Lang thang bên Sortitoutsi.net tìm đc cái này, xem SS vô cùng ấn tượng :rang:
Introduction
New year and new season – it’s time to once again try and analyse the real Arsenal tactic, and figure out how to implement it into Football Manager so it:
- Looks like the real deal in the 3D match engine
- Gives us similar match stats in comparison to real statistics
- Provides us with results and individual performances that match reality
This is done by comparing real life match analysis with the in game match analysis tool, tweaking every position until they correspond to each other. When they match, we start comparing the 3D match engine with real life videos of games played, to see if it looks convincing. Lastly, we play longer saves to see how players performances and ratings stack up. If they perform too different to reality, we either alter their positions/individual instructions slightly, or we alter them in the database.
The goal is to create the most realistic Arsenal experience in Football Manager 2011 possible, and this article is the first in a series of Arsenal-centric puzzle pieces that will eventually form the massive Arsenal Simulation Compendium – the most in-depth simulation guide for playing a realistic Arsenal save in Football Manager.
Differences From Last Year’s 4-3-3
At the beginning of last season, Arsene Wenger moved away from his trusted 4-4-2 he’d been using and tweaking since the late 90′s. This was a tactic that had won him two doubles, and had allowed his team to go unbeaten for 49 domestic games and coming within 20 minutes of winning the club’s first ever Champions League trophy. The most important key to the success of Wenger’s 4-4-2 was the wide forward positions that allowed wingers to cut inside and play as high (or, at times, higher) up the pitch than the two central forwards. Players like Overmars, Ljungberg and Pires thrived in these positions.
After the 2008/2009 season, Wenger decided to try something new. One reason was to finally conquer Europe, where Wenger tried a more exotic 4-5-1 to adapt more to continental tactics. Another reason was to create more freedom for Cesc Fabregas, letting him roam more without compromising on the numbers in midfield.
He took the best elements from the trusted 4-4-2 -> positional fluidity, inside cutting wide forwards, highly attacking full backs and a clever possession-based short passing approach, and created a 4-3-3 that worked quite well at first. The players definitely enjoyed it more, as it allowed them greater creative freedom and more passing options.
This first version of the 4-3-3 had a fairly flat midfield, with Song as the only defensive-minded midfielder in the middle, while Cesc and Diaby where attacking midfielders. Cesc started out in the right MC slot and was the most attack-minded. Diaby started out in the left MC slot and was the link between Song and Cesc – creating a skewed midfield.
This tactic had a couple of problems, especially defensive ones. The defensive line pushed very high up, due to the full backs replacing wingers on the flanks when they cut inside, and also due to our defenders not being physically blessed enough to fully control opposition players further down the pitch. Instead, they tried pressing high to reduce any threats early on.
You can view the analysis of last years 4-3-3 here
This year, Wenger has tweaked his 4-3-3 into a 4-2-3-1, and it now features a couple of key differences to the previous tactic:
- The midfield is now a 2-1 instead of a 1-2. Instead of having Song centrally as the only holding midfielder, Wenger has essentially skewed the midfield even more, finally turning Cesc into a proper AMC
- Defensive line has been pushed down, as Arsenal now have defenders better equipped to battle in the air closer to goal.
These changes (and some good signings for once) makes this years Arsenal formation a little bit more rigid, and as a result; better defensively.
Match Analysis Comparisons
So here is the first step; making sure the players move and behave like in reality, and the best way to do this is by comparing real match analytics with the in-game ones.
New for this season is that instead of showing you three example games for each player, I’ve combined the first ten Premier League games of the season and superimposed each and every position so you have a good overview of how they move over a prolonged period. I’ve also included both passing arrows and heatmaps that shows activity percentage.
Gael Clichy & Bacary Sagna (WBL & WBR)
True wing backs, covering the whole span of the flanks. Defending, attacking, assisting, crossing…the wing backs have a key role in the Arsenal 4-2-3-1.
http://dl.dropbox.com/u/1897235/fm/m...-tactic/dl.jpg
http://dl.dropbox.com/u/1897235/fm/m...-tactic/dr.jpg
Key Facts:
• Clichy spends 49.9% passing in the opponents half, Sagna 55.6%.
• Sagna passes (crosses) almost twice as much in the top zone than Clichy.
• Sagna covers the whole length of the pitch, Clichy plays wider.
Passing Stats:
• Clichy – 58.2 average passes per game, 87% completion rate.
• Sagna – 58.3 average passes per game, 85% completion rate.
The biggest difference between the two is that Sagna gets further up the field. Walcott covers a lot of area on the entire right flank (see below), and also moves into central positions, so the reason for Sagna being further up the field is because he’s replacing and overlapping Walcott throughout the game. On the left, Arshavin is more of a classic AML/AMLC, so the left flank is slightly more cautious.
