IGN member speak the truth about Football manager

[mantralux]

The FM match engine generates 45 minutes of football in approx. 2 seconds at the start of every match.

FPS have 5 people on screen who need to decide whether to duck behind cover or not.

It's hilarious that you actually think this is true.

EDIT: Sorry, that was too short to be deemed constructive. Let me put it this way: it now makes sense how you are dazzled by Sports Interactive and the code in FM.

[Ackter]

And it's incredible if you genuinely took that last sentence seriously.

[Ackter]

Oh while I remember, do you even know what legacy code is?

Do you also realise it's present in pretty much every PC and Mac game?

[mantralux]

Oh while I remember, do you even know what legacy code is?

Yes.

Do you also realise it's present in pretty much every PC and Mac game?

Yes.

Do you realise that the amount of legacy code can make a huge difference in how much freedom you have while coding? Have you ever coded anything? Do you realise that Sports Interactive have to, at some stage, re-write the game from scratch to keep up with new technology and ageing software? I mean, the comments you've made about FPS AI in this thread are mind-boggling. Is it the "friend from Capcom" that taught you about AI in first-person shooters?

[themadsheep2001]

And it's incredible if you genuinely took that last sentence seriously.

Obviously you were being very tongue in cheek, but interested to how know it differs between FM and and a shooter like say, Ghost Recon. Probably a very general question, but I'm curious.

[mantralux]

Obviously you were being very tongue in cheek, but interested to how know it differs between FM and and a shooter like say, Ghost Recon. Probably a very general question, but I'm curious.

Have to be way more specific than that. Obviously there's a huge difference when coding a turn-based game, as it's easier to let the user select X amount of actions before rules kick in. With a real-time game like a FPS, you need to instantly apply rules for a ton of actions impossible to predict beforehand.

[suikoden]

Just to brag I was 15th (out of a few hundred thousand players), before I stopped playing due to real world stuffs, on SEA server in Starcraft 2 so I am pretty good at the game and the AI is terrible. It doesn't need to be good. SC2 is about multiplayer. Lets move on to Diablo 3. Terrible AI also with crazy monster affixes dominating you in inferno instead of good AI. I also remember playing FEAR when it was released and how everyone was touting its AI as amazing and stuff. And it was for that 'era' with enemies trying to flank you etc...But it still wasn't that good. Stealth games like Thief and MGS has very basic AI also. Those type of games are more like puzzle games where you are suppose to know how the AI acts and it is up to you to solve the puzzle by getting from A to B.

FM has good AI. It should be better but it is not super crap. Still I hope it is a priority for SI along with the ME. I feel the tactics system is too restricting.

On another note I feel uneasy at all the accusations that SI are lazy and complacent just because they don't have any competitors. The games' industry is unique to other industries such as one that makes homogenous product in that you don't necessarily need a competitor to want to do better. I doubt those who bash SI on this actually know whether the developers are being lazy due to lack of competition or they are working hard because they are passionate enough. Until you know I think you should refrain from making such accusations.

My opinion is that they aren't being complacent. The dev cycle of releasing a game every year is incredibly short and the changes made year by year are decent enough for me. You can't expect revolution every year. Most of the yearly sports titles I play are far behind SIgames in the year by year differences. It is amazing that there are those that expect SI to go far and above the industry norm.

[suikoden]

Have to be way more specific than that. Obviously there's a huge difference when coding a turn-based game, as it's easier to let the user select X amount of actions before rules kick in. With a real-time game like a FPS, you need to instantly apply rules for a ton of actions impossible to predict beforehand.

It is not easier it is just different. Yes AI in FPS is really advanced in recent years. No longer are AI following patterns and responding with basic maneuvers. But they still follow a set rule. They don't instantly apply rules for tonnes of actions. They mostly work on pathfinding and line of sight, pre-processing of terrain and obstacles in making decisions. Not to mention FPS AI is still extremely restrictive and difficult. The AI of an enemy soldier is top notch in the first 20 or so seconds but after that you get a noticeable degradation of AI behaviour as the options open to the AI gets bigger and the queries become more complex. Which is why you still get a lot of FPS games giving you set pieces where you fight in a small area then move onto the next one. And where you can kill enemies fairly quickly. Quantity over quality. You don't get long shootout duels in FPS and when you do the 'boss' follows a pattern anyway.

In turnbase AI there are different considerations. There is the actions of the player to consider and also the simulated world. It isn't just the AI will do a list of xyz just because abc triggered it. The AI algorithm has to take into account priorities. What do I give up for that player or which castle should I storm? I need resource A, B, C but which do I farm first? It is the most basic of queries but in a simulated world the complexity of simple things ramp up very quickly eg. instead of a ladder you would have a weird looking tree

[wwfan]

Again; it's code. It's not magic. The talent lies in creating the algorithms that control what you experience as game behaviour. The true beauty in FM lies in what PaulC does with the match engine. As in; how 22 code elements should behave to make it look like an actual football match is taking place on your screen. But even the ME is being re-written from scratch, so obviously that wasn't as good as it could have been either.

I know it is code, not magic. But thanks for the condescension.

I'd agree that the ME is a work of genius. However, Paul has gone on record stating exactly how difficult to get the balance right in the AI squad building / transfer modules and that he's glad he doesn't work on them anymore. My own interpretation of AI squad building problems is that it rests somewhere in the recognition and reputation of players between the ages of 20 and 24. Unless they are already world class (which most are not) the AI teams tend not to play them, which means they don't develop properly. They won't got to lower rep clubs on loan for a few years prior to being good enough for the big clubs. They just sit in the reserves and stagnate. Users have a huge advantage as they can recognise the quality of these players, buy them cheap, develop them and profit monetarily or on the pitch. This seems to repeat in all iterations of FM. Perhaps you could suggest a solution? Or, if you think I'm wrong, explain which parts of the squad building / transfer modules are causing the imbalance?

The ME is also not being rewritten from scratch. I don't know where you got that from. Paul stated he needed at lease two development cycles (if not more) to work on core elements in the ME. He believed that he should do it now as the 2011/12 ME was well balanced, meaning it would remain robust across iterations and that it was a good starting point for these key changes.

[Kriss]

On another note I feel uneasy at all the accusations that SI are lazy and complacent just because they don't have any competitors. The games' industry is unique to other industries such as one that makes homogenous product in that you don't necessarily need a competitor to want to do better. I doubt those who bash SI on this actually know whether the developers are being lazy due to lack of competition or they are working hard because they are passionate enough. Until you know I think you should refrain from making such accusations.

My opinion is that they aren't being complacent. The dev cycle of releasing a game every year is incredibly short and the changes made year by year are decent enough for me. You can't expect revolution every year. Most of the yearly sports titles I play are far behind SIgames in the year by year differences. It is amazing that there are those that expect SI to go far and above the industry norm.

