Playing Out of Defence - Influence of Team Instructions

[CAE82]

In this post, I'll look at how team instructions (TI) affect playing out of defence, particuarly when it comes to goal keeper distribution and how players position themselves on goal kicks.

I'll use Man City and keep replaying the first game of the season, away at Brighton. I'll use 433 DM Wide with a default Gegenpressing style.

Control

The base tactic/lineup is below. For the 'control', I've removed all TI related to playing out of defence and transition instructions for the GK.

Passing Data:

Pass Map:

Goal-kick Behaviour:

Summary:

In this control match, Ederson has the lowest number of passes in the team. At goal-kicks, the team are compact on the left side of the field and he kicks long to Rodri (16, DM).

[CAE82]

Distribute to Centre Backs

We'll now add the TI - Distribute to Centre Backs.

Summary:

There is not a big difference in terms of passes, but it is clear the setup from goal kicks is completely different. The BPDs drop into the box (reflecting the rule change from a few years ago) while the wing-backs take up positions right on the touchline. The WBs are in plenty of space, potentially opening up lines of attack if the BPDs can quickly find them before they are closed down.

Edited January 30 by CAE82
Edit

[CAE82]

Distribute to Full-Backs

Summary:
Again, there is no huge difference in terms of the passing with the biggest difference being the line-up for a goal-kick. Ederson made more passes towards Gvardiol at LB, likely because he is left-footed and took the goal kciks from the left side of the box. The BPD are no longer in the box and the WBs are positioned more in-field compared to when distributing to the centre-backs. The DM and MC are also pushed ~10 yards further up the pitch, with the Brighton backline similarly dropped back ~10 yards into their own half when in the previous examples they were lined up on the half-way line.

Edited January 31 by CAE82
Edit

[CAE82]

Distribute to Playmaker

Summary:

In this example, Rodri (DM) acts as the playmaker and receives more passes (~ +20) than in the previous examples and the pass map shows Ederson playing more balls centrally. The players lineup at goal kicks similar to when distributing to full-backs, apart from Rodri (16) dropping ~10 yards deeper to pick up the ball in between the centre-backs.

[CAE82]

Distribute to Playmaker + Stay Wider Player Instruction on AMRL

In the above example, Rodri receives the ball centrally in a decent amount of space. However, the space ahead of him is fairly congested centrally, with Brighton having 6 players blocking off the passing lanes to the CM and Wingers ahead of him.

We therefore added the Stay Wider player instruction to Grealish (AML) and Foden (AMR)  who then take up wider positions at the goal kick leaving themselves in more space. If Rodri can then find a pass to them, they are in a good position to attack the space down the flanks.

** It should be noted that in the majority of the highlights, Ederson distributed the ball to the centre-backs despite the distribute to playmaker instruction. It is unclear if the 'Stay Wider' instruction interferes with his decision of who to pass to. **

Edited January 31 by CAE82
Mistake

[Skywalk3r83]

Very interesting. I always use PooD and Roll out to CB. I might change it to my FB after reading this. 

[RogerC]

Extremely interesting data! Really clear when you see it like this. Great tool to make changes during the match when the opposition doesn't allow ball progression!

[CAE82]

Distribute to Specific Team-Mate + Stay Wider Player Instruction on AMRL

Based on the previous example when Ederson seemingly ignored the 'Distribute to Playmaker' instruction and instead passed to the centre-backs, we used the Distribute to Specific Team-Mate TI and selected Rodri/DM.

At goal kicks, it can be seen that he now drops even deeper to just outside the D with the CB and WB pushing up slightly. He collects the ball in a lot of space and has chance to play a pass before he is closed down. Of course, a less able DM with poor first-touch and composure may end up losing possesion in a (very) dangerous position.

Grealish and Foden are in space, but are closed down fairly quickly by the Brighton full-backs. Nonetheless, City do have any opportunity to attack down the flanks with the WBs and AMRL.

[CAE82]

Play Out of Defence

The result here is somewhat surprising. The players line-up from the goal kick similar to how they do with 'Distribute to Centre Back' with the WBs on the touchline. However, Ederson kicks it to Gvardiol at LB who is in plenty of space and free to run down the flank.

Note how much further up the pitch Brighton are compared to when not using POOD.

** This may be as a result of Brighton using 'Prevent Short GK Distribution' in response to City using POOD **

It may therefore be beneficial to observe how the opposition line-up at GKs and if they are aggresively pressing as they are above then removing POOD and telling the goal keeper to distribute to the flanks to hit them on the break.

Below is how City lined up at a Brighton goal kick using the Prevent Short GK Distribution and maxing pressing. It can be seen that they are not as high as Brighton in the above example.

Edited January 31 by CAE82
Added new example

[Bahnzo]

One thing I've yet to figure out, is how to cover those FB's on the goal kick. I've seen the AI do it to me, but even with PI's set to mark those players with Tight Marking also ticked, they still setup like shown below with the wingers still playing far off the FB's. Winger's don't typically have the best Marking attribute, so maybe that's it, but I've certainly seen the AI have their wingers locking down my FB's on a goal kick, effectively marking them out of the game there. 

