Jack Joyce

It's hard to say honestly, it could be too strong but it could also make your press worse in some cases. The way most modern teams approach pressing is they stay in shape and wait for a pressing 'trigger', at which point they pounce on the opposition and try to win the ball back. The OI is meant to have two main use cases: Use on opposition players who have bad composure or ball playing ability, with the aim of forcing them in to cheap turnovers. Use on opposition players you absolutely don't want to give any time on the ball to e.g. Jorginho-esque playmakers. It really depends on your preferences which sort of triggers you want to use, but if you use the OI too much - you could end up making your team lose their structure too often chasing the other team around the pitch. When you've triggered a press, your team will leave their tactical positions to engage to try and win the ball back, which at times can be fine but at other times can be disastrous. Other than the structural issues that can occur, there's also the physical exertion of the team you have to factor in - if your team are told to trigger a press whenever a player receives the ball, they never get a chance to 'reset' and have a bit of a breather. We have a short-term condition model, which if expended means that your players won't be able to track back or run as well until they've had a bit of a break. It could be leading to defensive issues because your team are too tired to chase back or track their markers.

It's correct, you can't. Opposition instructions are intended to be used on a per-opposition basis, to react to specific strengths/weaknesses of the other team. Because of this they're not designed to be saved per-tactic. If you don't want your team to press the opposition backline, then that's what line of engagement is for really (mid or low block), and if you want to do a high press, then you can use the high press instruction rather than setting every opposition player to 'trigger press more often'. We'd only recommend using the pressing trigger OI on one/two key players in the opposition that you want to single out in a pressing trap. That doesn't mean it always needs to work this way though, and if you have good ideas on how to design/structure a new way of doing OIs then please do make a thread in the feature requests forum - but personally I think if there's something you want to tell your team to do, that varies per tactic rather than per opposition, then you want a new team/player instruction more than anything?

If you have some examples that you think look questionable, then please do start a bug report and we can have a look. Although wild results can happen sometimes e.g. when Liverpool's hopes of going the season unbeaten were stopped by.. a 3-0 hammering away to Watford.

This is the second time I've seen someone mention home/away game balance in their feedback, but we're yet to see any convincing evidence that there's an issue. According to our statistics, people have a win rate of 66% at home and 51% away from home so far in FM23, which is within the ranges you can see from real football above, especially considering that a lot of people don't adapt their tactics for away matches. The AI managers are a tougher challenge this year, which means that you might need to adjust your plans when you're away from home. Teams will have a plan to hurt you! Just to show the increased challenge this year - the overall win rate is 58%, when in FM22 it was 62%. It was easier to get away with just playing the same plug-and-play tactic every match before than it is now.

To an extent we do this, part of working on the ME is identifying how much we should be balancing the stats, and how much we should be focusing on 'gamefeel'. Our main priority is always going to be how the match looks and feels, and we have a lot of v knowledgeable coaches in our internal QA team who are constantly feeding back to us on a number of issues they've found. But stats are still very useful to see the bigger picture at times, and to assess the effect of fixes e.g. if you make a finishing improvement then how did that affect the global conversion rates? It's a very useful tool but never one we'd solely rely on without an eye test as well to back it up. As for certain stats, yes at the end of the day the FM gameworld has its own data provider (us!), which means at times we have different definitions to some real-world companies. OPPDA is one of them, since we tend to have higher numbers for defensive actions it means that the OPPDA number can be lower for us. But as long as the statistic makes sense within the context it is used (i.e. higher OPPDA = more intense pressing teams) then it should be fine, and we do factor this in when analysing our balance against the real world stats.

Absolutely, FM is an advanced simulation with thousands of players, staff, clubs etc. each of which with a huge amount of datapoints that affect their personality, playing ability, and more. Then we take this simulation and extrapolate it across save games that can be played for an indefinite amount of time. There's very, very few games that come close to that level of complexity, if any. It's what makes working on FM so fun and interesting, but also what makes it so difficult as well.

There's no such thing as a bug free game, and the vast majority of games are less complex than FM. We release the best ME we can within the timeframes that we have, then we use your feedback to iterate and improve so that each ME update is better than the last. This feedback is crucial, since you're all playing millions of matches, whereas we can only play a fraction of that ourselves. We do our best to take in as much feedback as possible, and make the right fixes to take the engine forward and improve realism with each iteration, which is very, very difficult! The ME will never be 'finished', we'll always have things we want to improve or add and that's part of the beauty of working on FM. But if we're saying that we can't release a match engine until it is bug free, then we'd never release one.