Finally, someone hits all the nails on the head, SI as an entity aren't lazy, as individuals they aren't allowed to be.

They try just as hard now as they did when there was some kind of competition.

They don't get everything right, they can't make us all happy because we have too many different personal priorities, nevertheless they won't stop trying to achieve both.

I certainly would like to see a vast upgrade of AI squad management which endures into a long career game, I'm well aware what a mammoth task that is because the AI isn't and can't be sentient.

[x42bn6]

The games' industry is unique to other industries such as one that makes homogenous product in that you don't necessarily need a competitor to want to do better.

Where did this one come from?

In every industry, there is a desire to improve, even in a monopoly.

But in industries like this one, you need competition to push everyone that last extra mile.

I'm sure SI work hard, but I'm sure they'd work harder if they had an equal rival (and/or rivals) where one slip-up could hit their paycheque substantially.

[CaptainPlanet]

On another note I feel uneasy at all the accusations that SI are lazy and complacent just because they don't have any competitors. The games' industry is unique to other industries such as one that makes homogenous product in that you don't necessarily need a competitor to want to do better. I doubt those who bash SI on this actually know whether the developers are being lazy due to lack of competition or they are working hard because they are passionate enough. Until you know I think you should refrain from making such accusations.

My opinion is that they aren't being complacent. The dev cycle of releasing a game every year is incredibly short and the changes made year by year are decent enough for me. You can't expect revolution every year. Most of the yearly sports titles I play are far behind SIgames in the year by year differences. It is amazing that there are those that expect SI to go far and above the industry norm.

Complacency is the wrong word really, but I can see why people use it.

Competition does help to up performance, the footballing world often cite bringing in players in certain positions to push those who think they are already trying hard as they can, to try that little bit harder.

Specifically when it comes to SI, when they are having development decisions it makes them ask different questions. You have to do different things to retain and gain customers from the lure of competitive products, compared to retaining or gaining them from nothing.

[heathxxx]

I would love a sparkling newly written game in 64 bit, better me, 3d engine and AI, I would also like to win the lotto, marry Ponelope Cruz and have a brewery...

That's rather frightening, as you're in Australia, thinking the same as me here in Spain, right down to Penelope Cruz and a brewery!

[swisso]

Where did this one come from?

In every industry, there is a desire to improve, even in a monopoly.

But in industries like this one, you need competition to push everyone that last extra mile.

I'm sure SI work hard, but I'm sure they'd work harder if they had an equal rival (and/or rivals) where one slip-up could hit their paycheque substantially.

Motivation by fear? I think the team is motivated enough by pride and passion for their product.

You only have to look at the Research team for the game, unpaid volunteers like myself, to know that there is passion out there to make it the best it can be.

[Äktsjon Männ]

Competition in this case could be a double edged sword really. Sure it could motivate SI to work even harder but on the other hand it would bring an even bigger need for shiny new features (read: pointless fluff to help sell copies). Ultimately it's a moot discussion anyway as there (probably) never will be a competitor. The time and resources needed to build a database that could compete with SI's is way too much for anyone to bother.

[Spurs_Q8]

What is 'AI' mean ?

[wally13]

What is 'AI' mean ?

Artificial Intelligence

[x42bn6]

Motivation by fear? I think the team is motivated enough by pride and passion for their product.

I'm sure the pride and passion are there, but competition would have even more of an effect. For example, tighter deadlines would have to be hit and slippages will be costly.

There are plenty of other factors that determine the quality of the end product, and competition is part of that.

There's nothing wrong with "motivation by fear" in this regard. If you are proud of your final product, you aren't going to want it to lose to your competitor. Which will make you even more motivated.

You only have to look at the Research team for the game, unpaid volunteers like myself, to know that there is passion out there to make it the best it can be.

Unpaid volunteers hardly have to worry about pay day in this regard. Passion is all you have to go on.

Competition, however, might have even more benefits to the consumer. For example, if there was a competitor with an even better database, SI might have to hire (more?) paid researchers. But of course unpaid researchers would still be needed (and welcomed).

[marty78]

All you genius' should get together and provide SI with the competition they need or apply to work for SI yourselves and help deliver this more advanced AI of which you speak. Should be easy enough for people with such knowledge. I can understand wanting an improved product but questioning the motivation of people you don't know is insulting.

I have no-one competing against me for my job yet I still manage to motivate myself to do the best job I can as do most of those I work with. The ones that don't soon get told or moved on.

[x42bn6]

All you genius' should get together and provide SI with the competition they need or apply to work for SI yourselves and help deliver this more advanced AI of which you speak. Should be easy enough for people with such knowledge. I can understand wanting an improved product but questioning the motivation of people you don't know is insulting.

Nobody is questioning the motivation of the staff. It is just that competition often brings out a little bit more in terms of the quality of the end-product, which is a boon for customers.

I have no-one competing against me for my job yet I still manage to motivate myself to do the best job I can as do most of those I work with. The ones that don't soon get told or moved on.

You might not have direct competition yourself, but your company has competition.

And indeed, I would bet you would work even harder if you did have someone competing for your job.

[wwfan]

I'm sure the pride and passion are there, but competition would have even more of an effect. For example, tighter deadlines would have to be hit and slippages will be costly.

There are plenty of other factors that determine the quality of the end product, and competition is part of that.

There's nothing wrong with "motivation by fear" in this regard. If you are proud of your final product, you aren't going to want it to lose to your competitor. Which will make you even more motivated.

I've never thought of it this way. You've completely opened my eyes to a new way of motivating creatives. Fear of losing their jobs is obviously going to motivate them to take innovative risks. Csikszentmihalyi, eat your heart out.

[mantralux]

I haven't seen anyone call SI lazy or complacent in this thread, but I guess you have to invent arguments when the discussion isn't going in the direction you want it to go.

[suikoden]

Where did this one come from?

In every industry, there is a desire to improve, even in a monopoly.

But in industries like this one, you need competition to push everyone that last extra mile.

I'm sure SI work hard, but I'm sure they'd work harder if they had an equal rival (and/or rivals) where one slip-up could hit their paycheque substantially.

Kinda wrong. Monopolies improve their products in line with profit maximisation. Mostly monopolies try to reach economies of scale and they try to reduce costs instead of improving their product. Sure they may improve their product but only to provide barriers to entry and at the most to keep the customer satisfied. The whole point of a monopoly is that there is no competition. You do not ever get a monopoly improving their product for the sake of improvement. So you are pretty wrong there.

You are also wrong about that last extra mile. That is an skewed mentality to take, expecting perfection and ultimate sacrifice. Except these are people with lives outside of work and family. Even the best indy developers would not push that last extra mile when they become established. It just isn't possible to keep going at that pace forever. No company in ANY industry does that. It is that sort of expectation that lets yourself down, not the company letting you down.

