Applying the 3-2-5 / 3-box-3 / W-M style to various Premier League teams

[Aron]

In recent months, I’ve started to enjoy playing the "3-2-5 / 3-box-3 / W-M" style across my FM24 saves. I enjoy the flexibility and the idea that you are the protagonist in the game. This trend explores the system and shares the way it can be applied to different Premier league teams, starting with Bournemouth.

The basic playing style

I’m not going to go too deep history into the story of the WM formation or the 3-2-5 or 3-box-3 that is very popular nowadays. Here is rather how I have decided to implement it in my saves.

 

The basic team instructions are shorter passing, lower tempo and short goalkeeper distribution (to the CBs). This allows the team a bit of time to get into the intended shape. We squeeze the space between the lines by increasing the defensive line without increasing the engagement line (with a faster back line you could move both up one step each). I also increase the pressing by one notch, but only for formations with double wing coverage for now. I make additional adjustments to pressing and transition if my team is better than average.

The player profiles 

 

The picture shows what I’m looking for in the player profiles for each of the in the in-possession roles of the tactic. At the back, I want to have three central defenders, enabling a solidity in the rest defense (and adding a set piece threat). The outer CBs should be especially good at passing the ball forward. The central CB generally is pressed by the opposition striker and so will be unable to pass forward anyway. In front of the back line, I want to have two hard-working ball winners that are also very good at progressing the ball. These players are responsible for the midfield battle and shifting the ball up the field to our five attackers.

The top half of the midfield box will have relatively similar player profiles to the bottom half, except with an increased focus on creating and taking chances instead of the focus on winning the ball (read: dribbling, finishing, off the ball > marking, tackling, positioning). The attack is flanked by two wide forwards, who will be stretching the opposition defense and have to be able to take on the opposition fullbacks 1-on-1 and get a good quality cross in. Spearheading the attack, we need a complete forward. They need to be able to drop deep and participate in the build up as well as get on the end of crosses.

Applying this to Bournemouth in FM24

 

I assessed the Bournemouth squad based on the player profiles. Underlined are players my assistant thought are PL quality and (in brackets) are players that can cover in that role but are better at another. As you can see, the squad is a bit unbalanced with a lot of coverage for the attacking roles but much less for defenders. With the help of the transfer market, I made some adjustments by getting rid of 7 attackers, mostly on loans with options (Tavernier, Semenyo, Cook, Smith, Marcondes, Christie & Fredericks). I also signed 3 cheap defensive players to cover for the back 5 (Dardai, Frances and Wharton).  

For me, there were three main insights when going through the players. (i) The best choice for left forward would be the young left back Milos Kerkez, but luckily he is able to play in midfield to avoid having to sprint up and down the full pitch all game, (ii) one of the two defensive midfielders is the right back Max Aarons, who can not play in the DM position, and (iii) one of the best attacking midfielder options, Justin Kluivert, can not play in the CM position. Now, we could re-train them, but these restrictions also help us set a structure for our tactic.

 

This is the resulting formation. Kelly, Zabarnyi and Mepham form the CB trio. Aarons steps up and joins Adams in midfield. Traoré and Sinisterra push up the pitch and join the others in a 3-2-5. Adjustments to the formation would be to shift the wingers up or down (W-s in AM and W-at in M) and the central midfielders up or down (AM-s in AM and MEZ-at in M), depending on positional familiarity and with the aim to balance the M and AM numbers.

A typical in-possession shape, with Mepham on the ball as the right-sided defender. The opposition full-backs are stretched very wide marking the wide forwards and it is now up to our wide defenders and the defensive midfielders to provide the progressive passes to the open spaces.

Results

Finished the season mid-table with season predictions and total wages having us as a bottom half competitor. So no miracle results, but promising that the playing style did not bring the team down. Let's also look at some statistics (with the help of the moneyball spreadsheet).

