An example of why three-man defences struggle against three-man attacks
Who to pick up the wingers with?
ZM rarely covers anything other than top-flight football, but with 2010/11’s Premier League action not starting until next weekend, here’s an opportunity to focus on a lower league game.
The match? Exeter City v Colchester United, and it provided with a brilliant example of how three-man defences struggle when up against the 4-3-3 system. This has been covered at length before on ZM, but a case study on the subject is overdue.
First, a quick summary of the game. Exeter went 1-0 up in the seventh minute with a penalty, but found themselves 1-2 down just five minutes later after two fine Anthony Wordsworth goals. It stayed that way for the next 79 minutes, before Richard Logan poached an equaliser three minutes from time. Full highlights here.
The most fascinating aspect of the game, however, was what happened in that goalless 79 minutes. Colchester lined up with a beefy central striker, Kayode Odejayi, and he was supported by two quick outside forwards in Wordsworth and Ian Henderson. It was a fairly typical 4-3-3 – like the system favoured by Jose Mourinho at Chelsea, for example.
However, Exeter were playing with a 3-5-2 / 5-3-2 system, rather than the four-man defence we’ve become accustomed to, which made for an interesting tactical battle. The problem here is that it’s difficult for that system to deal with wingers in a 4-3-3.
If the wingers are broadly picked up by the outside centre-backs, the back three can be easily stretched across the pitch, not to mention the fact that it’s quite risky in terms of pure numbers – almost all managers favour having a one-man numerical advantage at the back. If the wingers are instead dealt with by the wing-backs, that leaves the defending side 5 v 3 at the back – fine defensively, but creating a shortfall somewhere else on the pitch.
The traditional problems arose in this contest.
The basic situation at the back
Firstly, here’s Exeter’s three-man defence (in red and white) lining up towards the end of the first half, with Exeter kicking to the left. Fairly standard given the formations – three centre-backs marking the one lone Colchester striker.
The system as a whole
There was no change at half-time – here’s a more detailed look at how their side lined up defensively for the second half. The three centre-backs in yellow, the two wing-backs in light blue, and the three central midfielders in green, with the two strikers on the halfway line.
Scenario one: 5 v 3
Towards the start of the second period, the Colchester wide players were tracked by Exeter’s wing-backs, as shown below. This meant that it was a 5 v 3 at the back, and Exeter had two spare players in the centre of their defence doing little. Fine defensively, but with Exeter 1-2 down and needing a goal, this wasn’t particularly useful for their goalscoring ambitions.
Scenario two: 3 v 3
When the wing-backs pushed forward, the Exeter centre-backs moved to pick up the Colchester wide men. Colchester were happy enough with this, as it meant they were in a good position to break quickly and run at the Exeter defence. With no spare man to sweep up, as soon as one of their forwards beat an Exeter defender, they were (in theory) in on goal straight away.
Scenario three: 3 v 1
When the Colchester wingers dropped back in the defensive phase of play, this left 3 v 1 at the back. Again, we essentially have one more defender than we need – ideally we’d have one picking up the striker, and one acting as cover. Three defenders are not needed to deal with one striker, and meant that when attacking to try and find an equaliser, Exeter had 7 v 9.
The temporary solution
Eventually, Exeter realised they needed another defender dropping in to help when the situation was 3 v3 at the back, and therefore the right wing-back started to get back more quickly when Exeter lost the ball, creating a more preferable 4 v 3 when defending.
The permanent solution
The right-sided wing-back dropping in to help out worked quite well, so what was the natural outcome? A permanent four-man defence after some substitutions, with the full-backs responsible for the wingers, and 2 v 1 against the central striker. This freed up a player to get forward into a full-time attacking role, and sure enough, Exeter grabbed an equaliser.
Conclusion
A great example of the dilemma for managers who want to play a three-man defence. That system can work well, but generally only against a two-man strikeforce. Against three forwards, you either have a surplus (with the wing-backs dropping back) or a shortfall (with the wing-backs moving forward) in defence.