And I am not saying that competition wouldn't do SI any good. It will probably make them better. But I just think it is folly to think the product SI puts out is below par for lack of effort. I don't think the product is below par and I do not think there is a gross lack of effort either. Also one slip-up hitting their paycheque substantially comment is why a lot of games do not take risks. In that sense perhaps it is better FM does not have any real competition. It lets SI push the boundaries and not worry about having slip-ups. The caveat is, of course, that they are willing to be innovative in the first place.

[suikoden]

I haven't seen anyone call SI lazy or complacent in this thread, but I guess you have to invent arguments when the discussion isn't going in the direction you want it to go.

Kinda like dropping the whole coding thing for FPS when you don't have a reply. Like saying a baseless and wrong comment like turn base AI is easier and not extrapolating on that at all.

Not to mention there has been posts calling SI lazy AND complacent ("safer than safe" to mention one). RBkalle has a few. Please don't accuse people of inventing arguments unless you are going to get it spot on. It is many things not good.

[bumchin]

every game is different for each of us.this reviewer probably lives in a bedsit,five knuckle shuffle palace.

[wwfan]

I haven't seen anyone call SI lazy or complacent in this thread, but I guess you have to invent arguments when the discussion isn't going in the direction you want it to go.

If they aren't lazy or complacent, then they must be incompetent? Given the problem of AI squad building issue is so blatant and the solution so obvious, yet they continually fail to find it, it must be one of the above?

[pigfacemonkeyman]

Ok, I'm now taking bets on when Miles will post in this thread.

Monday morning is 2/5 favourite.

[suikoden]

Because there aren't any options. There is no competition.

And it's not like the game is unplayable. I still enjoy it, but it could be so much better.

Forgot to add this. You said it yourself they are being lazy or/and complacent. Or is the "there is no competition" suppose to imply something different? Would love to hear it.

[marty78]

Nobody is questioning the motivation of the staff. It is just that competition often brings out a little bit more in terms of the quality of the end-product, which is a boon for customers.

You might not have direct competition yourself, but your company has competition.

And indeed, I would bet you would work even harder if you did have someone competing for your job.

I would more likely be more motivated by more money or being recognized as the best in my field. If I was weaker than others at the same job and feared for my job I would look elsewhere to a job that suited my ability.

Why is there no realistic competition? Could it be because SI and Sega are the only ones prepared to put resources and effort into a single game for what profit that is worth? I would imagine the competition, such as EA, have considered trying to take it on and deemed it not to be worth it, what do you think? If it was just a case of getting their staff to work harder then there should be several competitors. You can work staff as hard as you want but without the knowledge/resources it is pointless.

I don't personally believe that this lack of competition should have any bearing on the motivation of SI to produce the best game they can year on year. Just an opinion though.

[swisso]

I'm sure the pride and passion are there, but competition would have even more of an effect. For example, tighter deadlines would have to be hit and slippages will be costly.

There are plenty of other factors that determine the quality of the end product, and competition is part of that.

There's nothing wrong with "motivation by fear" in this regard. If you are proud of your final product, you aren't going to want it to lose to your competitor. Which will make you even more motivated.



Unpaid volunteers hardly have to worry about pay day in this regard. Passion is all you have to go on.

Competition, however, might have even more benefits to the consumer. For example, if there was a competitor with an even better database, SI might have to hire (more?) paid researchers. But of course unpaid researchers would still be needed (and welcomed).

You're right of course in that competition is a factor in ones endeavour to improve and strive to be the best, but my opinion is that it's not a big a factor as you would have it in the case of Football Manager. I guess the difference in opinion comes down to my belief that the team at SI needs no more motivation to make FM the best it can be, other than their passion for Football and software development and their pride in their end product.

This point isn't regarding yourself, but i'm sure a lot of the Researchers wouldn't be as motivated to provide top quality data from U18s to the First Team squad if they didn't think the guys they were supplying it to didn't utilise it by producing the best game they could. A lot of us scout our clubs from top to bottom to provide the community with the best experience they can have in FM, safe in the knowledge that the guys in London are equally, if not more, driven in their goals

[x42bn6]

And I am not saying that competition wouldn't do SI any good. It will probably make them better. But I just think it is folly to think the product SI puts out is below par for lack of effort. I don't think the product is below par and I do not think there is a gross lack of effort either. Also one slip-up hitting their paycheque substantially comment is why a lot of games do not take risks. In that sense perhaps it is better FM does not have any real competition. It lets SI push the boundaries and not worry about having slip-ups. The caveat is, of course, that they are willing to be innovative in the first place.

Why is it better in that sense if there is no competition? If SI take a risk and slip-up, then there is always SI's competitor to fall back on.

If they aren't lazy or complacent, then they must be incompetent? Given the problem of AI squad building issue is so blatant and the solution so obvious, yet they continually fail to find it, it must be one of the above?

It's more like "there is no evidence to suggest that SI are lazy or complacent", which leaves that question open.

I would more likely be more motivated by more money or being recognized as the best in my field. If I was weaker than others at the same job and feared for my job I would look elsewhere to a job that suited my ability.

Which would benefit customers at the end of the day if a more appropriate person took your role.

Why is there no realistic competition? Could it be because SI and Sega are the only ones prepared to put resources and effort into a single game for what profit that is worth? I would imagine the competition, such as EA, have considered trying to take it on and deemed it not to be worth it, what do you think? If it was just a case of getting their staff to work harder then there should be several competitors. You can work staff as hard as you want but without the knowledge/resources it is pointless.

There's no competition because FIFA Manager's engine has been the same since forever and Championship Manager is now finished due to a bizarre pricing scheme and poor product. Nothing to do with the desire to commit resources or anything.

I don't personally believe that this lack of competition should have any bearing on the motivation of SI to produce the best game they can year on year. Just an opinion though.

It doesn't, but then again has there been any serious suggestion there's a lack of commitment or motivation? Only that competition can surely make things better, especially fairly glaring issues like squad building.

[suikoden]

"Why is it better in that sense if there is no competition? If SI take a risk and slip-up, then there is always SI's competitor to fall back on." That is assuming the competition is actually really good. A possible scenario is that both franchises stagnate.

I do respect your opinion that competition can surely make things better. Mine is that even if there was competition, it would be marginally better at best. I have seen other gaming franchises become bad once there is real competition (2/3 horse race) as they all release similar products and it gets quite crappy.

[wwfan]

It's more like "there is no evidence to suggest that SI are lazy or complacent", which leaves that question open.

To be honest, I don't think mantralux or yourself can escape the argument that you are accusing SI of being either lazy, complacent or incompetent. You have both stated that the problems (that we've all agreed the game has) are easy to solve. If they are, and SI aren't solving them, then, by definition, you are accusing them of not being up to the task.

Of course, they might be really difficult to solve, which suggests something else.

It doesn't, but then again has there been any serious suggestion there's a lack of commitment or motivation? Only that competition can surely make things better, especially fairly glaring issues like squad building.