These are possession-adjusted percentile ranks among PL CBs (99 players in total), with the average for these 8 metrics calculated along with the relative rank for that averaged metric. Zabarnyi and Wharton who played in the central defender role scored very low on progressive passes and possession won, as expected, as they will be hounded by the opposition striker and are the last line of defense. However, their tackle % and header % are cause for concern. The other 4 CBs (Senesi, Kelly, Frances and Mepham) played in the wide defender roles, on the left and right respectively. The main point of improvement among this group would for me be replacing Mepham with a better ball-playing defender. As for Zabarnyi, he is a young and very highly valued player so if a bid would come in I would take it.

Next, here are the numbers for the "DM/IWB" roles compared to PL LB/RB/DM players (192 players in total). The two starters (Aarons and Adams) are doing well enough for now. Billing does not really have the best attributes for this role (quite low marking, tackling and positioning) and so I would look to improve there, preferably targeting a player who can play in at least two of the related positions for the sake of flexibility (e.g. LB, RB, DM, CB).

In the attacking midfield (165 player sample, MC and AMC), the numbers are cause for concern across all four players. The complete lack of defensive and physical output and inconsistent attacking contributions mean that reinforcements will be needed. I would likely look at replacements for Brooks and Kluivert if better players are interested in a transfer.

Here are the wide forwards (127 players, LW and RW). Sinisterra has been a superstar this season, but he is of course only on loan at the club. The priority for the summer should be to sign him permanently. None of the other players have impressed. I would look for an upgrade on Scott, who is better fit for a central position anyway.

The strikers did okay (79 players, ST). Note that this player sample includes many players who played on the wing, so for a pure striker these are acceptable numbers.

To summarize, the top 3 priorities for this summer would be signing Sinisterra permanently, an upgrade on Scott on the wing and an upgrade on Mepham in the right sided defender role.

Conclusions

This wraps up Bournemouth. Next up, I think I'll try this with Burnley. Let me know what you think!

Edited January 24 by Aron

[bosque]

I like it! I use very similar team instructions but without the higher defensive line because I find that sometimes we end up compressing the space to much between the defence and midfield.

I'm not sure about 2 things. The assymetric formation, but only because I don't like to use formations the AI can't. The other is, why you need your IWB to have strenght? When team loses possession he will come back to the FB position instead of holding the mid, right? I'm asking because I don't have FM24, maybe in the new edition the IWB holds the middle a bit more before going to the FB position.

[naufal husain]

Interesting idea, looking forward to this

[Aron]

Just updated the main post with full season results, let me know what you think!

 

19 hours ago, bosque said:

I like it! I use very similar team instructions but without the higher defensive line because I find that sometimes we end up compressing the space to much between the defence and midfield.

I'm not sure about 2 things. The assymetric formation, but only because I don't like to use formations the AI can't. The other is, why you need your IWB to have strenght? When team loses possession he will come back to the FB position instead of holding the mid, right? I'm asking because I don't have FM24, maybe in the new edition the IWB holds the middle a bit more before going to the FB position.

The intention with the player profiles is to look for players that could play in multiple positions while still having the same role (e.g. IWB-s and DM-d). In this case, Tyler Adams can play both positions and has the attribute set needed for both, which meant that he could play where needed.

[Aron]

Burnley was a rough time

Applying this style to Burnley proved challenging. I will break down my attempt as follows:

1. Squad analysis: the Burnley squad was unbalanced with several restrictions
2. Tactics: no big innovations - input welcome!
3. Results: it could have been worse...

