By Adam Smith on November 25th, 2013 at 9:00 pm.
Football Manager is an enormously complex simulation. On a global level, the game tracks thousands of careers, ambitions and relationships, and on any given match day, weather, morale, skills and individual personal issues can contribute to moments of brilliance or abject failure. Talking to Sports Interactive’s director Miles Jacobson, I found that the simulation model is even more elaborate in some areas than I’d expected. Read on to find out about the game’s expanding narrative engine, how climate change is forcing the team to update the code that generates weather patterns, why the ugliest aspects of football have no place in FM and how a non-contract player’s family situation might prevent him from playing for your club.
RPS: You’ve mentioned in previous interviews that you have a database of thousands of features to implement eventually. But do you have an overall vision of where the game is going to be in two or three years? Is there a shape that it’s taking?
Jacobson: I tend to work two versions ahead. It used to be three but it’s two now because we’re managing to fit in a lot more each year, so there’s always an overall vision for the game. Whether that’s a year of revolution or of evolution – I think, certainly, the revolution years are going to be less and less because there’s so much in the game already that we’d rather look at evolving certain large chunks of the game each year.
When you’re working on an annually iterative sports title that’s based on real life, it’s kind of difficult to turn around and say ‘the whole match engine is going to be 5d!’. We can’t change the rules like that. So once 3d went in, that was the year we had Che Guevara postcards going out for the year of revolution. We’re on to seemingly smaller changes now.
There are loads of changes in the game this year that are actually revolutionary but they’re not obvious when you first start playing the game. Things like Financial Fair Play, which changes the way you play the game. And something that you picked up on in your review, which I really enjoyed, was the personalities of players coming out in press conferences and the various interactions that are possible.
The underlying system of that is quite revolutionary. To the person playing the game, at first glance, it seems like an evolution. But as you get further into the game, you realise it’s a lot more powerful.
Basically, we look at all of the feedback we’re getting from different areas of the game and that will help us to steer it. While I know what I want to do for the next couple of years, that doesn’t mean it’s fixed. If an area that I believed was important to concentrate on doesn’t receive any feedback, positive or negative, then we need stats to back up if it’s even worth changing it.
That’s one of the things that we did with the press conferences in the game. We were under the impression that very few people were using the press conferences – lots of people say they find them boring – but we put a process in the game that tracks how many people are using it, provided they haven’t turned off Steam’s data collection. Off the back of that, we found out that 50% of the press conferences in the game were being attended by the manager. So we thought, if we improve that part of the game for those 50% of players, other people might also start to use it as well. That helps to steer the game.
RPS: With that kind of stat-tracking, I think you picked up on about 10% of players using Classic last year? Have there been any changes this year?
Jacobson: It was just under 10%. But to be honest, I haven’t looked yet this year. We concentrate so much on getting the game out. You also can’t learn too much from a couple of weeks of stats.
RPS: Yeah, I think a lot of people might try the full version, or Classic, and then switch later on anyhow. Find the one they’re comfortable with.
Jacobson: I tend to first look at the stats around the middle of January. There are a lot of people who get the game for Christmas. Last year we had 45,000 activations on Christmas Day, so we make sure we have tech support working then! We had very few enquiries last year though, which was good because it meant we could all get a little bit drunk (laughs).
Because we know that it’s a big game for gifting, it’s important to get stats from those people as well, rather than just the hardcore who buy it on the day of release. They’re the ones who tend to get the most immersed in the game but we need to know what people who are playing more casually are doing as well.
RPS: Just to go back to the emphasis on personal interaction and character-based instructions and conversations, that felt like a large change to me. Almost a different philosophical direction. It moves the game away from being about managing a machine and turns it toward a sort of simulation/RPG. It’s easy to see why people think the loss of sliders and precise settings might remove complexity – has it been a hard sell?
Jacobson: Yeah, that’s why it wasn’t the number one thing on our press releases! If we turn round and say “we’ve got a new narrative system” most people aren’t going to give a crap until they see that in the long-term. It’s fair to say that this year is version one of what we want to do with that narrative system. If we’d had another three months there would have been a load more in there.