Can or surely? Different things. Also, in major doubt. There has never been any serious competition, so how can you have any idea what it might do? You are extrapolating a claim from a contested and extremely contemporary theory of motivation and presenting it as "truth".

[x42bn6]

I've never thought of it this way. You've completely opened my eyes to a new way of motivating creatives. Fear of losing their jobs is obviously going to motivate them to take innovative risks. Csikszentmihalyi, eat your heart out.

If you'd've read the initial quote, "paycheque" was being used literally as a metaphor for a company's profits.

Having a competitor is naturally bad for anyone in the company when it comes to job security. However, having a competitor means the company as a whole has to naturally take risks to differentiate themselves. Risk-reward and all that.

Which is what I mean by "motivation by fear" as not being wrong in this regard - the margin for error is slimmer when you have competition, which increases the pressure on the employees, but the end result is that consumers benefit. From a customer's perspective, that's a win.

Even though "motivation by fear" is a metaphor, even if it were true, you certainly don't motivate your employees that way. If the company is struggling, you don't motivate your employees with threats. A company has stakeholders and employees, and it's obvious that both sets of people have competing desires in the company, and hence need different messages. If there is competition, the stakeholders will demand increased profits and market share, but only an idiot CEO would use this as the sole motivational tool for his employees.

[marty78]

There's no competition because FIFA Manager's engine has been the same since forever and Championship Manager is now finished due to a bizarre pricing scheme and poor product. Nothing to do with the desire to commit resources or anything.

Why did these not provide competition given it is not a question of resources or desire? Why did Fifa not just create a new match engine given that they had the resources and desire to be a realistic competitor? They clearly had competition to motivate them in the form of SI/Sega. I don't understand why they failed given what I've learnt today in this thread.

[wwfan]

If you'd've read the initial quote, "paycheque" was being used literally as a metaphor for a company's profits.

Ah. The literal metaphor. So foolish of me not to pick that up.

Having a competitor is naturally bad for anyone in the company when it comes to job security. However, having a competitor means the company as a whole has to naturally take risks to differentiate themselves. Risk-reward and all that.

Again, you are confusing contemporary theory on motivation with historical evidence that disputes it. The contemporary theory is focused around the logic of market rationality, which, might, according to recent evidence, be a little flawed?

Which is what I mean by "motivation by fear" as not being wrong in this regard - the margin for error is slimmer when you have competition, which increases the pressure on the employees, but the end result is that consumers benefit. From a customer's perspective, that's a win.

Even though "motivation by fear" is a metaphor, even if it were true, you certainly don't motivate your employees that way. If the company is struggling, you don't motivate your employees with threats. A company has stakeholders and employees, and it's obvious that both sets of people have competing desires in the company, and hence need different messages. If there is competition, the stakeholders will demand increased profits and market share, but only an idiot CEO would use this as the sole motivational tool for his employees.

So, fear is another metaphor? Wow, I feel stupid now. What is it a metaphor for, by the way? Motivation by "worry"? Motivation through "increased pressure and shorter deadlines?"

Could you expand on "competing desires?" Is it only the stakeholders and employees, as grouped entities, that have different desires? Aren't employees actually stakeholders? Aren't we stakeholders? Are there any other groups that might have different desires?

[x42bn6]

To be honest, I don't think mantralux or yourself can escape the argument that you are accusing SI of being either lazy, complacent or incompetent. You have both stated that the problems (that we've all agreed the game has) are easy to solve. If they are, and SI aren't solving them, then, by definition, you are accusing them of not being up to the task.

I don't think I've said it's easy to solve, and mantralux has given a suggestion that sounds easy but of course will not be difficult to implement in practice.

Depending on how "legacy" the codebase is, easy solutions might simply have complex implementations. Markov decision processes, the suggestion I put out, are very easy to actually do but integrating that with the legacy codebase could be difficult. Imagine, for example, if squad building wasn't a module in itself, but a criss-cross of dependencies.

Is it an accusation of ineptness? Probably. Which is why we'd like developers to tell us why it's so hard. The mantralux suggestion of a very simple condition (I'd prefer it if it were a goal, if the AI is goal-oriented) would be to have two players per position - why wouldn't that work?

Of course, they might be really difficult to solve, which suggests something else.

I'd like to hear from developers why it's so hard. You can find algorithms for Markov decision processes quite easily. Does the codebase not use anything similar? If so, why is it hard to adjust?

Can or surely? Different things. Also, in major doubt. There has never been any serious competition, so how can you have any idea what it might do? You are extrapolating a claim from a contested and extremely contemporary theory of motivation and presenting it as "truth".

Championship Manager? Maybe not an equal competitor, but a competitor nevertheless.

When I used "surely" in that sense, I was trying to put out that "it can't make it any worse", i.e. P(game gets worse) = 0 (or is negligible).

[wwfan]

I don't think I've said it's easy to solve, and mantralux has given a suggestion that sounds easy but of course will not be difficult to implement in practice.

Depending on how "legacy" the codebase is, easy solutions might simply have complex implementations. Markov decision processes, the suggestion I put out, are very easy to actually do but integrating that with the legacy codebase could be difficult. Imagine, for example, if squad building wasn't a module in itself, but a criss-cross of dependencies.

Is it an accusation of ineptness? Probably. Which is why we'd like developers to tell us why it's so hard. The mantralux suggestion of a very simple condition (I'd prefer it if it were a goal, if the AI is goal-oriented) would be to have two players per position - why wouldn't that work?

I'd like to hear from developers why it's so hard. You can find algorithms for Markov decision processes quite easily. Does the codebase not use anything similar? If so, why is it hard to adjust?

I'm not a coder. Paul Collyer has told me it is hard. That is enough for me. But at least you are now being honest about your motivations. You think SI are inept in squad management AI coding. You think it the squad building issue should, theoretically, be easy to solve. You assume the legacy code is the reason it isn't. You'd like clarification. Fair enough.

Championship Manager? Maybe not an equal competitor, but a competitor nevertheless.

When I used "surely" in that sense, I was trying to put out that "it can't make it any worse", i.e. P(game gets worse) = 0 (or is negligible).

Ah, but now you've made your bed, you must lie in it. SI are never going to share any coding information with anyone if they feel it will give away a competitive advantage. They can't simultaneously tell their fan base all their technical secrets and keep an advantage over their competitors.

[x42bn6]

Why did these not provide competition given it is not a question of resources or desire? Why did Fifa not just create a new match engine given that they had the resources and desire to be a realistic competitor? They clearly had competition to motivate them in the form of SI/Sega. I don't understand why they failed given what I've learnt today in this thread.

It's possible they didn't have the resources to build the game from bottom-up. Let's not forget FIFA Manager is not a huge cash cow for them.

So, fear is another metaphor? Wow, I feel stupid now. What is it a metaphor for, by the way? Motivation by "worry"? Motivation through "increased pressure and shorter deadlines?"