1. Squad analysis

Similarly to Bournemouth, the Burnley squad is unbalanced. Here is my overview:

Again, the underlined names are players my staff thought were PL quality at the start of the game. For a team predicted to finish only one spot below Bournemouth, the overall quality is much lower (7 PL quality players compared to Bournemouth's 20). This is a concern when trying to play proactive football and using ~20 players. Spoiler: the results will reflect this. Let's work from back to front. The central defender role was okay. For wide defenders, only one of 4 is left-sided and that player was unable to play at LB at the start - so one reinforcement was necessary. In defensive midfield, we find our only defensive-minded player who can play LB - Connor Roberts. This indicates that we would invert him from the left as an IWB. He also needs cover, so this creates the second need for reinforcement. I also shifted Cullen to attacking midfield to reinforce there, despite his very limited goalscoring ability (6 finishing and 7 long shots). On the wings we see what Burnley have been spending their money on. There are 2 players in on loan and 3 new summer recruits, making it difficult to slim this area down. In addition, Charlie Taylor is more of a wing-back profile - more on him below. Up top, there is good depth with 3 pretty good options.

The Charlie Taylor dilemma

I did some testing with tactics to fit in Taylor to the W-M style. As left midfielder, he was okay. As left wingback, I was not able to make the W-M style work. He made his run forward very late, making the in-possession shape more of a 3-3-4 rather than 3-2-5 unless we kept the ball for an extensive period of time. In addition, the shape would require a number of rotations for the formation to be reasonable. I tested a 3-4-2-1 Wide, with Taylor as the attacking wingback on the left. Roberts would invert from RWB, with Brownhill moving forward as a SV-a. Ramsey would play on the left wing as an IF, moving inside to make space for Taylor. All of these rotations meant that it would take a long while in possession before the shape turned into the intended W-M. On top of all this, none of the other wide forwards could reliably contribute defensively from the LWB position. So I gave up.

Transfers

Based on the above restrictions and the £0 starting transfer budget, sales would be necessary to reinforce defensively. I tried to offload where possible and Vitinho, Zaroury, Muric, Gudmundsson, and Taylor left as a result, with Bruun Larsen (on loan) delegated to the reserves along with young Odobert. The money went to Malang Sarr (left wide defender) and Silas Gnaka (defensive midfielder inverting from the left).

2. Tactics

I made no big tweaks compared to Bournemouth, mostly out of fear of being overly risk-averse and letting opponents dictate the game. The IWB sits to the left instead of the right since one of the left-sided wide defenders (Delcroix) was unable to play LB. Instead of the asymmetric setup (as with Bournemouth) the midfield is all nice and symmetric. In home games against other bottom half scrappers I would swap to Positive and enable counterpressing.

Happy to welcome input here on how to adjust this further for underdog squads without compromising on the style!

3. Results

I was unable to work miracles and just barely stayed up, with a pre-season relegation prediction.

Let's also look at player output compared to the rest of the league. Thanks again to the moneyball spreadsheet. Again, it is based on the Premier League players "accomplished" in the relevant positions and with >450 minutes played. The data is also adjusted for possession.

Defenders (97 player sample)

Al-Dakhil impressively posts the worst statistical performance of all compared players. The players were not good at passing forward (possibly also due to the players in front of them not that good at off the ball movement). Also, the Headers Won % stands out to me. Let me know if you have seen similar results. These players are not that bad in the air attribute-wise. Overall, the improvements should be to the two backups in Delcroix and Al-Dakhil.

Defensive midfielders (198 player sample)

First off, I noticed a detail in this table - for Possession Lost/90, a low percentile is good and not bad as with all the other statistics. Adjusted it here (by inverting it) and will update the Bournemouth post accordingly.

As with the defenders, improvements here are needed - specifically on Cork/Massengo who start in the DM position. Progressive passing is better here than for the defenders however.

Attacking midfielders (168 player sample)

The attacking contribution from this group was not great, excepting Brownhill who posted 8 goals and 10 assists. Ramsey's lacking out-of-possession data stands out to me - he has reasonable work rate. Regardless, I would look to upgrade on Cullen specifically as he does not have the attributes to put balls in the net as is needed from this role.

