Metrics of change

Pogue Mahone said:
That counter press one is really interesting. Bumped the thread to share stats from here, which include that one.

What is counter-press, exactly? How does it differ from common or garden pressing? And how does it compare with life before ETH? This could be the most exciting metric of change in the thread so far!
Click to expand...
I can't quickly find counter-pressing stats for last season (whatever the concept means exactly), but I did find this article on a couple of aspects related to pressing:

https://www.premierleague.com/news/2636954

It doesn't give full tables for the EPL, but I suppose it's telling that United isn't mentioned aware, while that graph above suggests that they would be for this season.
 
Pogue Mahone said:
That counter press one is really interesting. Bumped the thread to share stats from here, which include that one.

What is counter-press, exactly? How does it differ from common or garden pressing? And how does it compare with life before ETH? This could be the most exciting metric of change in the thread so far!
Click to expand...
From the guy who made the chart: When a team regains possession in opponent’s half within 8 seconds after losing a ball in the same area of the pitch I count it as a "successful counter-pressing action". X-axis is counter-pressing itself (ie defending proactively), y-axis is how often successful counter-press lead to danger/chances.
 
  • Like
Reactions: Cheimoon
Surely 'pressing' and 'counter pressing' is something all teams should (collectively) be doing, all through the game, in all areas of the pitch?

"The opposition cannot score whilst you have the ball" (according to LvG)

"If he's not interfering with play... what is he doing on the pitch" (quote attributed to Danny Blanchflower)
:lol:
 
Pogue Mahone said:
That counter press one is really interesting. Bumped the thread to share stats from here, which include that one.

What is counter-press, exactly? How does it differ from common or garden pressing? And how does it compare with life before ETH? This could be the most exciting metric of change in the thread so far!
Click to expand...

Counter-pressing happens right after you've lost the ball. So if we're attacking in whatever attacking shape we have, we lose the ball, we press to try and immediately win the ball back rather than have to drop off into an organised defensive shape. In the case of those stats yer man counts it as a successful pressure if we win the ball back in the opposition half within 8 seconds of losing it in the opposition half.

Whereas regular pressing doesn't have to come right after you lose possession. For example if the opposition is in possesion for a while, we try to force them to play the ball to a certain area of the pitch, then when they do we trigger a press to try and win it back.
 
Pogue Mahone said:
That counter press one is really interesting. Bumped the thread to share stats from here, which include that one.

What is counter-press, exactly? How does it differ from common or garden pressing? And how does it compare with life before ETH? This could be the most exciting metric of change in the thread so far!
Click to expand...

It's the same thing as gengenpressing (high pressing immediately after possession loss when the game is in transition without allowing it to set).

Data is interesting and I feel like it's a little cherry picked. I don't think our pressing game is sophisticated enough to intentionally lay pressing traps and press only in certain scenarios which is what the data suggests.

If you look at the raw PPDA (passes allowed per defensive action) numbers we're pretty mediocre. However we do rank pretty well in high turnovers and high turnovers leading to shots. The pressing has definitely improved but I'd be a little skeptical about the data until the eye test says we're actually a really good pressing team.

My takeaway from that is to look for moments where we engage in the press during games and see if we changed something. We're already quite good at pressing from fixed situations (e.g., from opponent GKs when they attempt to play out the back) but that's the extent of it I feel.
 
Teja said:
It's the same thing as gengenpressing (high pressing immediately after possession loss when the game is in transition without allowing it to set).

Data is interesting and I feel like it's a little cherry picked. I don't think our pressing game is sophisticated enough to intentionally lay pressing traps and press only in certain scenarios which is what the data suggests.

If you look at the raw PPDA (passes allowed per defensive action) numbers we're pretty mediocre. However we do rank pretty well in high turnovers and high turnovers leading to shots. The pressing has definitely improved but I'd be a little skeptical about the data until the eye test says we're actually a really good pressing team.

My takeaway from that is to look for moments where we engage in the press during games and see if we changed something. We're already quite good at pressing from fixed situations (e.g., from opponent GKs when they attempt to play out the back) but that's the extent of it I feel.
Click to expand...

If you are talking about the above data, that's not what he is saying. It only talks abt pressing immediately (with in 8s) after losing the ball. In that his data shows we are both successful and productive with the subsequent attack.

Our regular pressing is not good enough, the type where the opposition is building from the back and we are trying to funnel into a trap. Agree that, that doesn't look anywhere near good. (personally think that's due to lack of a proper pressing striker and Eriksen not good enough)
 


For those wondering how those counter-pressing stats compare to last season.
 
sullydnl said:


For those wondering how those counter-pressing stats compare to last season.
Click to expand...

That would suggest we're tactically moving in a different direction to all other clubs. ETH is a dinosaur.
 
