he spends £24m on a never has been to solve it. He shows no signs of admitting he has got things wrong, until that happens and he stops blaming the players, he cannot change things and we will keep losing.

 

[broomhillblade]

I think we're missing Jack O'Connell big time - at both ends of the pitch.

 

[Hague Blade]

I honestly think part of the problem with Ramsdale, and I'm guilty of this myself, is that it's hard to shake the feeling of resentment at how much we paid for him. He was our player, we gave him away for next to nothing, and then we had to buy him back for somewhere between £12m and £18m. Hard not to feel a bit of a sense of expectation (even though it's relative peanuts, it feels like a lot for little ole Sheff Utd) and in the face of that expectation he's been no more than OK.

He obviously suffers in comparison to Henderson, but I think he also suffers as being perceived as not value for money. We're a tight bunch, and we expect a lot for 18 million.

 

[Yogi]

I honestly think part of the problem with Ramsdale, and I'm guilty of this myself, is that it's hard to shake the feeling of resentment at how much we paid for him. He was our player, we gave him away for next to nothing, and then we had to buy him back for somewhere between £12m and £18m. Hard not to feel a bit of a sense of expectation (even though it's relative peanuts, it feels like a lot for little ole Sheff Utd) and in the face of that expectation he's been no more than OK.

He obviously suffers in comparison to Henderson, but I think he also suffers as being perceived as not value for money. We're a tight bunch, and we expect a lot for 18 million.

I’d agree it’s the expectation which alters the perception of him those stats indicate that he has been as advertised a roughly average prem keeper. I know it’s popular at the minute to hold the price tag up and say that that money has been wasted but realistically we haven’t overpaid for an average prem keeper. Could we have gotten a 30+ year old at the same level for circa 5 mill probably but Ramsdale gives us a number 1 for at least 10 years so over that time period we’d have had to buy at least 3 if not 4 30+ year olds at a similar or slightly higher cost. If we choose to move on from Ramsdale in his prime at around 26-30 then we’d at least wipe our face if not make a profit it isn’t a bad investment.

 

[cooperblade]

Not necessarily though, it only accounts for the basics, for example it will look at where the shot was taken from, rather than where it could have been. If that makes sense. It's impossible to have a completely accurate system, no one really claims it is.
We were exceptionally good at being structured in the box, now we seem pretty poor at it.

It's possible to have a system that's (at least in part) responsible for some people being rich enough to buy football clubs.

The best models will look at pretty much every variable - was it on the player's strong foot, was it bouncing, was the player under pressure, where was the goalkeeper positioned etc etc. They have a team of people who study every chance in detail.

On a separate note, https://twitter.com/Jhdharrison1 is brilliant in terms of analysing goalkeeper performance

 

[Wincobank S9 Blade]

IMO it’s the expectation sometimes, that generation of blades that’s seen us in 4th division in the early 80’s (which quite frankly would have been horrendous) will see this not as bad because they’ve admittedly been through the blades rollercoaster of agony a lot longer I think most people would love us to be a solid PL team but with our wage structure 1 bad season could be costly I never imagined it to be this bad mind you but on grand scheme of things atleast we haven’t done a Portsmouth and I’m convinced Wilder will turn this around

 

[Revolution]

I never imagined it to be this bad mind you but on grand scheme of things atleast we haven’t done a Portsmouth

yet.

 

[Balham]

I honestly think part of the problem with Ramsdale, and I'm guilty of this myself, is that it's hard to shake the feeling of resentment at how much we paid for him. He was our player, we gave him away for next to nothing, and then we had to buy him back for somewhere between £12m and £18m. Hard not to feel a bit of a sense of expectation (even though it's relative peanuts, it feels like a lot for little ole Sheff Utd) and in the face of that expectation he's been no more than OK.

He obviously suffers in comparison to Henderson, but I think he also suffers as being perceived as not value for money. We're a tight bunch, and we expect a lot for 18 million.

This is it. I don't understand why we allow the purchase price narrative to be driven by the selling club. It's the same with Robinson (reported as £8m, more like £5m), Mousset (£10m!/£6m), McBurnie (£20m/£17m), Brewster (£23.5m/£18m). Our up-front cost of signing Ramsdale is somewhere between £10m and £12m yet we've saddled him with the expectations of a fee nearly double that.

 

[Revolution]

This is it. I don't understand why we allow the purchase price narrative to be driven by the selling club. It's the same with Robinson (reported as £8m, more like £5m), Mousset (£10m!/£6m), McBurnie (£20m/£17m), Brewster (£23.5m/£18m). Our up-front cost of signing Ramsdale is somewhere between £10m and £12m yet we've saddled him with the expectations of a fee nearly double that.

I have always thought that fees should be transparent, public information.