The situation is summed up by this game but is also evident at the highest level of football. Some managers have started to shift between three- and four- man defences depending upon how many forwards they are facing. Both Otto Rehhagel and Marcelo Bielsa both insist on a spare man at the back – that is to say, they want 4 v 3, or 3 v 2. Never 4 v 2, or 3 v 3. Check out Rahhagel’s shifts throughout Euro 2004, or Bielsa’s switch midway through this summer’s Honduras v Chile game for a more direct example.
Defensive tactics overall are still largely based around having one spare man at the back, whether in a three- or four-man defence* – if shifting between the two systems becomes desirable, then it is another reason why we may see the rise of the ‘modern centre-half’, quite possibly resulting in something approaching Mexico’s fluid system.
*With the slight caveat that when facing a system without wide forwards in a three-man attack, full-backs in a four-man defence count as ‘half’ a defender each.
An example of why three-man defences struggle against three-man attacks
Interesting stuff, I’ve been researching some BSC Young Boys stuff for the Spurs CL play off and they seem to play 3-4-3 which appears to be more of a 5-2-3 with full backs playing on the side of the midfield four. It will be interesting to see how they cope if Spurs play 4-4-2 with 2 direct wide players in Lennon and Bale.
Interesting. The first game is in Switzerland, so if Redknapp plays with one up front they might outnumber BSC in the middle and give them a lot of trouble.
But if he plays:
–Defoe—Crouch–
Bale————-Lennon
as a front 4 it would seem to play into the hands of a 5-2-3
I’d have thought that would play into Young Boys’ hands quite nicely? Interesting to see if Redknapp switches from the 4-4-2 in Europe this season.
Should be a very interesting match-up indeed.
True that 3-4-3 would be the favourite formation of YB-coach Petkovic and he’s used it almost exclusively for a couple of seasons. But they got criticized quite a bit for their defending, often unfairly, and have started to use other formations, too. Probably not because they got criticized a lot previously, but because they didn’t find effective roles for some of their new guys in this formation. That’s a bit strange, as you’d expect such a team to make transfers fitting into their primary formation and the coach really loves that 3-4-3. So, I’d say they realized after a poor start in the league they had messed up some of their transfers and started to try different things. They were not very successful with a 3-5-2, but then beat Fenerbahce with a 4-1-4-1 and have just won their first game in the league with that formation, too.
It will be really interesting to see what kind of formation he’ll chose against Tottenham. I’d say the 3-4-3 would fit very well against the Tottenham side you describe there. But they’ve as said been more successful with a 4-1-4-1 this season and from that point of view, I can’t see them using anything else. But then again, they’ve as mentioned used it mainly because it gave them success – well, that’s cool of course – but not because the coach loves it. So, after all, he still might consider the 3-4-3 again and it probably would be the best choice indeed, but he’d almost certainly take a beating if it won’t work out.
Nice analysis of a persistent problem and good to see you attending a lower league game.
Is one other solution not to push the ’spare’ centre back in front of the other two and utilise him as a deep midfielder, allowing the other midfielders to be more attacking? I’m not sure if (and no disrespect here) Colchester or Exeter had such ball-playing defenders to implement this system, though.
This is mentioned at the end of the article. Mexico used Marquez in that role.
Speaking of Mexico and Rafa Marquez, he made his debut yesterday with the New York Red Bulls against the Chicago Fire.
New York plays a 4-4-2, but did occasionally play with three in the back, with Marquez dropping from a central midfield position to between the two center backs. And the center backs are technically above average, at least for MLS.
Nice. Will watch NYRB soon.
Yes! I noticed a few times after Marquez dropped back, Ream (left CB) would find himself in acres of space, and confidently rampage forward 30 yards w/ the ball ala Roberto Carlos (not something a typical MLS center back would be comfortable doing, to be sure). I would have to believe this is a genuine tactical novelty in MLS.