"Fear" as in pressure from people with financial stakes in the company.

This pressure drives shorter deadlines and increased risks. If you fall behind, your competitor will gleefully accept.

So the full metaphor is "Motivation/pressure (in some form, hopefully positive) to meet shorter deadlines."

Could you expand on "competing desires?" Is it only the stakeholders and employees, as grouped entities, that have different desires? Aren't employees actually stakeholders? Aren't we stakeholders? Are there any other groups that might have different desires?

Employees might be stakeholders depending on whether or not they have a desire to see the company develop. A loyal employee might want to see a company develop because he cares, so will be a stakeholder. Some employees hold shares in the company, which means they have an interest in doing well and growing their investment. Some employees, of course, are in for their pay packet and nothing more.

Are customers stakeholders? In a lot of ways, I'd argue no, unless you hold shares in Sega or are an employee as well. If you really care about a company, then you could be considered a stakeholder as you too would suffer if it suffers.

Stakeholders hold stakes for different reasons and even stakeholders will have different desires. A loyal employee or caring customer wants to see a company do well both financially and reputationally. A shareholder cares more about financial results.

There are others, like the government, unions, regulators and auditors who would fall into the categories of "interested parties", but not all will have stakes in the company. In a lot of ways, some of these wouldn't be too concerned if the company failed, as long as the company wasn't too large, of course.

[x42bn6]

I'm not a coder. Paul Collyer has told me it is hard. That is enough for me. But at least you are now being honest about your motivations. You think SI are inept in squad management AI coding. You think it the squad building issue should, theoretically, be easy to solve. You assume the legacy code is the reason it isn't. You'd like clarification. Fair enough.

I'm not assuming the legacy code is an issue, but it's the first thing that comes to mind. One possible alternative could simply be a lack of unit/integration testing so there's no desire to stomach a refactoring, because you never know what breaks. There's a few other reasons why it's hard.

But in my experience, it's usually legacy code.

Ah, but now you've made your bed, you must lie in it. SI are never going to share any coding information with anyone if they feel it will give away a competitive advantage. They can't simultaneously tell their fan base all their technical secrets and keep an advantage over their competitors.

They don't have to tell us the coding secrets in detail. If they said "we use Markov chains/goal-based AI/neural networks/some common AI construct", congratulations, they've just told the world they've used one of the most useful technical constructs known in this area, which happens to be used by a large number of universities and companies. If they say they use some custom-built AI, then it will be a lot more interesting because I don't know why they'd pick something over things that simply work.

[thom]

For example, tighter deadlines would have to be hit and slippages will be costly.

There are plenty of other factors that determine the quality of the end product, and competition is part of that.

Tighter deadlines often have the opposite effect in terms of quality, actually.
You often hear people in this forum (possibly you included) claiming that FM would be better if it was released after the January transfer window, or even better every other year (the financial repercussions such a decision would have on the company have already been discussed countless times).
I very much doubt tight deadlines are a quality-determining factor.

As far as competition leading to a better product is concerned, we could all come up with examples supporting and contradicting that claim until we're blue in the face.

[marty78]

It's possible they didn't have the resources to build the game from bottom-up. Let's not forget FIFA Manager is not a huge cash cow for them.

So why should SEGA pump money into FM development when EA have little interest? Out of the goodness of their heart? I'm sure if they were happy to not make any money all sorts of things could be achievable and I'd love to hear SI pitch that idea to SEGA. Its as unrealistic as when people suggest missing a release to allow for extra development.

[x42bn6]

Tighter deadlines often have the opposite effect in terms of quality, actually.
You often hear people in this forum (possibly you included) claiming that FM would be better if it was released after the January transfer window, or even better every other year (the financial repercussions such a decision would have on the company have already been discussed countless times).
I very much doubt tight deadlines are a quality-determining factor.

That's what project planning is for. If deadlines are too tight, then it might be that SI need to hire more developers to fill the gap, because not filling the gap (or filling it badly) will result in the competition gaining an advantage.

So why should SEGA pump money into FM development when EA have little interest? Out of the goodness of their heart? I'm sure if they were happy to not make any money all sorts of things could be achievable and I'd love to hear SI pitch that idea to SEGA. Its as unrealistic as when people suggest missing a release to allow for extra development.

One day, I hope I can say it's down to customer pressure, where customers stop trying to hold back their demands for the game to improve and actually focus their energies on asking things of SI.

[thom]

That's what project planning is for. If deadlines are too tight, then it might be that SI need to hire more developers to fill the gap, because not filling the gap (or filling it badly) will result in the competition gaining an advantage.

I guess a company like Valve believe a quality product will always be successful, however long it takes to release it and however hard their competitors try in the meantime.
And they're probably right...

[godsjobsagoodun]

The FM match engine generates 45 minutes of football in approx. 2 seconds at the start of every match.

FPS have 5 people on screen who need to decide whether to duck behind cover or not.

Ackter, are you saying that the result of the 1st half is decided before its played, and not actually as i watch it?

[marty78]

That's what project planning is for. If deadlines are too tight, then it might be that SI need to hire more developers to fill the gap, because not filling the gap (or filling it badly) will result in the competition gaining an advantage.

Deadlines are not the only thing to be met. A budget will also exist so hiring new staff is easier said than done for most companies. Since there is so little competition would increasing the budget be worth the extra sales? Would a better AI win over than many new customers? I don't know, do you?

One day, I hope I can say it's down to customer pressure, where customers stop trying to hold back their demands for the game to improve and actually focus their energies on asking things of SI.

Is this forum not full of such pressure? Often with no foundation as well. See the wishlist threads for people "focusing their energies on asking things of SI" and doesn't this thread itself do this even if rather poorly?

10 characters

[thomit]

@godetc. If you or the AI don't change anything during the first half, it is. If you change anything during the first half, even the tiniest little thing, the rest of the half will be re-calculated. And so on. Repeat for 2.nd half.
This is what I have been lead to believe, anyway. It is the way it has to be, in order to bring you goal/key/extended highlights.

[CaptainPlanet]

A bit like pin the tail on the donkey, we should play put a red dot on the triangle where you think FM is.

[thomit]

Well, we know FM isn't particularly cheap, nor is it particularly fast ... so it must be good then

[godsjobsagoodun]

@godetc. If you or the AI don't change anything during the first half, it is. If you change anything during the first half, even the tiniest little thing, the rest of the half will be re-calculated. And so on. Repeat for 2.nd half.
This is what I have been lead to believe, anyway. It is the way it has to be, in order to bring you goal/key/extended highlights.

Cheers Thomit

[fmobande]

i feel happy, cause i made a thread with a decent amount of posts , anyways ..

[VonBlade]

I'm not defending SI, god knows the game has some serious issues, but can anyone name games that do successfully provide a long term challenge?