Last edited by a moderator:
The Irish Connection said:
Cool, so our counter press success rate is now best in the league?
Click to expand...

Well, over the sample size of games that is this season so far. I'm sure we'll wax and wane over time just like those other teams have, with some other team overtaking us at some point.

The key thing though is that's we're in or around where the best teams in the league have been this season and last, so there's been an unquestionable improvement in that one facet of the game at least. And I think that's backed by the eye test as well, as we've looked better in that regard.
 
sullydnl said:


For those wondering how those counter-pressing stats compare to last season.
Click to expand...


Spurs the easiest team to press in the league is interesting, guess its the workmanlike midfield.
Also interesting to see City keep the ball better with Haaland even if the actual press suffers. Nice graphic, huge improvement in United.
 
People can often make the mistake of thinking that counter-press is looking for a break (counter-attack) as soon as you win the ball, but indeed it means to win the ball as soon as you lose it - we could see that from Casemiro last night chasing a ball after his mistake forcing Charlton to go long and we won possession back.

One of the goals of counter-pressing can be to counter the opponent quickly while they are not organized at the back, especially if you win the ball close to their box (f.e. Liverpool and Klopp's 6 seconds rule) but some teams are instructed to hold shape instead and look to attack structured when they retain possession (ETH and Pep). That does not mean that they would not go for it if they can, as we could see in the early days of ETH when the team was not yet fluid in possession (Arsenal game) or last night against Charlton where there was a lot of space at the end of the match. Not rushing after we won back possession is also a sign that players understand and work on certain patterns of structured attack, so we can also take that as a metric of change.

Then again high press is not the only way of doing pressing (f.e. middle press) and often is misunderstood in the sense that people think you should be on the edge of the opponent's box as soon as they have a goal kick. But that is not correct, teams are instructed to set traps all over the pitch to get players that are not composed on the ball into situations where they have to find a difficult solution way over their limits.

Examples in our team could be setting a trap on the side where the byline is acting as your defender and you can target f.e. Wan Bissaka (as most teams did last season) or you can trigger a press in the middle as soon as Fred is receiving a ball with his back to the goal cutting his options on his left foot (also something we saw previously). You simply have to make that player available to the player in possession making him his best option when instead they are trapped into thinking that and when the ball is played towards him you trigger a press (not before because you will ruin a trap so timing and patience are essential).

What I would like to see from our team is a mixture of those two principles: press immediately after you lose the ball and if you fail to win it back, stay in shape/positions (f.e. 4-1-4-1) to set a pressing trap. If you get too open to retain position and shape after an unsuccessful attempt of winning a ball, go for a foul to help yourself and your teammates to reorganize. All in all, we saw something similar with ETH so that is probably why he was called as a hybrid between Klopp and Pep when we announced him as a new boss.

To conclude, IMO good pressing is stopping a ball progression, making the opponent go back or sideways, very good pressing is winning a ball back and great pressing is scoring from winning a ball back.
 
If I understand the above graphic, City and Pool are the teams that it's the hardest to counter-press against?
 
mav_9me said:
If you are talking about the above data, that's not what he is saying. It only talks abt pressing immediately (with in 8s) after losing the ball. In that his data shows we are both successful and productive with the subsequent attack.

Our regular pressing is not good enough, the type where the opposition is building from the back and we are trying to funnel into a trap. Agree that, that doesn't look anywhere near good. (personally think that's due to lack of a proper pressing striker and Eriksen not good enough)
Click to expand...

Yup agreed, just trying to interpret that metric a little more and see if we can make a story with data that lines up with the eye test too.

1. Our PPDA is pretty average, we allow a bunch of passes from the opponent before we commit a defensive action.

2. Our # of high turnovers forced are really good - that means we're doing something right in forcing the ball to turnover high up the pitch.

3. The chart above says we're good at regaining the ball <8s after losing the ball.

So the only interpretation I can come up with it is that we're good at counter pressing in certain situations but not others? Or maybe that we are good at counter pressing but only employ it in certain phases of the game (e.g., when we're chasing a game or when it's 0-0 and sit back quite a bit in others)
 
Teja said:
Yup agreed, just trying to interpret that metric a little more and see if we can make a story with data that lines up with the eye test too.

1. Our PPDA is pretty average, we allow a bunch of passes from the opponent before we commit a defensive action.

2. Our # of high turnovers forced are really good - that means we're doing something right in forcing the ball to turnover high up the pitch.

3. The chart above says we're good at regaining the ball <8s after losing the ball.

So the only interpretation I can come up with it is that we're good at counter pressing in certain situations but not others? Or maybe that we are good at counter pressing but only employ it in certain phases of the game (e.g., when we're chasing a game or when it's 0-0 and sit back quite a bit in others)
Click to expand...