Substitutes: Kieran Gibbs is pushing Clichy for a first team place, and is the definite option if Clichy is out. If Sagna is unavailable, Eboué usually plays here, as there aren’t any other natural replacements for the DR slot. Havard Nordtveit has been playing the DR slot in the U18 team, so in the worst injury crisis he could be used there.
Thomas Vermaelen & Laurent Koscielny (DLC & DRC)
No matter who is playing in central defence, there is a difference between the DLC and the DRC positions. The DLC position is the more advanced one, pushing forward and trying to press higher up, while the DRC position is covering behind him. Normally this behaviour would be controlled by the type of player in the position, but as evidenced by the chalkboards (they take both Vermaelen and Koscielny into account for the DLC slot), it’s purely position based.
http://dl.dropbox.com/u/1897235/fm/m...tactic/dlc.jpg
http://dl.dropbox.com/u/1897235/fm/m...tactic/drc.jpg
Key Facts:
• Almost 1/10 of all passes from the central defenders are made in the opponent’s half.
• DRC spends about double the amount of time in the wing back slot than the DLC.
• DLC covers a wider area (laterally), while DRC sits slightly deeper.
Passing Stats:
• Vermaelen – 28.3 average passes per game, 99% completion rate.
• Squillaci – 33.7 average passes per game, 95% completion rate.
• Koscielny (DLC) – 25.0 average passes per game, 92% completion rate.
• Koscielny (DRC) – 30.0 average passes per game, 95% completion rate.
• Djorou – 33.0 average passes per game, 97% completion rate.
When the wing backs push up, the central defenders stretch wide to cover both the DC and DL/R positions. Koscielny (DRC) is spending almost 20% in Sagna’s defensive space (DR/WBR), and Vermaelen (DLC) is spending around 10% in Clichy’s defensive space (DL/WBL).
The DRC position is better suited for slower and taller players (Koscielny, Squillaci), while the DLC position is better for short, quick and more creative players (Vermaelen being the only one in the squad fitting this description at the moment).
I also have to mention the Djorou-Squillaci partnership – to be honest I think this duo provides more stability than the Vermaelen-Koscielny partnership. Squillaci is a traditional no-nonsense defender, and Djorou has the physical traits of a bull. Sure he’s not the quickest defender, but this partnership has led to less confusion in the backline.
Substitutes: If Vermaelen is unavailable, Koscielny takes his place, and Squillaci takes the DRC position. So far, Squillaci has always played in the DRC position. Mobile and creative defenders go in the DLC slot, while solid defenders go in the DRC slot. Players are preferred in the following order: Vermaelen, Koscielny, Squillaci, Djorou and lastly Song if there are no other alternatives available.
Jack Wilshere & Alex Song (MLC & MRC)
It’s still unclear whether or not Jack Wilshere is regarded as a first team regular. Diaby is still highly regarded by Wenger, and Denilson has been getting some action lately…but apart from Wilshere, Denilson and Diaby, there aren’t many players able to convincingly play in that MLC position.
http://dl.dropbox.com/u/1897235/fm/m...tactic/mlc.jpg
http://dl.dropbox.com/u/1897235/fm/m...tactic/mrc.jpg
Key Facts:
• More than 1/4 of Wilshere’s passes come from the AMC line.
• Half of both Song’s and Wilshere’s total passes come from an MC position.
• Song and Wilshere have very similar distribution overall, but Wilshere plays wider, and higher up.
• Wilshere has almost double the amount of passes in the AMC line than Song.
Passing Stats:
• Song – 59.1 average passes per game, 92% completion rate.
• Wilshere – 39.5 average passes per game, 91% completion rate.
• Denilson – 67.0 average passes per game, 94% completion rate.
• Diaby (MRC) – 51.5 average passes per game, 93% completion rate.
Song is still more of a classic defensive midfielder, even though he constantly travels up the field.
Song touches the ball more than Wilshere (Wilshere only does 2/3 of the average amount of passes as Song), and is more compact in the middle. These two positions do swap around a lot with each other, filling in when needed. More about this in the ‘Connections to Total Football’ section below.
Substitutes: There are three players currently competing for the MLC slot – Wilshere, Diaby and Denilson – and they seem to be preferred in that order at the moment. In the MRC slot, the only “natural” replacement for Song is Denilson, but he’s not a defensive midfielder. In pre-season, Emmanuel Frimpong really showed some class, and was regarded as the backup for Song before he tore his cruciate ligaments in mid-August.