There aren't many that need the AI to make long-term strategic decisions, Civ springs to mind as one, and of those I can't think of any that don't cheat to provide a false sense of challenge. Civ and other 4X games do resource cheating, RTS games do both resource/diplomacy/fog-of-war cheating and the ability to micromanage a hundred units at the same time, FPS games do the range/accuracy/health/ammo cheating, and they aren't having to provide a long-term challenge. So although FM is poor, when it comes to keeping things on the level it's probably at the head of the pack. And yes I know the AI does insta-scouting FOW cheating.

[Smac]

It is unnecessary to have to know why competition might spur innovation in order to suggest that it may well do. Unfortunately it seems that many seem to think it is required. It is actually a lack of this (and other) knowledge that is a good reason to suggest competition in the first place.

I don't think mantralux or yourself can escape the argument that you are accusing SI of being either lazy, complacent or incompetent. You have both stated that the problems (that we've all agreed the game has) are easy to solve. If they are, and SI aren't solving them, then, by definition, you are accusing them of not being up to the task.

Perhaps, but I'll take a whack at it...

You are suggesting that SI have all possible ingredients to make the best game possible, which implies that outside entities / events could have no impact on the set of "ingredients". I think that this is the assumption being made, not that SI must be "lazy, complacent or incompetent". No, it is precisely that one cannot anticipate what will come out of working with uncontrolled outside influences that makes it so appealing when confronting what appear to be intractable hurdles. If I cannot solve a problem by myself, I ask a friend for help, etc.. Competition can have influences much like cooperation in that solutions to difficult problems emerge from the process that can not be anticipated by any individual party.

As has been rightly pointed out, competition can also destroy things. There is not guarantee of progress, let alone for any particular entity. Doesn't mean it couldn't inspire SI to new heights though.

I think those that have supported the idea of stronger competition for FM have fallen into the trap of trying to defend the "What would that do?" and "Why would it work?" questions. It isn't necessary (or likely possible) to know. It may well be the case that no amount of competition, large or small, will spur the football manager market to new heights. And it almost certainly would be the case that if strong competition to FM arose that it would be an uncomfortable and risky impingement on SI. Its one of those things where it's "be careful what you wish for" must apply. Still, like any relationship, you can't know it'll be horrible without having it.

[x42bn6]

Deadlines are not the only thing to be met. A budget will also exist so hiring new staff is easier said than done for most companies. Since there is so little competition would increasing the budget be worth the extra sales? Would a better AI win over than many new customers? I don't know, do you?

Like I said, that's what project planning was for, and I used the example of hiring more developers as an example of what could be done. Evaluating the cost benefits will of course fall into this equation.

Is this forum not full of such pressure? Often with no foundation as well. See the wishlist threads for people "focusing their energies on asking things of SI" and doesn't this thread itself do this even if rather poorly?

Is it full of pressure? Not enough, personally. There's a lot of defending of SI, though, although that makes no sense from a customer perspective.

It is unnecessary to have to know why competition might spur innovation in order to suggest that it may well do. Unfortunately it seems that many seem to think it is required. It is actually a lack of this (and other) knowledge that is a good reason to suggest competition in the first place.



Perhaps, but I'll take a whack at it...

You are suggesting that SI have all possible ingredients to make the best game possible, which implies that outside entities / events could have no impact on the set of "ingredients". I think that this is the assumption being made, not that SI must be "lazy, complacent or incompetent". No, it is precisely that one cannot anticipate what will come out of working with uncontrolled outside influences that makes it so appealing when confronting what appear to be intractable hurdles. If I cannot solve a problem by myself, I ask a friend for help, etc.. Competition can have influences much like cooperation in that solutions to difficult problems emerge from the process that can not be anticipated by any individual party.

As has been rightly pointed out, competition can also destroy things. There is not guarantee of progress, let alone for any particular entity. Doesn't mean it couldn't inspire SI to new heights though.

I think those that have supported the idea of stronger competition for FM have fallen into the trap of trying to defend the "What would that do?" and "Why would it work?" questions. It isn't necessary (or likely possible) to know. It may well be the case that no amount of competition, large or small, will spur the football manager market to new heights. And it almost certainly would be the case that if strong competition to FM arose that it would be an uncomfortable and risky impingement on SI. Its one of those things where it's "be careful what you wish for" must apply. Still, like any relationship, you can't know it'll be horrible without having it.

There's reason to believe that competition will not at the very least destroy all competitors. In the past, Championship Manager has been somewhat of a challenger and they didn't end up taking each other down. Then there's FIFA and PES, where the competition has led to both sides trading blows for years.

[fmobande]

Like I said, that's what project planning was for, and I used the example of hiring more developers as an example of what could be done. Evaluating the cost benefits will of course fall into this equation.



Is it full of pressure? Not enough, personally. There's a lot of defending of SI, though, although that makes no sense from a customer perspective.



There's reason to believe that competition will not at the very least destroy all competitors. In the past, Championship Manager has been somewhat of a challenger and they didn't end up taking each other down. Then there's FIFA and PES, where the competition has led to both sides trading blows for years.

why is people talking about competition like its a big factor in this thread and FM, this thread its mainly addressing about how this game fails to stay consistent in the long term seasons, because of poor A.I. the user from IGN only pointed out the competition statement just to add value to his complaint

[edgar555]

why is people talking about competition like its a big factor in this thread and FM, this thread its mainly addressing about how this game fails to stay consistent in the long term seasons, because of poor A.I. the user from IGN only pointed out the competition statement just to add value to his complaint

This thread has come a long way since you posted the OP (and have subsequently added nothing to). It's gone down the route it has because a number of people believe that the lack of competition is one of the reasons why the AI is poor. If you'd actually taken the time to read your thread you wouldn't need to ask the question. Well done for starting what has been for the most part a stimulating and interesting conversation, but all you did was post a user review and then leave everyone to it, so don't congratulate yourself too much.

[marty78]

Is it full of pressure? Not enough, personally. There's a lot of defending of SI, though, although that makes no sense from a customer perspective.

The first page of GD has several threads either suggesting how to improve things or moaning that the game is broken in some way and there is also the bugs forums. There is a lot of defending SI though, which I admit gets a bit silly in some cases. I completely agree that SI should strive to improve every area of the game but I have no reason to believe they are not already doing so. I have only played the last four iterations but I believe they have improved each time and hopefully that trend continues.

I believe SI has no competition because they deliver an excellent product rather than that they make a product below standard because they have no competition. I also believe that every area of the game could be improved and hope that they will.

[wally13]

This thread has come a long way since you posted the OP (and have subsequently added nothing to). It's gone down the route it has because a number of people believe that the lack of competition is one of the reasons why the AI is poor. If you'd actually taken the time to read your thread you wouldn't need to ask the question. Well done for starting what has been for the most part a stimulating and interesting conversation, but all you did was post a user review and then leave everyone to it, so don't congratulate yourself too much.