I’d wait and see. Marquez still has to get familiar with his team mates, and his fitness is not quite ready to deal with the heat and humidity. RBNY isn’t done signing players either…
Ream is a promising CB. It will be good for the MLS if he stays, but he has the potential to move to Europe.
It wouldn’t surprise me to find out that Ream grew up playing on a youth team where one of the CB’s would consistently get forward and support the attack. It has been common for the past 2 decades for youth teams to field their most technical defenders in the center in a sweeper/stopper arrangement where one of the two is expected to move into midfield and make a 3-5-2 during attack, kind of like the German teams of the 70’s. Only in the last 5 years have coaches really started moving away from this.
Yeah, I thought Marquez looked great for the Red Bulls. He may prove to be an even better addition then Henry. He was really able to dictate the tempo of the whole match from his deep midfield position plus he helped out a lot in defense. You saw how much of an impact he had because after he came off the lost a real lot. It was some of the most tactically interesting things I’d ever seen in MLS.
Yes, I think you’re right there, on both counts.
Interesting article. Although I have seen some clubs in Scotland (a good example is Hamilton Academicals) succesfully employ a 3-5-2 system against a 4-3-3.
The 3 centre backs would sit slightly deeper than normal, leaving less space for the opposing wingers to exploit. It also allows them to match 3 on 3 in the middle of the park, while keeping two out and out strikers up top.
I do agree though, at the top level, the modern centre back appears to be the answer. We wont see many ball playing centre halves in the SPL this year though!
I’ll look out for Hamilton’s games this season!
From what I’ve seen and heard of Hamilton (not an incredible amount, admittedly, outside of their games vs Celtic last season) they seem to play a 3-4-3 system – which allows them to have even more of a numerical advantage in midfield (especially if the the two outside forwards drop back to mark the opposition full-backs.) All the match reports I’ve seen of them describe them as playing a 343 (http://www.telegraph.co.uk/sport/football/leagues/scottishpremier/rangers/6962530/Hamilton-Academical-3-Rangers-3-match-report.html) and (http://scotlandonsunday.scotsman.com/stmirrenfc/St-Mirren-0–0.6110976.jp)
Nevertheless, I’m not sure how that’d work against a 4-3-3. I’ll need to watch out for them next season
Indeed they were never tested against a 4-3-3 because teams in Scotland are extremely adverse to playing another other than a 4-4-2 or a defensive 4-5-1 which is as far away from the 4-3-3 as you can tactically make it.
I want to see my team Celtic start with the 4-3-3 as soon as possible.
The Accies mix it about quite a lot, they also played 433 themselves quite a lot last year. Will be interesting to see how they shape up next year after losing a few key players (McArthur, Weslowski, McLenehan, Antione-Currier) but theyve brought in a good couple of signings too. Good few kids coming through this year too.
I’m actually a Celtic fan. MG, what do you think of the 3 centre mids for celtic this year? I would like to see ZM do a Celtic game and compare it to a top 4 Prem team, I think the glaring difference would be the lack of a ball winner in the middle. Scott Brown is a preverbial headless chicken, he just wants to fight people. We need another Lennon, Lambert or Peter Grant in there.
Configuration of personel is also an influence; there are rare example of attack-minded managers, such as Bielsa and LaVolpe, who rarely field three pure centre-backs in such a back line instead fielding defensively sound full-backs (such as Gary Medel) in the wide-centre-back position; this way, they can avoid the problem of redundancy by simply tasking either one of the wide-backs to join the attack (when facing a lone striker) and then you still have 2 v 1 at the back. Of course, it is extremely risky, and only the paciest, best-drilled teams can execute such movement without inviting a switch change in play or a counter-attack.