From what I understood counter pressing is pressing right after losing the ball. Which we are good at. Regular pressing we are not.
 
mav_9me said:
From what I understood counter pressing is pressing right after losing the ball. Which we are good at. Regular pressing we are not.
Click to expand...

Regular pressing patterns from goal kicks etc. are much easier to coach though. We were pretty good at it even under Ralf and the eye test says we're good at it under Ten Hag. Any random coach like Moyes can coach a good, structured mid block / low block.

Counter pressing successfully is much harder. You need to identify the right moments, 3-4 players around the ball need to move together to cut out options and force the turnover. Game is in a dynamic state so you can't just do what the default is for your role (i.e., CDM = I'm dropping). You have to have the entire picture in your head and make the right decision.

My opinion is that we only engage in counter pressing only some of the time and it works quite well when we do.
 
mav_9me said:
If you are talking about the above data, that's not what he is saying. It only talks abt pressing immediately (with in 8s) after losing the ball. In that his data shows we are both successful and productive with the subsequent attack.

Our regular pressing is not good enough, the type where the opposition is building from the back and we are trying to funnel into a trap. Agree that, that doesn't look anywhere near good. (personally think that's due to lack of a proper pressing striker and Eriksen not good enough)
Click to expand...

I mean we don't really try to press from the front off of goal kicks though, we press from a mid block instead. Watch how Martial and the forwards are instructed to press, they shepherd the ball into situations where a defender has to go for a long option instead of sprinting around trying to actually win the ball. Once the ball goes into the midfield is when we contract the space and try to win the ball back.
 
We certainly don't press like city and pool have done over the last few years, at least to my eyes from watching games. I think however ETH knows that you can't implement that over night, it takes a while to do that. Games have been coming so thick and fast for a long time now that ETH probably can't train for that right now. I think next season we will see a further significant improvement in our pressing start....WW coming in and Ronaldo recently departing might bump those stats up a bit as well.
 
sullydnl said:
Well, over the sample size of games that is this season so far. I'm sure we'll wax and wane over time just like those other teams have, with some other team overtaking us at some point.

The key thing though is that's we're in or around where the best teams in the league have been this season and last, so there's been an unquestionable improvement in that one facet of the game at least. And I think that's backed by the eye test as well, as we've looked better in that regard.
Click to expand...
Great to see, and long may it continue.
 
Teja said:
My opinion is that we only engage in counter pressing only some of the time and it works quite well when we do.
Click to expand...
Yeah, that's one thing to keep in mind. Those graphs don't say how often United counter-press, only that it works very well when they do. (Which in itself is good of course!)
 
Cheimoon said:
I'm not sure how you mean this, but in general: if a team is doing better than a statistical overview would suggest, you have to wonder if those stats are missing something or if the team is going through a lucky spell and can be expected to start dropping off any time now.
Click to expand...

It was tongue in cheek.

But your first point is the one I was alluding to. Statistics can't be wrong because they are facts. But people's interpretation, supposition and forecasting on those statistics can be massively flawed. And with something like sport with such a huge 'human element' like form, luck, confidence, injury etc forecasting is always prone to human error.

It's like the statistic of team not winning at an away ground for 40 years (or whatever). The statistic is true. But it has absolutely no bearing at all on the result of the next match. That's human error using a false statistic to measure a future result.
 
Mr. MUJAC said:
It was tongue in cheek.

But your first point is the one I was alluding to. Statistics can't be wrong because they are facts. But people's interpretation, supposition and forecasting on those statistics can be massively flawed. And with something like sport with such a huge 'human element' like form, luck, confidence, injury etc forecasting is always prone to human error.

It's like the statistic of team not winning at an away ground for 40 years (or whatever). The statistic is true. But it has absolutely no bearing at all on the result of the next match. That's human error using a false statistic to measure a future result.
Click to expand...
Sorry, I'm never sure with the way many people just dismiss stats. :)

Yeah, it's true that stats can never predict outcomes of individual games. It's like people thinking that it's less likely you'll roll another 6 if you just threw dice 10 times and it came out 6 each time. On the other hand, stats can indicate what sort of general trend you can expect over the next series of game - while taking into account context obviously. (like injuries, level of opposition, etc). But I suppose I'm saying nothing new here either, then. :)
 
There have been three distinct phases of Ten Hag’s tenure. The first was Brighton and Brentford where we were gung-ho on his ideals with disastrous results due to lack of suitable players and no adaptation period. The second was from Liverpool to City, where we were stodgy and pragmatic in the main, grinding out results without controlling games. Everton onwards, where Casemiro has been fully integrated and is giving us control in games has seen our underlying metrics improve enormously.