There were probably more cuts in that area of the game than in any other area, in terms of what we were able to do in the time that we had and what we wanted to do. A large chunk of the 350 features that I’ve mentioned that are down for FM 15 are in that area. That’s an area that will expand.
We’ve always had good AI in the game but this takes it to another level. We’ve always been known as a game that people become immersed in, a world that people escape into, and we always look for ways to make that more believable. A lot of the changes that occur in the game are due to that. If you look at the new tactics system – we’re getting a few complaints from the hardcore wanting to use the sliders – but that old system had no real connection to the reality of football.
You don’t get people talking in training, or at pre-match talks, saying ‘today I want you to be two notches higher in creativity’. They talk in the realms of player roles and player instructions. I’m very fortunate that as part of making the game I get to go to a lot of clubs at all levels of the game, visiting training sessions and post-training meetings were they’re setting up tactics for the weekend, and this is how people talk about the game. We want to add believability, which is really important when you’re making a simulation.
RPS: When you talk about trying to be more like real football, are there parts of the real world of football that you wish you could ignore? Like the increasing importance of agents in recent years. Could any of that hinder the enjoyment of the game?
Jacobson: Certain things around the financial model, definitely. Financial Fair Play has made us look at how things really work rather than how we’d like them to work. So this year you might start off with a transfer budget of half a million quid, and set out to make some free transfers. Then your budget is gone and the reason for that is that agent fees and other aspects still come out of the transfer budget.
RPS: Do people ever report things like that as a bug?
Jacobson: I’ve had quite a few people on Twitter asking why it’s happening and I explain it and they see that it makes sense. There were real life managers who failed to understand these things. They went out and signed a bunch of players and in January asked for another £7 million but they’d spent it all on wages and agent fees. That stuff comes out from talking to people inside football but it has made things a little bit more complicated for people who don’t have maths degrees.
RPS: There are other sides of football that you don’t simulate in the game – I’m thinking of the ugly side of the game, particularly incidents of racial abuse that do, within the rules of the various football authorities, have an effect on clubs as well as individuals. Are you quite happy that you can eliminate that from your parallel football world?
Jacobson: As far as racism and homophobia goes, we’ve been a big supporter of Kick It Out and have had their logo in the game for about 14 years. There are other aspects as well, such as hooliganism and match fixing, that are part of the game in real life that we don’t want in Football Manager. Even if we did want to include them, we couldn’t because there are legal ramifications.
It’s not something that I’d like to see in our game because it’s not something that I like to see in the real world. From that perspective, I’m quite glad that it doesn’t occur in our parallel universe. And it’s also perhaps a way that we can help to steer society a little bit. If we did ever have incidents of racism, hooliganism or homophobia in the game, the punishments would be a lot harsher than they are in real life because I believe they should be. But it’s probably good that all of it is kept away because you wouldn’t want your club banned from competition because of half a dozen idiots.
RPS: It’s something that interests me because when a simulation is so in-depth there are some things that are necessarily left behind and identifying those things can be informative.
Jacobson: There are some things that happen to regen players that can’t happen to real players. You’re more likely to have a regen who doesn’t show up for training because they’ve been out on the lash, or there’s a possibility of a referee being pushed over by a regen. We can have a bit of fun with things like that later on but the real negative elements don’t belong in our more utopian world.
RPS: One of the things that I don’t think I’ve ever seen a proper discussion of is the actual nitty-gritty of the match engine. A lot of games give direct feedback when something happens. An RPG might roll a dice for an attack and show the results. How many calculations are occurring during an FM match?
Jacobson: Millions upon millions (laughs). We’ve actually optimised a lot of it this year but the basic match engine code is a little bit like spaghetti. There’s so much stuff going on in there and it’s the only area of the game that we have to do speed tests for with each new version. One tiny change can upset the balance completely.