Was always going pear shape in here once x42bn6 got his 2 pence worth in

[fmobande]

This thread has come a long way since you posted the OP (and have subsequently added nothing to). It's gone down the route it has because a number of people believe that the lack of competition is one of the reasons why the AI is poor. If you'd actually taken the time to read your thread you wouldn't need to ask the question. Well done for starting what has been for the most part a stimulating and interesting conversation, but all you did was post a user review and then leave everyone to it, so don't congratulate yourself too much.

lool you dont need to get annoyed just because i asked rhetorical question , i wasnt following the thread because unlike some people i havent really been online since i posted that thread , i didnt think there would be so many posts, but well done for you to be part of this interesting conversation , im sure you can continue posting all day long with your 1000+ posts while some other people actually do something with their time instead of spending it with by having a discussion on a games forum

[McMaster#28]

There are so many wrong things I this thread I don't know where to start.
I'll admit I'm not the biggest fan of FM now-a-days, but it has almost nothing to do with the AI.

And the idea that the match AI is one the reason FM is "bad", is laughable. Paul is a genius in that department, end of.


Also there seems to be a lot of confusion between the Match AI and the AI in other modules.

[ham_aka_stam]

lool you dont need to get annoyed just because i asked rhetorical question , i wasnt following the thread because unlike some people i havent really been online since i posted that thread , i didnt think there would be so many posts, but well done for you to be part of this interesting conversation , im sure you can continue posting all day long with your 1000+ posts while some other people actually do something with their time instead of spending it with by having a discussion on a games forum

If you want to pick on people for their post count, you're really picking on the wrong person.

Now discuss the topic at hand or don't say anything at all.

[fmobande]

If you want to pick on people for their post count, you're really picking on the wrong person.

Now discuss the topic at hand or don't say anything at all.

he had a go at me just because i joined the conversation , i was discussing the topic at hand till he made a unnecessary comment about what effort i put on this thread, he got annoyed for no reason, but sorry if i pointed out the number of posts he said , ill leave it and stop before this thread goes off topic.

[Svenc]

Why did these not provide competition given it is not a question of resources or desire? Why did Fifa not just create a new match engine given that they had the resources and desire to be a realistic competitor? They clearly had competition to motivate them in the form of SI/Sega. I don't understand why they failed given what I've learnt today in this thread.

The team doesn't create a single coherent "engine" as such - there are multiple ways FIFA Manager calculates a match, one of them being an action packed 2x5 minutes sim based on an oldish FIFA Soccer Engine, but there are also completely different calculations for other modes of play. None of them are very compelling to say the least - currently you cannot even influence the basic style of play as all your team would ever do is playing a direct attacking game. This and similar quirks have been a trademark with this developer and designer for more than a decade. And the most puzzling of all is a German press that collectively insists on tactical and management depth that just isn't there. Taking a quick look at Metacritic on whatever iteration will tell you that pretty much every single glowing review hails from German speaking countries year in year out.

If EA really want to tackle the international markets one day, I can see them approaching Sega or SI rather than Bright Future in Cologne really catching up given their track record. Big IF. The truth is that whilst FIFA Manager may not be big news outside of its cozy German comfort zone, it is selling about 130,000+ copies* each in the German speaking territories year in year out with no "real" competition whatsoever interfering. Yes, you can buy Football Manager, but it is not officially released nor advertised nor anything. Certainly a rather comfortable situation to have around, given the arguably rather limited resources being allocated to FIFA Manager - the expensive exclusive licenses for the Bundesliga help to keep FIFA Manager afloat in Germany, but their main use lies in a much bigger game, and that much bigger game is FIFA Soccer.


Anyway, speaking about competition or lack thereof - in Germany things are evidently much worse off. You can see how a lack of competition skews people's expectations, in my opinion as evident in the often puzzling claims made in articles being published all around. It is well documented that many an editor doesn't even know anything else but FIFA anymore - nor do fans of manager games.

* number taken from an article dating a few years and iterations back.

[wwfan]

You are suggesting that SI have all possible ingredients to make the best game possible, which implies that outside entities / events could have no impact on the set of "ingredients". I think that this is the assumption being made, not that SI must be "lazy, complacent or incompetent". No, it is precisely that one cannot anticipate what will come out of working with uncontrolled outside influences that makes it so appealing when confronting what appear to be intractable hurdles. If I cannot solve a problem by myself, I ask a friend for help, etc.. Competition can have influences much like cooperation in that solutions to difficult problems emerge from the process that can not be anticipated by any individual party.

From my own personal experience, SI do ask outside sources for help when they are confronted with intractable problems. But that's by the by. I'm not trying to suggest that they are, in any way, perfect. I am confronting the condemnation made my the two most vocal critics in this thread and asking them to clarify. You've given them a possible out, but I think they've already accumulated too much rope. That you had to provide it speaks volumes. I'm afraid you are being far too creditable.

As has been rightly pointed out, competition can also destroy things. There is not guarantee of progress, let alone for any particular entity. Doesn't mean it couldn't inspire SI to new heights though.

I think those that have supported the idea of stronger competition for FM have fallen into the trap of trying to defend the "What would that do?" and "Why would it work?" questions. It isn't necessary (or likely possible) to know. It may well be the case that no amount of competition, large or small, will spur the football manager market to new heights. And it almost certainly would be the case that if strong competition to FM arose that it would be an uncomfortable and risky impingement on SI. Its one of those things where it's "be careful what you wish for" must apply. Still, like any relationship, you can't know it'll be horrible without having it.

Well said.

[wwfan]

I'm not assuming the legacy code is an issue, but it's the first thing that comes to mind. One possible alternative could simply be a lack of unit/integration testing so there's no desire to stomach a refactoring, because you never know what breaks. There's a few other reasons why it's hard.

But in my experience, it's usually legacy code.

They don't have to tell us the coding secrets in detail. If they said "we use Markov chains/goal-based AI/neural networks/some common AI construct", congratulations, they've just told the world they've used one of the most useful technical constructs known in this area, which happens to be used by a large number of universities and companies. If they say they use some custom-built AI, then it will be a lot more interesting because I don't know why they'd pick something over things that simply work.

You see, at first glance, here I have to believe you, as I don't have the knowledge to argue the point. My only sources of reference are Paul Collyer, who has told the forums how difficult it is, and my own friends, who are successful software designers/coders, who also say it would be horribly difficult.

However, I can extrapolate. You are suggesting that it should be theoretically relatively simple to program a robust squad building / transfer AI that extends over 20 odd seasons. You've offered a coupe of theoretical solutions. However, you have forgotten practice. For this to be programmed, it needs to be tested. For it to be tested, you need to have constant simulations of 20 seasons or more, some being played by users, all being evaluated by testers, that find and feed relevant data back to the coders in order for them to correct code. Not looking so easy now, is it?