Wide forwards (129 player sample)

Relatively, the best group so far. I would attribute the lack of npxG primarily to the overall lack of chance creation in the team (low progressive passing, low open play key passes especially). There is a chance that bigger teams come in for Koleosho, but otherwise this is not where I would look to improve the squad.

Strikers (72 player sample)

Attribute-wise, this group should have been good enough for this level but the statistical output was bottom tier. Some of it could be contributed to subpar service from the rest of the team, sure. If money allows, I would look to upgrade here.

Conclusion

This was painful. Let me know your thoughts on setting up the tactic and what could explain the statistical outcomes. If I would continue with Burnley, my top transfer priorities would be: (i and ii) two new defenders, and (iii) a new DM. Next up, Crystal Palace.

 

[bosque]

I'm really liking this posts beacuse you are not focusing on super-tactics but trying to roleplay more with real life systems.

[Mutumba]

Everyday I come onto this site in hope to find threads like this one. Ab-so-lut-ley brilliant. Really looking forward to follow this<

 

What do you want the specific positions to achieve? I mean, like the DMs for example. Should they basically hold position, or run with the ball, or arrive late in the penalty area and so on? Should they be a DLP-d type or more Anchors? 

[Aron]

14 hours ago, Mutumba said:

Everyday I come onto this site in hope to find threads like this one. Ab-so-lut-ley brilliant. Really looking forward to follow this<

 

What do you want the specific positions to achieve? I mean, like the DMs for example. Should they basically hold position, or run with the ball, or arrive late in the penalty area and so on? Should they be a DLP-d type or more Anchors? 

Thanks!

I mention a bit about it in the "Player profiles" section in the first post. In this basic setup, the DM's primary job is to progress the ball to the attackers (by carrying or pinpoint passes) and then sit tight in the rest defense and win the ball back when we lose it.  My gut feeling is that they should be better out of possession than in possession.

Not entirely sure which specific attributes have been key for success. Comparing Cork and Adams in my two attempts, Adams outperformed Cork in every (relative) metric except for pass completion % and possession lost/90. Attribute-wise, the main difference is physical - Adams is a much faster player. While speed is not a key attribute for the DM/DLP/Anchor role, it still likely had an effect here. A faster player would have an advantage in (i) in possession, moving away from markers into open space to receive passes and free up time to turn and pass forward, and (ii) getting into position to make an interception or tackle when the opposition has the ball. That might be what we see with Adams and Cork: by the time Cork receives a pass, he can only pass it backward, leading to a higher pass completion rate and less possession lost.

Adams v Cork (Jan 2024)

Edited January 31 by Aron

[R0ca]

This is a great thread, I hope OP continues it. 

[Vizzini]

On 01/02/2024 at 15:31, R0ca said:

This is a great thread, I hope OP continues it. 

Echoing this, hope there is more to come! Brighton would be a fun side!

[limafx]

first im sorry for my bad english, your topic encouraged me to do a tatic like the image base and this is the result, i test in Benfica in 2027(my first season in the club) , there aren't many players from the original team in the squad, I won't post statistics and related things as I'm still in the middle of the season.

[Aron]

Crystal Palace time

I tried some new things to meet the high expectations from the club and supporters ("Top half" expectations with "11-14th" season preview). Here's how it went.

1. Squad analysis

The squad assessment had to be harsher than before to achieve a top half finish: underlined are players my staff thought were "Good Premier League level" or better, compared to the Bournemouth and Burnley assessments where "Decent" was enough. In addition, players with asterisks* are out with long term injuries at the start.

Defensively, the squad is fine. Andersen and Guehi are great. I ended up retraining Holding so he could play right-sided defender from a RB position. The aging Tomkins and Ward are slow and I avoided playing them. In defensive midfield, Mitchell and Clyne can invert from the LB position.

In the attack, there is a lack of depth made worse by the two long-term injuries. I used Doucoure in attacking midfield to help, but even then there are only 7 players combined for the 4 attacking midfield and wide forward roles, of which 2 are injured. Up front, Edouard and Mateta are both capable.

