Statistical Domination 2023/24

[repman 642]

Got the xG graphs from the last 2 games here

The Blackburn game was almost the perfect example of why game state matters, and how the graphs can at least demonstrate that a bit better than looking at the final tally. I think anyone watching the game knew there was a stark difference before and after the red but this makes it even clearer.

Onto Cardiff and there's a few things to note here, the first one is that when making these I only get the data from every individual shot, this means that the final xG score might sometimes be inflated. Idah's winner is the perfect example of this situation arising where there are 2 shots in the same attack. Rowe's shot was worth 0.47xG and Idah's follow up was 0.74, overall models will probably credit Norwich with a lower xG than I've shown because out of that attack there could only ever be one goal, whereas the graph adds both which gives it a value of 1.21xG. 

The other thing I do find amusing is how a result can impact people's perception of chances, Hwang's miss in the first half was given an xG of 0.8, when Idah missed that chance against Swansea it was 0.48 but you only really heard about one over the other.

Cardiff game was a very weird one overall I felt, we created some really good chances but certainly in the first half we were poor and maybe Cardiff sat back a bit much in the 2nd half? (obviously the full backs coming on had a big impact too).

I held off posting these in case Wagner ended up going so I could also make the 5 game average chart to show a broader picture, that will wait a little while longer I guess.

 

[hogesar 9,817]

On 05/11/2023 at 22:41, ZLF said:

These are the whoscored equivalents equally placing KM relatively high.

But hogesar, yiur defencenof both kenny and Gibson is that they are better than sara and duffy;  are these stats reliable or not (I don't believe they are, like all data items they are pretty useless on their own and huge elements of context need applying)

What these data views don't show is just how inadequate the squad are when out of possession or the impact of a team selection that gifts space to opponents.

 

Sorry ZLF, only just seen this!

No, I don't think Kenny is better than Sara. But I do think they don't work as a two and it's not really Kenny's fault that they don't work as a two - I think Sara's heatmaps that I showed previously demonstrates just how high he ends up playing - and when we get caught with him that high we only have Mclean in sometimes 1 v 3 overload against him - he will then get criticised (rightly so) for ball-chasing but we have little option at that point.

I do think Gibson is better than Duffy though. In pretty much every aspect other than perhaps aerial duels. 

[hogesar 9,817]

10 minutes ago, repman said:

Got the xG graphs from the last 2 games here

The Blackburn game was almost the perfect example of why game state matters, and how the graphs can at least demonstrate that a bit better than looking at the final tally. I think anyone watching the game knew there was a stark difference before and after the red but this makes it even clearer.

Onto Cardiff and there's a few things to note here, the first one is that when making these I only get the data from every individual shot, this means that the final xG score might sometimes be inflated. Idah's winner is the perfect example of this situation arising where there are 2 shots in the same attack. Rowe's shot was worth 0.47xG and Idah's follow up was 0.74, overall models will probably credit Norwich with a lower xG than I've shown because out of that attack there could only ever be one goal, whereas the graph adds both which gives it a value of 1.21xG. 

The other thing I do find amusing is how a result can impact people's perception of chances, Hwang's miss in the first half was given an xG of 0.8, when Idah missed that chance against Swansea it was 0.48 but you only really heard about one over the other.

Cardiff game was a very weird one overall I felt, we created some really good chances but certainly in the first half we were poor and maybe Cardiff sat back a bit much in the 2nd half? (obviously the full backs coming on had a big impact too).

I held off posting these in case Wagner ended up going so I could also make the 5 game average chart to show a broader picture, that will wait a little while longer I guess.

 

Awesome @repman and thanks again.

Understand your shot-recording xG. It's not too dissimilar to Sofascore:

I actually disagree a bit re your Cardiff assertion though. I didn't think we were anywhere near as poor as made out in the first half. I think the goals quite understandably led to a lot of negativity by the end of the half but we were against a very in-form team away from home and stats looked like this:

 

They'd scored 2 goals from decent but not exceptional chances. I do think considering Warner and Fisher aren't ready yet and Placheta isn't a LB it wasn't as bad as described at the time (Wagners decision to play them is a different level of bad of course, though).

