I suspect the 'treatment table' is more bed-like than table-likeIs a treatment table a bed? (Cue canario complaining that this is a football forum, not a semantics forum .........)
The Stats Thread
- Thread starter DUNCAN DONUTS
- Start date
-
Please bear with us on the new site integration and fixing any known bugs over the coming days. If you can not log in please try resetting your password and check your spam box. If you have tried these steps and are still struggling email [email protected] with your username/registered email address
-
Can anybody dig up the most amount of clean sheets we have had in a season? I'm wondering if we stand a chance to better it this year.
According to David Freezer in the Pink it is twenty in '74/'75, and 36 is the fewest number of goals conceded, in '71/'72. Records could be set in a number of areas this season it seems.
Last edited:
Judging by some of the modern novels my wife reads, there are some things that happen on tables that I'll wager never happen on Colney treatment tables .............I suspect the 'treatment table' is more bed-like than table-like
I think we are on fifteen so far this season? To get to twenty-one clean sheets from here feels like a tall order.
Keeping goals against under 36 feels doable and would be fantastic.
I think those records Freezer came up with concern our performance in the second tier only, so not sure if they are outright records. 94 points is also another record that could be challenged.
Here is the latest Championship scatter graphic from Experimental 3:6:1:
and here, for comparison, is the equivalent from March 2019:
The lines dividing the four sectors in each graphic are the median values for the x and y axes. Ben Mayhew's summary in both cases, very good in attack, close to average in defence.
You must log in or register to see images
and here, for comparison, is the equivalent from March 2019:
You must log in or register to see images
The lines dividing the four sectors in each graphic are the median values for the x and y axes. Ben Mayhew's summary in both cases, very good in attack, close to average in defence.
I don’t know how they’re calculating the median, because while it’s difficult to tell it only looks like ten teams below the line.
It’s falsified, though, because the defences are significantly stronger this season versus 18/19. We were 1.3 then, 1.1 now.
I don’t know how they measure “1.1” either. We’ve conceded (joint) fewest. So it’s all very dubious to me.
Yeah I was trying to work that out unless Krul's/ McGovern's saves mean that the defense looks better because there are less goals scored due to good keeping?
Isn't it a stat thing that you can gave fewer teams on one side of a line? It just means that a few teams are a long way above average and are balanced by more being just below. Rather like GD where a couple of teams have big positives and larger numbers smaller negatives.
Spot the odd one out (hint: it's not Swansea 
):
Rankings are from football xg. (xG ranking high to low; xGa ranking low to high; Net xG ranking high to low)
):Rankings are from football xg. (xG ranking high to low; xGa ranking low to high; Net xG ranking high to low)
You must log in or register to see images
Luton?Spot the odd one out (hint: it's not Swansea
Rankings are from football xg. (xG ranking high to low; xGa ranking low to high; Net xG ranking high to low)
You must log in or register to see images
Spot the odd one out (hint: it's not Swansea
Rankings are from football xg. (xG ranking high to low; xGa ranking low to high; Net xG ranking high to low)
You must log in or register to see images
The trouble is Robbie this is still unconvincing. When our xGa is close to 50% higher than actually conceded, it suggests maybe the xGa stat is not quite working. Swansea and Watford are also miles off their xGa.
And you have told us before that you do not believe this is poor finishing - right? Because otherwise you would not argue when we point to our xG being six goals higher than our actual goals scored that it is nothing to do with poor finishing...
It’s all a bit inconsistent. Which really just highlights why xG and xGa are massively flawed and, while interesting, don’t really tell us something concrete.
Especially as they’re also vulnerable to manipulation - for example, recently you have often changed how you use the xGa stat by arguing that one big chance xG is worth more than lots of little xG chances (even though that would defeat the object of having xG percentages as if they are correct than 75% made up of multiple small chances should have exactly the same odds of a goal as one shot with a 75% chance...)
But I can also manipulate the xGa stats for Norwich across matches even if we swallow the idea of xGa being accurate - for example, over the last six matches our average xGa per match is 0.867. That would put us top of your table by a massive margin on a seasonal basis.
But that’s only recent form you say? Well Ok, let’s take our last 15 matches. Well our xGa for that period 0.872 - so only a fraction higher. So our excellent defensive form has actually been for an extended period of time!
But that’s less than half the season I hear you say - Rob you’re picking the best periods to show us in a better light. Well ok, let’s go back 20 matches then. Ah! Well now we are talking - our xGa has shot up to 0.97 per match! But hang on a second... That would still see us second in the defensive table... So still pretty good.
Even if you back 25 matches we are still almost exactly 1 goal xGa a game. If you want to get clever and look at only our possible opponents next year, in that same period of xGa is 1.06 against top six teams.
The point being we all knew our form at the start of this season was shaky and our defence wasn’t great, but since January in particular it has really clicked. Stats shouldn’t be looked at in a vacuum, they should always be contextualised, their flaws acknowledged. Lies, damn lies and football stats.
How strong is our defence?
If you look at the top five leagues in Europe, all pro tiers, across the last 20 matches played only Padova in the Italian third tier has conceded fewer goals than the ten we have conceded (they have conceded nine). Lille and Man City tie for third place having conceded eleven in their last 20 matches.
If you look at the last 15 matches, we are top (seven conceded), jointly with Padova and Lille.
That roughly coincides with Giannoulis joining.
If you look at the top five leagues in Europe, all pro tiers, across the last 20 matches played only Padova in the Italian third tier has conceded fewer goals than the ten we have conceded (they have conceded nine). Lille and Man City tie for third place having conceded eleven in their last 20 matches.
If you look at the last 15 matches, we are top (seven conceded), jointly with Padova and Lille.
That roughly coincides with Giannoulis joining.