Really, if the analysis comes down to numbers, we’re missing 75% of what makes tactics fail or succeed. For example, in 5v3 scenario, that backline is way too flat for where the ball seems to be (wide left). There needs to be more depth to that backline and that problem exists whether you play 3,4 or 5 in the back. In the 3v3 scenario, the correct interpretation is actually 1v1, not 3v3, because the CB has no cover behind him. The CB should be pushing up and dropping back to utilize the offside law to force the attacker to play “honest”, or the two outside backs should be rotating behind the CB to make a triangle to give some depth. In the 3v1 scenario, again, the defense is too flat for where the ball seems to be, and the CB and LCB should be covering each other and the RCB to give this depth and make it more difficult for the forward to get a run on goal.
The value of the 3 back system is not in defending, because as is pointed out, a back 3 becomes a back 4 or 5 on defense just as much as a back four becomes a back 3 on attack. The back 3 gives more defensive solidity in the attacking phase than a back four with two attacking fullbacks, and allows a team to get up to 7 players into the attack without losing defensive structure, in a way that an attacking 4-4-2 or 4-3-3 doesn’t, if both fullbacks have license to get forward.
Exactly – a nominal back 3 or back 4 are both a problem if the players are not positionally aware and move to create and resolve numerical imbalances, both on attack and in defense. While the article does an excellent job of displaying the tactical questions a 3 man back line asks of a manager, we shouldn’t assume that there are no answers.
Scenario 1 5v3: one of the wingbacks shouldn’t be on the back-line, but that is determined by position of the ball.
Scenario 2 3v3: the CB(sweeper) should step up, and a midfielder should probably be filling the space in front of him. But, if this were a back 4 with attacking fullbacks who both got forward, this would at best be 3v3 anyway until the fullbacks returned to their position.
Scenario 3 3v1: What is wrong with this? Isn’t 3 in the back during attack the next tactical evolution? When Rafa Marquez drops between 2 CB’s, don’t you really just have 3 CB’s in a line?
The problem with a 3 man defense isn’t exactly the “3″, but what it means for the other positions. For example, a 3-5-2 has a problem for the side midfielder, because he needs to provide defensive cover on the weak side and width in attack, and he often faces 2v1 on the flanks. However, a 3-4-3 doesn’t necessarily have these problems.
Well, having 3 v 1 at the back isn’t particularly useful when you’re chasing a goal. Marquez invariably drops between two centre-backs who push wide and are good on the ball, you rarely just have three ‘central’ centre-backs marking one central striker.
Numbers have nothing to do with it, in truth. It’s about space and how it gets used, offensively and defensively. The team that uses space the best, or closes it down on defense the best, is the team that usually wins.
Mourinho apparently is training his Madrid team to play in three formations – 4213, a 442 diamond and also a formation with a three man defense, mainly to be used when a goal is needed with little time for more patient play. It will be interesting to keep an eye on how often he uses the last option.
I believe at times with Chelsea he utilised that third option – correct me if I’m wrong but normally removing a centre-back and chucking him up top to make his old 4-3-3 which had a lone striker, into more of a 3-3-4 (or 3-3-2-2 but erm that sounds awful).
Will ZM cover the Community shield game?
I didn’t think there was too much to write about, to be honest. And the best player on the pitch described it as a ‘pre-season friendly’. So probably not. Just looked like more of the same from last season, from both sides.
ZM, I used what I’ve learned from you guys in a tournament with my youth team a couple weeks ago. We lined up in our recently favoured 4312 but quickly switched to our other favoured formation, a 433, when we saw that the other side lined up in a 352. We won 5-2 and pretty much all the things that you guys bring up are the reasons we scored. I even wrote a ‘ZM style’ write up after the match. Watching my side decimate a similarly talented side was so gratifying, it showed the theory work in practice. The other manager blamed the loss on ‘lack of heart’ but I couldn’t help but feel bad for his players, they played with effort but never had a chance once we made the switch unless the changed something. Anyway, just wanted to say that my friend and I whom I coach with love your site and it has helped us practically in our matches. These things even work in Canadian youth soccer! Thought you’d like to know.
I was toying with the idea of sending you guys my post game write up but you probably have more important things on your plate.
Sorry for late reply! Send it to me, I’d love to read it – email is on the ‘contact’ page above.