Note possession didn't change much between 1st and 2nd half, only a % either way so whilst I think Cardiff came out quite negative they weren't just 11 men behind soaking up possession.

One thing to note that's not really conclusive but our sweet-spot for results is approx 55% possession. IF we have over 60% we're normally struggling to break teams down and there's little counter opportunity (losing to Rotherham, fortunate win vs Stoke at home, Plymouth debacle). IF we have under 50% possession our xGA is higher than our xG.

But 55% with a couple % either way has been all our best results and performances - Birmingham, Huddersfield, QPR cup, Millwall, Hull.

 

 

[Robert N. LiM 4,664]

There's a separate thread on this, but thought I'd link Opta's predictions here, since the article includes the detail that we are almost exactly matching our expected points so far this season, which I thought was interesting. And that Ipswich are massively outperforming theirs, which is giving me hope that Daniel might reel them in. Also suggests that Coventry and Preston are both in slightly false positions at this stage.

@hogesar @repman

https://theanalyst.com/eu/2023/11/championship-predictions-2023-24-november/

Edited November 21, 2023 by Robert N. LiM

[Nuff Said 5,246]

2 hours ago, Robert N. LiM said:

There's a separate thread on this, but thought I'd link Opta's predictions here, since the article includes the detail that we are almost exactly matching our expected points so far this season, which I thought was interesting. And that Ipswich are massively outperforming theirs, which is giving me hope that Daniel might reel them in. Also suggests that Coventry and Preston are both in slightly false positions at this stage.

@hogesar @repman

https://theanalyst.com/eu/2023/11/championship-predictions-2023-24-november/

Interesting that Leicester are over-performing too.

[Robert N. LiM 4,664]

32 minutes ago, Nuff Said said:

Interesting that Leicester are over-performing too.

Indeed. Based on nothing but my desires, I chose to interpret this as Leicester having quality players in key areas (so likely to continue outperforming xG), and Ipswich just being lucky...

[Midlands Yellow 4,118]

It’s a fantastic thread, mostly enjoyed by 3/4 posters. 

[Robert N. LiM 4,664]

15 minutes ago, Midlands Yellow said:

It’s a fantastic thread, mostly enjoyed by 3/4 posters. 

Thanks for continuing to bump it.

[Midlands Yellow 4,118]

Just now, Robert N. LiM said:

Thanks for continuing to bump it.

You’re welcome, it’s a fascinating read by very interesting posters. More please. 

[Nuff Said 5,246]

11 hours ago, Midlands Yellow said:

It’s a fantastic thread, mostly enjoyed by 3/4 posters. 

And that’s a problem because…?

[hogesar 9,817]

34 minutes ago, Nuff Said said:

And that’s a problem because…?

He doesn't understand any of it.

[hogesar 9,817]

Thanks @Robert N. LiM

I suppose it backs up the eye test that we've generally deserved to win the games we have won (maybe Stoke at home apart?) and generally deserved to lose the one's we've lost (Rotherham away, whilst not a great performance, was an outlier in Rotherham scoring 2 low xG chances).

 

[Robert N. LiM 4,664]

3 hours ago, Nuff Said said:

And that’s a problem because…?

To be honest, I kind of miss the xG-sceptics clogging up this thread by demonstrating their innumeracy and/or insecurity.

[hogesar 9,817]

Ok so, seems a bit pointless now but QPR:

Let's dispel the myth that QPR created lots. They didn't. They had a lot of the ball, we were rubbish but relatively compact - very little came through the middle which has been our weak point. We forced them out wide and generally dealt with crosses. That's as good as it gets and of course the argument is any manager can make you (more) solid if you surrender all attacking intent.

 

 

Now the pain of last night.

We took a 2-0 lead. Not "undeservedly" so, but with fortune and luck involved. We had little of the ball but Watford gave it away in dangerous areas and we were pretty incisive. The corner goal was good to see, Watford are one of the worst at defending set pieces this season.

Hwang's goal was a great strike and Sara's quick thought of playing it first time into his path gave him the opportunity. So, 2 assists for Sara (I mention that, because that's all he generally offered).