We're currently working towards an update before Christmas, we'll be updating before March. Communication is really important - even if its not always the news people want to hear. But believe me, everyone on the team is working hard on improvements and has been throughout the beta period and hopefully you'll see that work soon!

With the full version of FM23 now available, we thought it might be helpful if we gave you a quick update on our short-term plans for the match engine. While the current match engine is the same one that featured in the Early Access Beta, this doesn't mean that we’re not working on improvements. In fact, we're currently in the process of taking on board the feedback we received from Beta players and looking to incorporate this in a forthcoming major update. We have however, included some animation tweaks with the aim of improving immersion, which as an example, means you'll be less likely to see keepers making odd looking saves with their feet. Making changes to the match engine is a delicate and complex business. Changes that we make often result in unforeseen knock-ons that we need to chase down and fix... and those fixes then often have their own knock-on effects as well. In a recent internal update we had an engine that ‘felt’ really good and played really nicely… except that central defenders were racking up 200+ passes in some matches, which is unrealistically high. As soon as we discovered that we had to 'go back in' to address that issue without, of course, losing the parts we liked. Although we're happy with the engine in the current game, we’re confident that we can improve on it with the right update, but we need a bit more time to make sure we balance things out. Thanks for all of your feedback so far, it’s been and continues to be incredibly useful. The Match Team

See how it looks in the full release and if you're still seeing issues, start a bug thread with one/two save game examples we can dig in to as we may need some more information. But as far as I can see managers are getting random formations based on the hierarchy we set up.

The managers react to situations/scenarios much better in terms of how they use roles, team instructions and mentalities. In terms of formations, the 2nd formation is generally only really used when their team doesn't fit the 1st preferred formation. Whereas the attacking/defensive formations tend to be used a lot more depending on the situation. Our main focus this year was roles, TIs and mentalities - we do have plans to improve how managers use formations but it forms part of larger thinking for the future, and that's all I can say on that for now.

In short, the player instruction text is an adjustment on top of the team level instruction. e.g. when it says 'more direct passing' on the player, it means 'more direct than the team level', which can still be fairly short/medium passing if your team is told to play short. it's something we want to clear up and improve in the future, but the team-wide instructions do have an effect on the player, the PIs are just an adjustment on top of it. Hopefully that makes sense

When there's no data available for a manager's preferred formation, we randomise through the most popular formations in football. 4231, 433, 442 and I believe 523DM tend to have the highest chances of being picked. However 532 isn't even a possible formation for the AI managers anymore in FM23, so I'm not sure where this comment is coming from that it hasn't changed at all? Especially when FM23 doesn't have an editor out yet. There's the 5122DM in FM23, which is absolutely a valid formation and isn't necessary defensive - but it's not one of the most popular choices and as far as I'm aware, hasn't been for a long while. For reference I just checked 3 random managers with no data in game - and none of them get 5122DM/523 as a preferred formation.

I think you make a good point (in bold) about players not squaring up the defender enough, and it is something we've also felt internally. It's definitely something we want to improve in future, but it's a very difficult thing to implement. If you could start a bug report with clear examples where a player could stand up the defender face to face, that would help us a lot. But the other point to do with dribbling PIs - I'm not convinced that being able to force your players to play in an unrealistic way is necessarily the right direction for the game. In real life there must be managers telling their players to take on their man 'more often', and yet no player has more than 2.5 dribbles into the box per game or 4 successful dribbles per 90. If you could tell a player to dribble to the extent that they dribble far more than anyone else in real world football, I don't think it really helps us recreate real football which is our ultimate goal, especially when you consider that AI managers use the 'dribble more' instructions as well. Please do start this bug and I'll have a look. But we probably need to steer back to this being a general feedback thread, as it's become a bit of a focused discussion on dribbling.