When looking for players to fill the gaps in the attack, I weighted mental attributes high - with the idea that intelligence is just as important as raw skill or physique if you want to control the games. Let me know your thoughts on this. Specifically, I looked for good levels of Fir, Pas, Tec, Ant, Cmp, Dec, Fla, OtB, Tea, and Vis. The players also had to be relatively flexible due to the injury situation. The additions ended up being Miranchuk (right wide forward and attacking midfield) and Livaja (left wide forward, attacking midfield and striker).

Well into pre-season, bids started coming in for Andersen and eventually I had to accept a £47M offer. To replace him, I similarly looked for an intelligent defender and signed Nehuen Perez who has excellent mentals.

2. Tactics

Similarly to the other two teams, I did not make any drastic adjustments. Notably, the right wing has switched the Mezzala from Attack to Support, with the Winger switched the other way. The idea is to have the right-sided Mezzala drop just a bit deeper to meet the right-sided defender. I am not sure if that had any impact however. In home games against non-top 6 sides, I switched to a Positive mentality.

3. Results

Looking at the final table, it becomes clear just how hard it is to achieve a "Top half" finish in this league. After accounting for the big 6 sides, there are only 3 spots left and the competition is hard with sides like Newcastle, Aston Villa and Brentford and Brighton. As a team in that group, you still get streaks of good and bad form, and the bad streaks can get you sacked quickly. We barely get there and the board ends up happy.

Now for the player evaluations.

Defenders (100 players, Tomkins does not make the 450 minute minimum)

New signing Nehuen Perez does not impress. I do see a pattern here: the wide defender stepping in from a RB/LB position tends to perform best (case in point here: Richards and Holding putting up excellent numbers from the RB position), while the player in the central defender role performs worst. At least in terms of the statistics we look at here. We can argue that the central defender has the hardest job both in terms of finding a forward pass as well as winning the ball back as the last man. Ferguson and Ward are not really PL level and that is anyway where I would look to improve.

Defensive midfielders (207 players)

Mitchell and Clyne do great inverting from LB. This category is dominated by fullbacks overall (as they pick up more defensive actions). Comparing Lerma and Hughes to other pure DMs in the list, the potential for improvement is mainly in terms of passing. The likes of Endo, Casemiro, and Partey all score much higher in progressive passes and open play key passes.

Attacking midfielders (168 players)

Note: Miranchuk played as wide forward in 11 out of his 36 appearances. The same could also be said for many others in the sample. With that in mind, 93rd and 100th out of 178 is still barely okay when trying to finish top half. Relative to the sample, these players performed about in line with my Burnley attempt. Doucoure is a better fit for DM and Ahamada is not at the level. Franca never picked up form after returning from injury either. I would look to improve here.

Wide forwards (138 players)

Olise performed great after returning from injury. The others were not as impressive. For new signing Livaja, the approach to prioritize mental attributes did not work out. He could however possibly be moved to striker instead. Schlupp and Ayew both are declining and should be replaced.

Strikers (85 players)

I might need to revisit the scoring for this group to account for many wingers being included in the sample. Erling Haaland ranks 62nd below both of my strikers as he "only" has his 98th percentile npxG/90 to show off. Meanwhile, creative wingers who contribute more defensively score high. Looking purely at npxG/90, Mateta ranks 21st. Improvement should be possible here.

Conclusion

Some patterns have emerged across these three teams. Let me know how you would better assess the performance of the central defender, especially. And your thoughts on mental attributes for wide forwards. Next, Brighton as requested! (quick edit: Brighton's board expects a mid table finish with a team predicted for 8-10th. That's so much less pressure than the Palace situation)

Edited February 13 by Aron

[bosque]

I'm really liking this! I would like to se how you create the box with the Brighton 4231.

Where do you find how your coaches rate the Premier League level of your players?

Edited February 13 by bosque