Andrey Arshavin & Theo Walcott (AML & AMR)
As previously mentioned, Walcott covers more ground, due to his tenacious nature. In these chalkboards you can clearly see how the wide forward positions work – the wide forwards start from an AML/R position, but moves up and cut inside to become central forwards. This behaviour is also tactically encouraged by having a “false 9″ – a central striker that drops down to a FC position to draw central defenders from their positions.
http://dl.dropbox.com/u/1897235/fm/m...tactic/aml.jpg
http://dl.dropbox.com/u/1897235/fm/m...tactic/amr.jpg
Key Facts:
• As corners are counted as passes, Arshavin’s stats in the top left zone are misleading.
• 1/3 of Arshavin’s passes come from an AML/AMLC position.
• Overall, both flanks have very similar distribution, Walcott cutting inside slightly more.
Passing Stats:
• Arshavin – 30.6 average passes per game, 71% completion rate.
• Nasri (AMR) – 61.0 average passes per game, 79% completion rate.
• Walcott – 16.7 average passes per game, 74% completion rate.
• Rosicky (AMR) – 37.0 average passes per game, 87% completion rate.
In FM terms, the players usually operate somewhere between the AML/R and AMLC/RC positions, but push forward and generally play as high (or higher) up as the FC.
Walcott, or the AMR slot, is more of a traditional forward. This is true even if Nasri is playing in this position. Arshavin, or the AML slot, is not as high up the pitch.
Also worth noting is that Walcott is seeking a more central position this year. He’s moving into channels and often end up in a FC position, forcing the actual FC to replace Walcott on the right flank. We could mimic this behaviour through the swap instruction. Once again there is more info about this in the ‘Connections to Total Football’ section below.
Substitutes: Nasri and Eboué frequently replaces Walcott on the right flank, while Rosicky seems to be the first choice to replace Arshavin this year. However, when both Van Persie and Chamakh are fit, one of them might be placed on one of the flanks. Vela has also been used to replace Arshavin towards the 80 minute mark.
Cesc Fabregas & Robin Van Persie / Marouane Chamakh (AMC & FC)
This year, Cesc has been moved from a MRC position to an AMC position, probably because he had such an impact at the end of last season when he started playing in a more advanced role. The AMC position enables him to roam around, and as you can see in the chalkboard, he is the player with the widest cover of the field, playing everywhere from DMC to AMR to AML to AMC.
http://dl.dropbox.com/u/1897235/fm/m...tactic/amc.jpg
http://dl.dropbox.com/u/1897235/fm/m...-tactic/fc.jpg
Key Facts:
• Fabregas covers the whole width of the pitch, in both the AMC and MC line.
• Both Fabregas and the FC have very similar distribution patterns in the AMC line.
• The FC distributes most of his passes from an AML position.
Passing Stats:
• Cesc – 62.5 average passes per game, 83% completion rate.
• Chamakh – 21.7 average passes per game, 85% completion rate.
• Nasri (AMC) – 56.3 average passes per game, 83% completion rate.
• Diaby (AMC) – 44.5 average passes per game, 95% completion rate.
Cesc just has a 100% free role, and the key to making this work is that other players in the team replace his position if he moves into theirs. That’s why Song and Wilshere often end up close to goal – it’s usually because Cesc has come deep to get the ball, so either Song or Wilshere has replaced Cesc in the AMC slot to keep the structure of the formation intact. Once again an attribute of total football.
Robin Van Persie and Marouane Chamakh are very similar in their movements, and the latter has definitely been brought in to provide a mirror replacement for the former. Van Persie and the FC position is so important for Wenger’s tactic to work, as the false 9 role is designed to allow for the wide forwards to push up and offer attacking options at the same level as the FC.
In essence, the AMC and FC play fairly level in terms of position on the pitch, forming what could be perceived as an attacking bank of four in a 4-2-4.
Another interesting thing to note is how the FC position follows the same pattern as the wide forwards – on the left side the position is more withdrawn and on the right side it’s more advanced. A clear indication that the FC covers for the AML/R positions when they cut inside.
Substitutes: Preferred substitute for Cesc is Diaby, Nasri or Aaron Ramsey. In the FC slot, Chamakh replaces Van Persie, but if both of them are out, the best bet to fill this slot would be Bendtner.
Formation
Now that we have all the individual positions analysed, the Arsenal tactic can be described as many things. It can easily be a 4-5-1, 4-2-4, 4-2-3-1, 4-2-1-3, or even some more exotic shapes like 2-4-1-3 or 3-3-1-3. But if we just take the raw data and position the players in their average positions, the formation looks like this:
http://dl.dropbox.com/u/1897235/fm/m.../positions.jpg
It’s a tie between a 4-2-3-1 or a 4-2-1-3, but for simplicity’s sake we’ll just call it a 4-2-3-1.
Some interesting things to note: all the sideways motion of the central players (MC’s and FC) should be familiar for anyone who has watched Arsenal the last 10 years or so – those movements are causing the “pass the ball into goal” effect so often displayed on bad days.