We were 2-0 up from just 0.36 xG - we hadn't created loads but we'd taken one corner chance and Hwang's strike is just "one of those" (equally similar to Rotherham away and their first goal - they happen).

At 2-0 up but having not had control of the match (although we'd pressed wide areas pretty well Sara was still wandering and the gaps between Mclean and Sara were far too large) the Hwang injury was a shame but the opportunity to try and change the flow of the game. Watford were poor on the ball but still better than us and having plenty of it - why we didn't opt for a midfielder who can collect the ball in spaces and play simple passes to retain possession is beyond me - Gibbs / Nunez seemed ideal. Instead we opted for un-fit Barnes.

Despite being 2-0 up and Sara having two assists, both of them were having poor games. Mclean had a dreadful one. And when that happens, we invariably have no out - Mclean is tasked with both somehow covering the back 4 and the gaps when our Fullbacks push all whilst on the ball expecting to drive past or beat the first press to open a passing opportunity. This is why my view that just replacing him with a natural CDM doesn't solve our problems. It needs a change from the manager where we either discipline Sara into playing closer to Mclean and not drifting or we play 3 midfielders centrally either with a CDM and double-pivot in front or an alternative.

The most concerning thing is our two central midfielders lost possession 22 times between them (12 Mclean, 10 Sara). Comparatively, Livermore lost possession 5 times (the only really comparable midfielder as Kaymebe and Kone offered an attacking intent / threat that our midfielders don't). 

My personal view is I don't see how this changes without a change of manager. Injuries are a problem but we are making it far more of one than we need to.

 

[hogesar 9,817]

I'm hoping @repman can bring himself to provide graphs but in the meantime:

You can see how we, whilst not having loads of the ball, were controlling the game for the first half of the first half (ha).

Unfortunately, as soon as they put pressure on us it was all far too easy. And most concerning is how passive we came out in the second half - the manager isn't having an impact at half time, or if he is, it's a negative one.

 

Match ratings, if anyone is bothered anymore!

[repman 642]

Here's the graph for the QPR game, I haven't done Watford yet as the data often gets updated a day or two after the game so it's best to hold off slightly.

I think the graph perhaps makes the game look more exciting than it was, that's mainly down to the y axis values being so low.

Still, it reflects the game pretty well in my view, QPR have clearly improved since sacking Ainsworth but it seems they've focused on the defence first and will look to try to improve the attack later. You can't fault that plan really as if you stay in games there's always a chance you'll nick a point. It's the same thing that Wagner has been preaching recently, but I'm yet to see any evidence we're actually getting better defensively. QPR were quite proactive in their defending, pressing us high and keeping us away from their goal (that of course did lead to our goal), whereas our plan is simply to drop deep and hope we can defend the box. So far we're a long way from that.

[Canaries north 151]

1 hour ago, hogesar said:

I'm hoping @repman can bring himself to provide graphs but in the meantime:

You can see how we, whilst not having loads of the ball, were controlling the game for the first half of the first half (ha).

Unfortunately, as soon as they put pressure on us it was all far too easy. And most concerning is how passive we came out in the second half - the manager isn't having an impact at half time, or if he is, it's a negative one.

 

Match ratings, if anyone is bothered anymore!

I'm a bit of a luddite so dont know how these figures are put together, but despite the 2 assists from Sara how did he get a better score than Livermore. He held the midfield and broke up play so well. Neat and tidy with the ball as well. The eye test for me does not agree with the scores. 

[hogesar 9,817]

Thanks @repman although QPR was a more noticeable shift than just dropping deep (which you're right, we did).

We successfully funnelled QPR down either flank and think this is largely responsible for their low xG. We won the initial balls in the area which were often aerials. We were poor to the second ball which is a common issue for us but it meant most their efforts were speculative.

Now, how much of that is due to QPRs lack of ability / confidence is an important part of that. 

Equally, they attained a 2.2 xG vs Stoke and 1.6xG was before the sending off. So maybe not as bad as we thought.

[hogesar 9,817]

1 minute ago, Canaries north said:

I'm a bit of a luddite so dont know how these figures are put together, but despite the 2 assists from Sara how did he get a better score than Livermore. He held the midfield and broke up play so well. Neat and tidy with the ball as well. The eye test for me does not agree with the scores. 