The match engine team used to be two people and now it’s eight or nine, plus people on FMO working on engine optimisation. The amount of calculations isn’t quite infinite but it’s as close as you can get! We’re simulating every quarter of a second of football – it used to be an eighth but we’ve just changed the time splices down. Every quarter of a second, every movement of the ball and every movement of every player is calculated based on every other position, their stats and the status of the match.
I saw some interesting conversation about EA saying that with FIFA 14 on next-gen consoles, because of the extra power they can have multiple players jumping to challenge for a header. Now that hasn’t ever been a problem for us because we’re not on the console side of things. You can watch a ninety minute match in FM and apart from a few niggles that we’re working through, it does look like a real game of football. Lots of passing in midfield rather and reorganising of defensive shapes rather than just end to end stuff,. We have all of that going on.
RPS: I’ve been doing some bizarre experiments, making players with extreme stats, either 1 or 20 in odd things, and then throwing them into teams to see how they work. You can see their worst impulses pulling them around the pitch like the strings on a marionette.
Jacobson: I’m glad that you have done experiments like that! Some people argue that it doesn’t make a difference and that we rely on randomisation or predetermined results. But it’s all running real-time in front of your eyes and every tactical change will make a difference, even if it’s communicated badly to players who don’t have the motivation or football intelligence to implement it properly.
All of the freak things that can happen such as referee mistakes or injuries affect things as well obviously. And the opposition managers have their own artificial intelligence as well.
RPS: That’s a hard area to read but it seems to have improved quite a lot this year. It seems more reactive.
Jacobson: It was tweaked considerably. Managers and opposition scouts will learn about the formations you’re most likely to play and then try to counteract that. They’ll also try to make educated guesses as to which players will be filling each position. If you’re the favourites to win they’ll act differently.
RPS: Here are some slightly sillier questions, but they’re about the wider simulation and the integrity of the world. How do you work out things like the weather on match day? Do you have a database of regions and pull a seasonal figure from there?
Jacobson: We have weather patterns for every town and city in the world. So it’s based on longitude and latitude. Any place that has a football club, we have a specific weather pattern for that stadium. There are a bunch of patterns inside the game, which work seasonally, but we also look at altitude. So players in certain parts of Mexico, where the stadia are much higher, are affected by altitude much more, becoming tired during a game if they’re natural fitness is low.
All of that is taken into account. The weather patterns are being changed at the moment because that’s happening in the real world. They’ve been stagnant, in-game, for a few years, but with the changes in climate in certain places – later winters in the UK for instance – we have to mimic all of that stuff as well. It’s a huge system and it’s a side of the game that people will never see because we don’t allow it to be edited.
RPS: That’s more than I expected but I think it all goes back to how convincing the game is. People often don’t realise how much FM is recreating and how credible the simulation is. Small additions like regional referees and retiring officials add to that integrity.
Jacobson: One of my favourite little things is that referees will only officiate lower league games regionally. A referee from Southampton won’t travel to Carlisle for a non-league game. But that’s the same with non-contract players – if one has just left a club in London, he’s unlikely to join a team in the North of England because they’re not going to move their family across the country without the guarantee of a contract. So travelling to play would cost more than they’re earning from matches. Why would they do that? Those little touches make the game believable.
These are the things that attracted me to the series. I didn’t work on the first Championship Manager game and the big difference between what Champ Manager did and the other management games did is that they built a world. Some games would reset at the end of every season. You could have a team of five star players and when you got promoted, they were all two star players again. You had other games were other results were all the roll of a dice and there were no leagues outside your local league.
Champ Manager wanted to be a whole universe with everything running in the background. Unfortunately, technology hasn’t caught up with what we want to create yet. You’d need a NASA supercomputer to run every league in full detail effectively, but that’s what we’re striving to do. Those little touches are important to us and when we talk about 1,000 new features in each iteration, people might think some of them are so tiny, but they do make a difference. Not everyone realises how important that difference is when they just read the list.
Part two will follow later this week, with a conversation about open world games and Football Manager’s possible influence outside the management genre, as well as Blur, music fanzines and modding.