Core Of The Formation
When trying to emulate a real life tactic, I usually try to figure out the core of a formation to better understand it. The “core” are the players who makes the most passes in the game. Most football teams today are broadly divided into two groups – the distributors and the receivers – and it’s usually balanced evenly (5-5).
To figure out the core, we take a look at where the most frequent passes are made, and from which positions. This is important for two reasons, and by figuring out where most of the passes come from, we can:
- Draw conclusions on what player instructions and preferred moves we need to apply.
- Better understand how the tactic works. Example: if most passes come from the wide forwards, we know that in FM terms, we need to set the passing instruction to “Down Both Flanks” in order to have the wingers more active.
In the Arsenal 4-2-3-1, most passes come from Clichy, Sagna, Song, Wilshere and Cesc, and if we take a look at it like this…
http://dl.dropbox.com/u/1897235/fm/m...c/corefive.jpg
…we see that at the core, Arsenal are distributing passes from a fairly flat midfield five, sacrificing some defensive stability by pushing the full backs up to form a 2-4-1-3 at times. The reasoning behind it is to have the benefits of a 5-man midfield – being able to outnumber the opposition in midfield increases your possession…but also have the benefits of a 3-man forward line – creating triangles and balancing the numbers in attack vs defence.
The way it works is that by distributing passes from a width of a 5-man midfield, Arsenal are able to bring versatility to their passing. The full backs offer more direct passing and one-two’s on the flanks, and the midfield can switch play from flank to flank and serve through balls for the front three.
For comparison purposes, we can take a look what the core of Manchester City looks like:
http://dl.dropbox.com/u/1897235/fm/m...ty-passing.jpg
A completely different approach. Arsenal are trying to distribute the bulk of their passes through width, while Man City are trying to distribute their passes from a defensive center. More about the Man City tactic in another article, coming soon.
Possession Much?
Another thing we have to do in order to re-create Arsenal’s style of play in FM is to figure out how and why Arsenal have so much ball possession. There are a couple of reasons:
Another thing we have to do in order to re-create Arsenal’s style of play in FM is to figure out how and why Arsenal have so much ball possession. There are a couple of reasons:
- Positional awareness – making yourself available for a pass. According to Wenger himself, a player should always have at least three clear passing options, at all times. The result is that Arsenal can keep possession for a long time (on a good day), since there is always an option to pass. Keyword being triangles.
- Probing play – patience is key. Contrary to popular belief, Arsenal plays a short passing game, with some players allowed to be more direct if they think it’s necessary. They’d rather keep possession than take a gamble.
The first point is very interesting to analyse. If we draw lines between every players short pass options, we get a whole lot of triangles:
http://dl.dropbox.com/u/1897235/fm/m.../triangles.jpg
By creating lines like this, we can see why the core have the most passes: they have the most options available, both to receive and distribute. The midfield trio all have at least 6 short passing options at any given time, and that makes it easier for them to distribute passes. It also allows them to move more freely, since even if one midfielder is out of position, the other midfielders have at least 5 clear options.
Now the most important thing to note here is that the default 11.1 database will NOT get you realistic possession and pass completion results. The reason is that pass completion and possession are mainly controlled by two attributes: passing and decision. Both of them have to be pretty high to get the 60/40 possession and 85% completion rate results we need to make it realistic. So if you want realistic results; apply the DS2 database, which is a custom Arsenal database. More details here.
To make this work in FM, we need to make sure the players in the database have good enough Off The Ball, Passing, Anticipation, Decision and Position attributes, and we also need to set the passing to short across the board.
Real life passing stats:
• Liverpool 1-1 Arsenal = 511 passes in total, 88% pass completion, 64% possession
• Arsenal 6-0 Blackpool = 554 passes in total, 89% pass completion, 54% possession
• Blackburn 1-2 Arsenal = 458 passes in total, 84% pass completion, 64% possession
• Arsenal 4-1 Bolton = 523 passes in total, 89% pass completion, 66% possession
• Sunderland 1-1 Arsenal = 305 passes in total, 84% pass completion, 46% possession
• Arsenal 2-3 West Bromwich Albion = 545 passes in total, 83% pass completion, 65% possession
• Chelsea 2-0 Arsenal = 398 passes in total, 86% pass completion, 52% possession
• Arsenal 2-1 Birmingham = 534 passes in total, 88% pass completion, 58% possession
• Man City 0-3 Arsenal = 612 passes in total, 90% pass completion, 65% possession
• Arsenal 1-0 West Ham = 501 passes in total, 87% pass completion, 66% possession
So on average, Arsenal has 60% possession, 86.8% pass completion rate and 494 passes per game. We won’t get that amount of passes per game in FM2011, due to it counting passes slightly different than the Guardian Chalkboards, but with the custom DS2 database you should definitely come close to the 60/40 possession and 86% pass completion rate.