Most of the data posted so far in this thread is for the season as a whole; taking the above into account, is it any surprise that the underlying statistics are mediocre? They are an average of the three phases and as such, inevitably don't show anything special.

Below is a chart of our xG, xGA and xGD as a 5 game moving average. Each data point is an average for the specified opposition and the four games preceding them. This was done to reduce the variance of individual fixtures as xG’s power is in the longer rather than shorter term. The difference before and after City is pretty stark. Villa looks to be the outlier, though maybe it had something to do with who we had playing up-front for the last time.

YK1n7Dr.png


Lots of talk about pressing in this thread and the same principle applies here (apologies if the formatting is confusing, couldn't think of a better way to do it):

UPTO AND INCCITYPOSTCITY
ValueLeague RankValueLeague Rank
PPDA (passes allowed per defensive action)
15.09​
14th
11.42​
7th
OPPDA (opposition passes allowed per defensive action)
11.49​
12th
17.33​
3rd
DC (completed passes within 20 yards of goal, excluding crosses)
58​
joint 6th
104​
2nd
ODC (opposition completed passes within 20 yards of goal, excluding crosses)
70​
16th (!)
48​
3rd


We've gone from one of the worst pressing and easiest to play through teams, to one of the best (only just below some combination of City, Liverpool or Arsenal for all but PPDA) and that's happened in the course of 5 months, without a suitable centre forward or goalkeeper. Ten Hag is elite.
 
Last edited:
  • Like
Reactions: Cheimoon
mav_9me said:
@sifi36 excellent post
Click to expand...

Thanks mate. The analysis in my post is way more simplistic than I’d like largely because football data is almost all paid for. I’d love to be able to dig into line heights, pressing intensity and location and much more (I think they would all show the same upward trend based on the eye test), but aside from understat and FBREF, data is crazy expensive and therefore unavailable.
 
Whilst it's not a metric, pass maps are interesting ways of looking at a team's play.
  • Size of circles represent number of passes
  • Colour of lines represent how much threat comes from that link
  • Position of circles represent average position passes for a specific position are made from and received
Pass map and comparison with City last season:

ufCRqPN.jpg


  • Little to no threat from the right
  • Do not play out from the back and centre backs do very little meaningful progression
  • No width high up the pitch
  • Play much further from opponent's goal
  • Do not stretch the pitch at all, all players within the middle third in possession

Same for this season (scales are different so compare to City rather than last season):

qFqRGJF.jpg


We actually look like a modern team! Not quite an elite one, as can be seen in the comparison, but getting there.
  • Wide players get much higher and wider
  • Goalkeeper now involved in build-up
  • Players are much further apart, we stretch the pitch in possession much better
Things that still need work:
  • Someone to create from the right half space, Eriksen and Bruno both spend most of their time towards the left
  • More threat from the right hand side (this may be linked to the previous point)
  • A strong passing pivot player, we still build through the wings mostly (this is risk management, losing the ball out wide in build up is less risky when the ball turns over). This doesn't necessarily mean Casemiro is an issue. At Madrid, this was Kroos in early build-up, Casemiro would push further forward, and they'd swap once the ball had progressed into the middle and final third.
  • Centre-backs more likely to go wide rather than through the middle, again this links to the previous point
Thoughts?
 
  • Like
Reactions: Cheimoon
Opta has a statistic called shot-ending sequences started; in other words, ball recoveries that lead to chances, and Casemiro is second in the league on this. As an aside and an insight into Ten Hag’s style, United have two others in the top five – Fred (first), Fernandes (fifth).
Click to expand...

I wonder if/when we last had a player in the top 10 for this particular stat, never mind 3 of the top 5!
 


Perhaps not strictly the thread for it but how refreshing for us to be the reference point in a stat comparison like this.
 
Somehow I knew we'd win yesterday even after they scored. This is the biggest change for me.
I'm also expecting us to win at Anfield, something I sometimes wasn't even confident about under SAF.
 
bosnian_red said:


Someone posted it in the match thread. We havent conceded since open play while Casemiro, Martinez and Varane have been on the pitch. Incredible really.
Click to expand...

Works here too... Impenetrable triangle. Now let's hope it stays true tomorrow
 
Was looking at the goals scored in each half. In 41 matches this season, 28 times United scored two or more goals in a match (14 PL, 3 FA, 6 EPL, 5 Furopa)

3 times scored 2 goals in the 1st half (v Forest EPL and EFL, v Newcastle in EFL)
18 times 2 or more goals in 2nd half
5 times 3 or more goals in 2nd half
Once 4 goals in 2nd half

In 2021-22, in 49 matches, United scored 2 or more 18 times. 4 times 2 or more in the first half, 10 times 2 or more in the second half.
 
Last edited:
  • Like
Reactions: roonster09
You must log in or register to reply here.