3 hours ago, CAE82 said:

 

 

 

Thanks for doing this @CAE82, that was an interesting idea and kinda neat to see how the AI reacts to different situations there. Definitely gives some thought to how to tweak playing out of the back. 

Edited January 31 by Bahnzo

[RogerC]

5 hours ago, CAE82 said:

Play Out of Defence

The result here is somewhat surprising. The players line-up from the goal kick similar to how they do with 'Distribute to Centre Back' with the WBs on the touchline. However, Ederson kicks it to Gvardiol at LB who is in plenty of space and free to run down the flank.

Note how much further up the pitch Brighton are compared to when not using POOD.

** This may be as a result of Brighton using 'Prevent Short GK Distribution' in response to City using POOD **

It may therefore be beneficial to observe how the opposition line-up at GKs and if they are aggresively pressing as they are above then removing POOD and telling the goal keeper to distribute to the flanks to hit them on the break.

Below is how City lined up at a Brighton goal kick using the Prevent Short GK Distribution and maxing pressing. It can be seen that they are not as high as Brighton in the above example.

How can one do what Brighton does in that top image? Or is it something only the AI can do? I mean pushing up so much.

[whatsupdoc]

Cool thread 

[Bahnzo]

3 hours ago, RogerC said:

How can one do what Brighton does in that top image? Or is it something only the AI can do? I mean pushing up so much.

It's using 'Prevent Short GK Distribution'  in the defensive options. 

[CAE82]

10 hours ago, RogerC said:

How can one do what Brighton does in that top image? Or is it something only the AI can do? I mean pushing up so much.

I’ve not managed to replicate thus far. I need to check OIs and formation as even by using Prevent Short GK Distribution the City attackers where not as high. Brighton were playing 4231 so possibly the addition of the AMC makes them push up higher. Or maybe their front line had PI of Close Down More. 

[Fox-7-]

Great thread!

17 ore fa, CAE82 ha scritto:

Play Out of Defence

The result here is somewhat surprising. The players line-up from the goal kick similar to how they do with 'Distribute to Centre Back' with the WBs on the touchline. However, Ederson kicks it to Gvardiol at LB who is in plenty of space and free to run down the flank.

Note how much further up the pitch Brighton are compared to when not using POOD.

** This may be as a result of Brighton using 'Prevent Short GK Distribution' in response to City using POOD **

It may therefore be beneficial to observe how the opposition line-up at GKs and if they are aggresively pressing as they are above then removing POOD and telling the goal keeper to distribute to the flanks to hit them on the break.

In this case Ederson plays at LWB, a quite direct pass despite the POOD TI.

Does his role as a SKa affect this by having take more risks hardcoded?

Did you try if changhing the GK role/duty affects the kind of passes he plays in the build-up from goal kicks?

[CAE82]

3 minutes ago, Fox-7- said:

Did you try if changhing the GK role/duty affects the kind of passes he plays in the build-up from goal kicks?

Not yet. Plan to try different roles/duties and investigate other TI. The distribute quickly/slow pace down likely have effects here e.g. outfield players having more time to get back into the structured positions. Although I’d guess this is more likely for when the ball is still in play e.g. catching a cross. Some GK obviously just flop onto the floor for a few seconds wasting time while the outfield players push up. 
 

I may also look at PPM e.g. uses long throw to start counter-attacks. 

[Bahnzo]

15 hours ago, CAE82 said:

I’ve not managed to replicate thus far. I need to check OIs and formation as even by using Prevent Short GK Distribution the City attackers where not as high. Brighton were playing 4231 so possibly the addition of the AMC makes them push up higher. Or maybe their front line had PI of Close Down More. 

This is my team, playing as a 442, defending a goal kick like Brighton. It's "prevent short GK" which does this I'm pretty sure. In this game, I'm playing a low block, so that's not what causes them to push so high up. It's got to be Prevent short GK. 

[CAE82]

@Bahnzo I’ll have a look at some point and investigate different combinations. It is clear in my example above, City do not push up as high despite having the ‘Prevent Short GK’ instruction enabled. So there must be something else involved. For example, the opposition defensive line. Maybe if they take up deeper positions at a goal kick, it allows the opposition to push further up. The strikers seem to align with the central midfielders, so may depend on if they are DM or MC etc.

Basically it needs more analysis on how team instructions affect a teams positions when defending a goal kick. 

[CAE82]

New 'Control' Tactic/Match

I've now taken over Brighton so I can control both teams formations and instructions. In the new control match, both teams are playing the same 'vanilla' 4231 DM. All instructions were removed (or sliders set in the middle).

The City goal-kick setup is below. De Bruyne (AMC) pushes up to join Haaland and City line-up in a 442-like shape. Ederson kicked long to De Bruyne who knocked it on for Haaland to actually go on and score!

Brighton have a similar setup (without the same outcome!)