ZM or anyone who have seen Krasic play, I just want to ask your opinion on him since it seems like we are getting him. To be honest I don’t follow the Russian league and have only watched Krasic one time only. They keep saying he is the “new Nedved” but I doubt that. They make look the same on screen haha but Pavel is one of a kind.
You do this, yet ignore Chelsea vs. Man United. I love the website and all, but even if it was essentially a friendly, it is still a glamour tie. I’d of thought such an articulate blogger would love to analyse how well Man United’s midfield 2 fared compared to Chelsea’s 3. But, each to their own. I mean, I guess the blogger isn’t looking for the readers satisfaction, no?
I’d much rather ZM continued the fantastic coverage of these ‘fringe teams’, teams that we all know many of the squad and catch the odd game but aren’t as familiar with as our home teams. The Spartak vs Rubin Kazan was a breath of fresh air as is this article.
Personally I’d much rather see more interesting coverage from these types of games than a meaningless game between two sides we see play 50+ times a season.
Anyway, great to see another mention of Marquez, really is one of a kind. I became mildly obsessed with Mexico’s set-up during the World Cup, mostly theorising on how it could be successfully integrated into the plans of a top club side. It had been a good 3 years since I had seen such refreshingly original football.
as a fan/followe.supporter of ZM and his fantastic website, I fully support any and all articles he publishes. Yes, Man U-Chelsea has been the biggets game of pre-season so far, but it won’t bring anything new in tactical terms. the variety and selectiveness by which ZM picks games/tactics to analyse is what adds quality to his work.
ZM, I cannot fully grasp the game you’re talking about because my knowledge on the teams (and their players/managers) is minimal, but it seems Exceter lacked the flexibility and fluidity sides like Mexico, Chile & Uruguay showed in the WC2010. I dont know, but doesn’t it seem easier to create this kind of system in club football rather than international? i mean, the manager has atleast 6 weeks of pre-season! i would have liked his players to have been drilled to be proactive rather than reactive, knowing when they are disadvantaged or in the driving seat.
Anyhow, I think Europe as a whole is missing out on the Copa Libertdores. it would be awesome if you could analyse the 2 finals on this site. It’ll give the readers a view on the game in Latin America, with 2 different yet equally entertaining football ideologies being represented (Chivas of Mexico, Internacional of Brazil).
Many thanks!
Ah, well that’s the beauty/awfulness of the site.
If you don’t like the content, you’re entirely welcome not to read the site.
Your final sentence has earned you a permanent place in the moderation queue, congratulations! Don’t even think about being so rude again.
Hahaha!
Good article. I was wondering if you were going to comment on any lower league football after I watched the Leeds v Derby game and Leeds’ 4-4-2 couldn’t cope with Derby’s 4-2-3-1 with the wide men drifting into the space in the middle between midfield and defense.
Actually Leeds played a 4-5-1/4-3-3 for the majority of the game. The problem for Leeds was that they didn’t have a midfielder tracking the runs of Commons.
Mexico’s system was very interesting and I would love to see it in club football ,but there aren’t many players who are as good as Marquez( I am a big fan !) in the role that he played for Mexico ,also you need “wingbacks” like salcido to pull it off (another player who really caught my eye).
I think Spurs is a team with a set of players who can try this system and it would work for them
————-Dawson—–Bassong———–
—————–King/Palacios————-
Assou-Ekoto ——————————Bale
————Modric——–Huddlestone———
Lennon———Crouch———- Defoe
:note that Lennon will be playing much higher up the pitch than usual almost as a very wide striker while defoe will look to cut inside making space for Bale however the only problem is that assou ekoto has to play on the right and he wont get much width due to Lennon eating up the space but it could work at least in Europe!
352 or 532.The key is having a ”marquez” and having a ”Zambrotta or Carlos”(I don’t think Ashley Cole for example will be as effective as a wing back).Its my favorite formation cause of the flexibility.