Youre right. Basically, assists and goals have a high impact on ratings. If you took them away he'd likely have a lower rating than Livermore.

It's just because assists and goals are easily ranked. Livermore being tidier and positionally better might have helped reduce our chances but it's a little intangible. 

I agree though, Livermore was better overall for me too.

[repman 642]

https://twitter.com/louorns/status/1729851451659235808?s=20

Interesting graph here (that I hope can be seen) showing midfielders in the championship. Obviously with us it's interesting that both Sara and McLean are in the top left quadrant, which backs up the eye test - good ball players but lacking defensively. 

Obviously I think both those players are being hung out to dry by our system, but it's also a clash of profiles, I've little doubt that Sara could easily take over McLean's progressive production if we added a lesser footballer beside him to win the tackles. 

[hogesar 9,817]

For all Sara's good, I'm pleased Hucks is seeing the same as I am (although they were both terrible yesterday)...

[repman 642]

1 hour ago, hogesar said:

For all Sara's good, I'm pleased Hucks is seeing the same as I am (although they were both terrible yesterday)...

The thing is Sara has won more tackles than Kenny has this year. I'd happily agree that Sara is leaving Kenny with too much if Kenny was actually doing a lot himself, but he just isn't.

 

Kenny is further ahead in interceptions but it's hardly an elite defensive profile. We rely on Sara to do too much already, it makes more sense to find someone to compliment his weaknesses than detract from his strengths to support players who aren't as good as him.

[king canary 7,705]

1 hour ago, hogesar said:

For all Sara's good, I'm pleased Hucks is seeing the same as I am (although they were both terrible yesterday)...

To me this is a bizarrely backwards way of looking at Sara. 

Everything we see from him stats wise shows a highly creative player who is central to everything we do going forward. He's clearly a cut above anyone else in the team in these aspects of the game. To use a parma term, he's a weapon at this level. 

So it falls on the coach and the sporting director to put our best player in a position to succeed. So that means giving him less to do off the ball, less defensive work, more cover in behind etc etc. We have to prioritise that. Even if it means moving others around or even shifting formation. To be getting annoyed at Sara for not doing enough tracking back and defending is to entirely miss the point of having that sort of player.

Kenny can be shoehorned into defensive midfield because, to be blunt, he isn't special in his natural position, he's just fine. Sara shouldn't be being shoved into the role he's in because he absolutely can be special when given the correct platform. 

I really hope that the second the January window opens up we've got something lined up to make this happen, otherwise its just a waste of a very good player.

[hogesar 9,817]

1 hour ago, king canary said:

To me this is a bizarrely backwards way of looking at Sara. 

Everything we see from him stats wise shows a highly creative player who is central to everything we do going forward. He's clearly a cut above anyone else in the team in these aspects of the game. To use a parma term, he's a weapon at this level. 

So it falls on the coach and the sporting director to put our best player in a position to succeed. So that means giving him less to do off the ball, less defensive work, more cover in behind etc etc. We have to prioritise that. Even if it means moving others around or even shifting formation. To be getting annoyed at Sara for not doing enough tracking back and defending is to entirely miss the point of having that sort of player.

Kenny can be shoehorned into defensive midfield because, to be blunt, he isn't special in his natural position, he's just fine. Sara shouldn't be being shoved into the role he's in because he absolutely can be special when given the correct platform. 

I really hope that the second the January window opens up we've got something lined up to make this happen, otherwise its just a waste of a very good player.

I get you and agree with you but fundamentally the point is that simply recruiting a CDM doesn't solve the tactical problems we are having. The CDM will still be overloaded and facing 2v1 and 3v1 situations. That's because Sara isn't a traditional CM or CDM. Certainly not in a 2.

[king canary 7,705]

10 hours ago, repman said:

Kenny is further ahead in interceptions but it's hardly an elite defensive profile. We rely on Sara to do too much already, it makes more sense to find someone to compliment his weaknesses than detract from his strengths to support players who aren't as good as him.

This is a perfect summary.