Edited February 2 by CAE82
Resize

[CAE82]

Adding POOD

Not Preventing Short GK Distribution

Both teams now have POOD enabled and are not preventing short GK distribution (PSGD). City players take up the positions that we have seen before.

We'll use the pitch markings (boundary between dark and light green stripes) as an axis. Brighton's goal-line will be 0, so the keeper Steele is stood at about 2.5, their defensive line is about 6.5, and their furthest forward player is about 11.

Preventing Short GK Distribution

Both teams will now POOD and PSGD. The image below has been rotated so the teams are playing the same way as in the image above. With the addition of PSGD, the GK and Defence are in pretty much the same position and the furthest forward player is about 11.5. Based on the Brighton pitch dimensions, each stripe is around 7.5 yards, so about 5 yards further forward and closer to the DM. Probably the biggest difference is the positions of the wingers who are now probably half the distance (closer) to the full-backs. So it is clear that PSGD (rather obviously) closes down the DMs and WBs to try and limit the options of the GK when distributing. Interestingly it is the City left DM who steps up and gets closer to the Brighton AMC rather than him necessarily being closed down by the Brighton player.

Edited February 2 by CAE82
Added new example

[CAE82]

Influence of Block

Both teams are POOD and City have a low-block which seemingly has no effect. This is no surpise really as the block is classed as an out-of-possession instruction and here City are in possession.

Below is when Brighton have a High Block/Press. It seems this does not really affect the Brighton positioning.

Below is when Brighton have a high press, pressing much more often, PSGD and stepping up more.

It's clear their defensive line is now positioned on the half-way line and the attackers are ~10 yards further upfield, with the striker inside the D. I repeated the above with City playing a much higher defensive line, but that did not affect the position of the defensive line and the Brighton were still just outside the box.

Edited February 2 by CAE82
Another example.

[CAE82]

13 hours ago, Bahnzo said:

It's got to be Prevent short GK

After doing these different experiments then I agree. Although I've never seen Haaland get as close to the D as Brighton/Fati, I've come to the conclusion that it is basically randomness. There is no real difference in terms of their atributes (marking, work rate, pace etc. - indeed Haaland is better) and they are playing the exact same tactic/roles. When you watch the highlight, the players clearly do move towards the D. How close they get could be due to their starting position and the time is takes the goal keeper to take the goal kick. Obviously the closer they are at the start and the more time they have to get there, they'll more likely end up being in the D when the kick is actually taken.

[CAE82]

Player Marking

I tried using player marking instructions to see if that will get the AMRL/AMC/SC to get closer to the opposition (particuarly trying to get the wingers to lock down the full backs)

Unfortunately it doesn't seem to do much, even with PGSD. The Brighton AML (22) is still far from Kyle Walker (2) and similarly the AMR (7) and the left-back (24)

I tried again without PSGD but kept the AML and AMR marking the opposition WB tightly. I also used the OI to always mark the WB tigher. The starting position was no closer at a goal kick but it does help once the kick is taken.

Ederson once again passes to Gvardiola at left WB.

March (7) then quickly closes him down, forcing him to play it back to Ederson.

Ederson actually plays it to Walker at RB

Who is quickly closed down by the AML (22)

So it seems to stop the opposition from playing out of the back, it can be useful to instruct your wingers to tightly mark their opposite full-back.

Edited February 2 by CAE82

[Bahnzo]

7 hours ago, CAE82 said:

Player Marking

I tried using player marking instructions to see if that will get the AMRL/AMC/SC to get closer to the opposition (particuarly trying to get the wingers to lock down the full backs)

So it seems to stop the opposition from playing out of the back, it can be useful to instruct your wingers to tightly mark their opposite full-back.

I've tried so many different ways to try to get them to mark the wingers from the goal kick, but have never been able to. Like I've said, I know I've seen the AI do, so next time I see, I'll have to look. I wonder if the marking attribute has anything to do with it?

But....yeah, having your wingers mark the WB's can be very useful. I try to do it whenever there's a dangerous WB. 

[whatsupdoc]

"Prevent short distribution" is very hit and miss it seems. Here's my team's interpretation (we're in white) right before we concede: 

[Cloud9]

4 hours ago, whatsupdoc said:

"Prevent short distribution" is very hit and miss it seems. Here's my team's interpretation (we're in white) right before we concede: 

You will struggle to high press with a front 3 line this. I would recommend dropping the AF for the 4-3-3 in favor of a supporting striker capable of pressing and holding up the play. 

With both wingbacks bombing on, I'd recommend dropping the CM(a) altogether (two IF is more than enough goal scorers), for a hard working variant. A CAR would provide cover for the CWB(a) (which is the most attacking WB you could choose), in a Henderson covering for Trent manner.

Your frontline are also physically weak. Arda Guler is emblematic of the problem, he will not press well because he doesn't have the attributes to press well. If you want a properly coordinated press you're going to need some soldiers in the frontline

Edited February 3 by Cloud9