It's a delicate balance, and one we always strive to improve with each iteration of the match engine. People will also have their own expectations/preferences for what is a realistic number of dribbles, or even what a dribble actually is! But in general, dribbles into the box aren't super common even at the top level in reality: The best dribblers in the Premier Division this season still all average less than 2.5 carries into the penalty area per match. That's not many! Especially when you consider that when you're watching on key/extended highlights, you're not seeing all of the attempts in the first place, only ones that lead to a good scoring opportunity. This is also reflected in the successful dribbles per game stat in general (but providers do vary on the definitions here) e.g. according to FotMob, no player has made more than 4 successful dribbles per 90 in the Premier Division this season. In fact, only 4 players have made more than 3 per 90. The reality is that nowadays, teams do recycle the ball a lot more from wide areas than they used to, teams have grown more patient over the years as they try to work a better goalscoring opportunity. You don't see many Ronaldinho-esque players in the modern game, it tends to be coached out of them now at the highest level. That's not to say there isn't improvements that can be made to dribbling in general though, in terms of feints, skills etc. which is absolutely something we want to improve one day. And if you're seeing clear situations where a player should cut inside with the ball, please do report them in the bugs forum so we can continue to tweak/balance throughout the year.

As I've said before, if you see players that regularly play at the base of a 4231/5212 in real life, with no ability to play DM in FM, then it's a research issue to be reported on the research forum.

The difference between a 4231 and 433 is fairly arbitrary, to the point where some coaches wouldn't even consider the 4231 a separate formation, instead seeing it as a variant of a 433. Pep typically plays a 433 system with two advanced central midfielders, which we'd call a 433, not a 4231. But if you wanted to play a 4231 with a roaming DM, then you can use either the roaming playmaker or the segundo volante role. As for players like Fabregas - it's very common for advanced playmakers to transition deeper as they get older e.g. Fabregas, Cazorla, Eriksen, and you have the DLP + RP roles, so I don't think it's all that limiting.

You would use a 433 with two advanced CM roles, with Rodri behind and IWBs.

We absolutely will require a save game with the set piece tactic set up and with the CB scoring a significant number of goals for us to investigate, especially since we've had many people say that they're using near post corners with no such results. With no tangible evidence at all, we have nothing to go by here. @toby14

The central midfield players in a 523 system need to act as a pivot to allow for the wingbacks to venture in to attacking areas. e.g.: Neves and Moutinho Jorginho and Kovacic Hojbjerg and Bentancur In these double pivot partnerships, you may see that one of them has more of a license to roam forward, through the use of a roaming playmaker or segundo volante role (which is by all means just a box-to-box midfielder that starts at DM). However, they are a double pivot so are fundamentally different to playing at CM. There's players who can play at CM that you wouldn't expect or want to play at the base of a 523, because of the increased defensive and positional responsibilities involved.

For the most part I agree with you, and we did add 424 (with DMs) as a selectable formation for AI managers, as well as 4222DM. But we decided to keep 442 flat as an option since 442 is such a diverse + flexible system, that we were wary of removing that one entirely. That may change in the future but for now it remains.

Just to clear things up: There were no Match Engine changes in the most recent update. It's not even possible for us to update the ME without changing the ME version number (which you can see via the game status screen). If we did, all your past matches would diverge which would be a major issue! Having said that, what rp1966 said above is true, factors outside of the match can feed into the match in different ways. However, since the only change in the last update was to player development, I think it's very unlikely there's any difference. It's more likely that your team lost 1/2 games and lost confidence, and then you've struggled to turn it around for whatever reason.

This is one of those cases where positions/formations are fairly arbitrary labels, but: 4411 - A 4231 in a mid/low block phase will take up this shape. 4141 - A 433 in a mid/low block phase will take up this shape. It's important for us to have consistency in how formations are set, which is why these are not used anymore. However as you suggested, based on the instructions/tendencies the manager has they are able to create these shapes through roles and instructions during a match.

Yes, AI managers have less formations available for selection through the researchers, and this is absolutely intentional. Less formations doesn't necessarily mean less tactical diversity though, as we've greatly improved the way AI managers choose their roles and team instructions. Bielsa has gone on record as saying there's only 10 formations in football. and any system you play is some variation of these 10 formations. We still have more than that! "Tactical systems are not of extreme importance, as the positions are not fixed, and player's roles will change within match context."

If you think you're seeing overly-defensive setups against you consistently then please report some examples in the match AI bugs forums, but: 532 isn't really inherently 8 defensive players depending on the roles, if you use wingbacks then thats 4 attackers already, but if you use mezzalas/advanced playmakers then they'll be advanced too. The AI will look to create a defensive/attacking balance through roles like this, and will consider how different roles work together.