Look at Sara's attacking outputs this season. Most key passes pg, most assists, 5 goals, best crosser, best at longer balls etc etc. 

Any sensible manager is looking at this team and building almost everything around getting the most out of him. Is he great defensively? No. So you build your structure around that. You give him a base behind him even if it means others are out of position. 

I don't love these ratings but whoscored have him as the 5th best player in the Championship this year. His key pass stats are second in the league. It is criminal we're asking him to play as a deep midfielder, it is even worse that we're doing this without even a proper DM next to him. He's being wasted because our coach and previous sporting director are woeful at assembling a midfield. Sara is being put in the same position as Gilmour was two years ago.

[Mason 47 1,420]

A lot of our midfield issues come about because we are married to this 'double false 9' system.

Was great when we had the specific players for it early on, but since then we've seen Idah, Hwang, Gibbs, Nunez & Sara all try to play the second striker role and do absolutely nothing with it. It's a very niche system

I'd be playing Sara as the furthest forward of a 3 in midfield for the time being. Available to get back and help out, which he is capable of, but not have him playing a primary defensive role.

[hogesar 9,817]

21 minutes ago, Mason 47 said:

A lot of our midfield issues come about because we are married to this 'double false 9' system.

Was great when we had the specific players for it early on, but since then we've seen Idah, Hwang, Gibbs, Nunez & Sara all try to play the second striker role and do absolutely nothing with it. It's a very niche system

I'd be playing Sara as the furthest forward of a 3 in midfield for the time being. Available to get back and help out, which he is capable of, but not have him playing a primary defensive role.

Yep, even Gibbs alongside Mclean with Sara further up would be helpful.

Basically, a 4-2-3-1 but without shoe-horning a striker into the pseudo-10 role.

Edited November 30, 2023 by hogesar

[BigFish 1,989]

11 hours ago, king canary said:

To me this is a bizarrely backwards way of looking at Sara. 

Everything we see from him stats wise shows a highly creative player who is central to everything we do going forward. He's clearly a cut above anyone else in the team in these aspects of the game. To use a parma term, he's a weapon at this level. 

So it falls on the coach and the sporting director to put our best player in a position to succeed. So that means giving him less to do off the ball, less defensive work, more cover in behind etc etc. We have to prioritise that. Even if it means moving others around or even shifting formation. To be getting annoyed at Sara for not doing enough tracking back and defending is to entirely miss the point of having that sort of player.

Kenny can be shoehorned into defensive midfield because, to be blunt, he isn't special in his natural position, he's just fine. Sara shouldn't be being shoved into the role he's in because he absolutely can be special when given the correct platform. 

I really hope that the second the January window opens up we've got something lined up to make this happen, otherwise its just a waste of a very good player.

That is certainly a tactical choice, and one that Wagner is not making. Point remains that the weakness in Sara's game is why he plays here and what will limit his career. It certainly isn't working as we currently set up so sticking an extra body in midfield and freeing him from the defensive side of the game is worth a try.

[BigFish 1,989]

7 minutes ago, hogesar said:

Yep, even Gibbs alongside Mclean with Sara further up would be helpful.

Basically, a 4-2-3-1 but without shoe-horning a striker into the pseudo-10 role.

Thing is in possession he is most effective exactly where he is playing, you don't want to move. Out of possession and in transition he is a liability. I think an extra body in midfield (433) or at the back (343) works better. Means either leaving out a striker/false 9 or one of the wide players.

[hogesar 9,817]

48 minutes ago, BigFish said:

Thing is in possession he is most effective exactly where he is playing, you don't want to move. Out of possession and in transition he is a liability. I think an extra body in midfield (433) or at the back (343) works better. Means either leaving out a striker/false 9 or one of the wide players.

Yep and you have to accept he's going to get caught up the field sometimes. 4-3-3 or 4-2-3-1 isn't that relevant in terms of fluidity as out of possession you can always drop pseudo-10 Sara deeper and even on the ball if Sara is deep Gibbs can move into the space.

4-3-3 makes sense and after all, if we are to believe some of the stuff from our previous season(s), it's where Webber was trying to build the team towards.