A really interesting analysis of my team, and one I broadly agree with (having been at the game). Exeter are an interesting team, as Paul Tisdale has, in seasons gone by, played 3-4-3, got promoted from the Conference with a mixture of 4-1-4-1 and 4-1-3-2, and promotion from League Two using 4-5-1 / 4-3-3. He does use 4-4-2 but we’ve never had many wide players on the books (until this season), so it’s not an overly common formation.
One aspect from Saturday which I noticed, that also did much to stimmy Exeter, was Colchester’s adjustment to City playing the ball from the back. In the first half (and early in the second), Paul Jones, the goalkeeper, was obviously under instructions to release the ball quickly to the wing-backs – the majority of passing attacks started there.
However, soon into the second half, Colchester worked this out and the two wide front players quickly closed down Tully and Goldbourne on the flanks, meaning Jones either had to go to a central defender or, more often than not, was forced to kick upfield, where City had little aerial threat. Because of the reasons you’ve described, there was a gap between attack and defence, meaning the second ball often fell to Colchester and Exeter couldn’t mount any kind of attack, until substitutions changed the formations a little.
Really enjoyed reading this!
Well Roy Hodgson made the point that a back-three can be forced into a five-at-the-back scenario too easily, whereas the back 4 gives you the option of leaving 2 or 3 defenders behind depending on the movement of your full-backs. Plus, in the case of the back four, these positioanl adjustments occur symmetrically or at least in a predictable manner. For the back three to become four, you need to have one of your wide-mids/wing-backs dropping back which is easier said than done (what happens if the opponents suddenly shift the play across to the other flank?). Alternatively, you can follow the approach taken by England in 1996 with the defensive midfielder dropping in whenever one of the wide-centre-backs moves out to form a makeshift back four. Hodgson said this is the only scenario whereby he would contemplate fielding a back three; arguably this is harder to drill than is the more fluid movement of a back four
Does that last idea from Hodgson not sound familiar to the Mexico system of Marquez who, strictly speaking is a 3rd centre-back but in a way is another midfielder? The blurry lines of formation discussion!
Crucial difference: Marquez slots in between/steps out from two centre-backs.
In Hodgson’s example of England ‘96, Paul Ince would drop back alongside three centre-backs (whereupon one of them would move out to the English flank that was being attacked).
You’ll notice how in the case of Mexico, the defense does not lose its symmetry (two centre backs + Marquez, 2 wing-backs).
Under the model that Hodgson cited, the defence is in danger of getting sucked across to one side unless the defensive midfielder slots into the correct position and his movement is co-ordinated with the backs.
Hodgson’s system, at least in the back, sounds a little like Van Gaal’s Ajax and the way the DM would move in and out of the back line to convert it from 3 to 4.
The easiest way to make it work is to give the DM the role of always dropping straight back into the middle to cover this space, and the CB (the guy in the middle) always has the job of sliding over to cover as second defender for the side CB who is on the flank being attacked.
But, when the DM slots back he vacates the critical space in front of the defense which may not be filled by another midfielder during quick counter attacks.
Right, except that the England team were using wing-backs, or wide-midfielders (two from McManamman, Le Saux and Anderton)in a fairly distinguishable 3-5-2.
Van Gaal’s midfield was a tight-trio, and the incorporation of the libero (Blindt or F. De Boer) converted it into something of a diamond.
That England shape was a 3-5-2 or a 3-1-4-2, if you will.
Van Gaal’s midfield in a 4-3-3/3-3-1-3 was a tight-trio, and the incorporation of the libero (Blindt or F. De Boer) converted it into something of a diamond.
Fantastic piece, very well explained.
I’d much rather read this than (another) article about a “glamour tie”!
you havent talked about how overlapping full-backs on the opposing team can cause trouble
you should be a football manager.
I think this is a poor example of a 3-5-2. Exeter must have some horrid ball retention if they allowed Colchester’s two wide men to play even with their striker and not punished them for their numerical inferiority out wide. And what the hell is the central midfielder doing leaving his back three exposed like that?