You are also failing to take into account balance problems. If the AI is slightly too aggressive in perfecting its squads, the transfer market gets flooded with money, destabilising the gameworld economy. Not aggressive enough, then the transfer market is totally stagnant, which also destabilises the gameworld economy. Get the balance wrong and the game because totally unplayable long-term. Consequently, major risk taking in this area should be avoided.

And even with the above, why should they let their competitors know which process they are using? Even if it is an industry standard? If there are more than one, you aren't going to tell potential competitors which one you know works best for football management simulations.

Employees might be stakeholders depending on whether or not they have a desire to see the company develop. A loyal employee might want to see a company develop because he cares, so will be a stakeholder. Some employees hold shares in the company, which means they have an interest in doing well and growing their investment. Some employees, of course, are in for their pay packet and nothing more.

Are customers stakeholders? In a lot of ways, I'd argue no, unless you hold shares in Sega or are an employee as well. If you really care about a company, then you could be considered a stakeholder as you too would suffer if it suffers.

Stakeholders hold stakes for different reasons and even stakeholders will have different desires. A loyal employee or caring customer wants to see a company do well both financially and reputationally. A shareholder cares more about financial results.

There are others, like the government, unions, regulators and auditors who would fall into the categories of "interested parties", but not all will have stakes in the company. In a lot of ways, some of these wouldn't be too concerned if the company failed, as long as the company wasn't too large, of course.

Now this sums up my real issue with you. Although I might accept your knowledge of coding as being robust and thus your observations legitimate, the above is massively problematic. I don't have to rely on others to tell me this, because it is my area of expertise. Your definition of stakeholders, although stabbing in the right direction, is pretty poor. However, your understanding of organizational motivation, hinted at again here and detailed in your other posts, is non-existent. Unfortunately for you, this undermines any faith I might have in anything else you are saying.

[Smac]

From my own personal experience, SI do ask outside sources for help when they are confronted with intractable problems. But that's by the by.

That is the kind of thing I'm trying to get at in defense of competition. Competition, or any sort of only partially voluntary interaction with outside forces can sometimes move a process beyond those areas we're amenable to addressing. Problems we seek answers to on our own, be they solved internally or through consultation, are potentially vastly different than problems we find ourselves having to solve because the actions of others have forced our hand in some way. This is why I wholeheartedly believe SI do everything they can to improve the product already and have little trouble accepting the above comments that "We are our own biggest critics" yet still think (hope? fantasize?) things could be improved further, beyond where SI are willing to go on their own initiative! The ideas are not contradictory.

I'm not trying to suggest that they are, in any way, perfect. I am confronting the condemnation made my the two most vocal critics in this thread and asking them to clarify. You've given them a possible out, but I think they've already accumulated too much rope. That you had to provide it speaks volumes. I'm afraid you are being far too creditable.

Well, frankly I care more that the idea of competition doesn't get thrown under the bus than about whether or not you're asking them to get rope or step on the nice wooden platform we have here A poor line of reasoning proves and disproves nothing. "An absence of evidence...." and so forth.

[wwfan]

That is the kind of thing I'm trying to get at in defense of competition. Competition, or any sort of only partially voluntary interaction with outside forces can sometimes move a process beyond those areas we're amenable to addressing. Problems we seek answers to on our own, be they solved internally or through consultation, are potentially vastly different than problems we find ourselves having to solve because the actions of others have forced our hand in some way. This is why I wholeheartedly believe SI do everything they can to improve the product already and have little trouble accepting the above comments that "We are our own biggest critics" yet still think (hope? fantasize?) things could be improved further, beyond where SI are willing to go on their own initiative! The ideas are not contradictory.

Well, frankly I care more that the idea of competition doesn't get thrown under the bus than about whether or not you're asking them to get rope or step on the nice wooden platform we have here A poor line of reasoning proves and disproves nothing. "An absence of evidence...." and so forth.

I've no gripe against competition. Some viable competition might or might not improve Football Manager. My issue with the claims in this thread is that 'as there is no competition, SI aren't pushing themselves'. This has been defended via two arguments.

Firstly, there is the claim that the problems of long-term squad management and transfer AI are easy to solve. Those making this claim obviously have some grasp of theory, and perhaps some practical experience. However, from what I can tell, they haven't worked on any AI project of this scale. They then get confronted by others with industry knowledge who argue that the practice of AI coding is far more difficult than they are positing. We then get into the intractable knot common to theory v practice debates. I can't debate this, although I can choose, based on previous experience, whom to trust.

The group arguing that SI should be better at coding then, for me, completely undermine their legitimacy by making strong claims about competition and motivation that are, at best, hugely contestable and are, at worse, total nonsense. For example, I briefly mentioned Csikszentmihalyi earlier. Even adding his theories into the mix completely undermines the strength of the competition=motivation argument, let alone looking at the complexities of post-Schein organizational psychology and related critical debates. The over-simplification of their claims on competition and motivation suggests to me that they are equally over-simplifying their discussions on coding.

If I add the above to my own subjective decision on whom I should trust, then I find the critique faintly ridiculous. That's not to say I don't think the competition/motivation argument is completely invalid. Merely that it being presented as gospel is hugely problematic, which undermines any other claim made.

[CaptainPlanet]

Wasn't the development of the FM 3D engine brought forward as a direct result of CM doing it?

I'm not using that as an example of motivating or complacency, more that competition influences development decision making.

[wwfan]

Wasn't the development of the FM 3D engine brought forward as a direct result of CM doing it?

I'm not using that as an example of motivating or complacency, more that competition influences development decision making.

Not according to SI. I have strong reason to believe that.

[keeper tom]

Fair points by Neil Brock here. It would be great if the AI was more clever when it comes to transfers because the likes of Man City spend alot of their money on players that cant even make the team. Also if a team is short on CB's the AI should notice this and try to buy a player in this position.

[CaptainPlanet]

Not according to SI. I have strong reason to believe that.

Fair enough.

I'm kind of hoping that if there was a competitive product, that some of the decisions made about what to spend time on developing would have been different and I would still be interested in buying new versions of the game instead of clinging on to the last version that I enjoyed.

[wwfan]

Fair enough.

I'm kind of hoping that if there was a competitive product, that some of the decisions made about what to spend time on developing would have been different and I would still be interested in buying new versions of the game instead of clinging on to the last version that I enjoyed.

Competition might generate innovation. It might not. I'd argue that you'd struggle to find any serious research on managing innovation and creativity that highlights the level of competition as being a key motivational factor. It is a factor, of course, but it should never be the core focus of a creative industry. Indeed, in the work I've done with creatives, explicit competing for critical acclaim actually harms the performance of the company, as it sucks up resources on meeting minor demands and current fashions at the expense of genuine innovation.

[CaptainPlanet]

Competition might generate innovation. It might not. I'd argue that you'd struggle to find any serious research on managing innovation and creativity that highlights the level of competition as being a key motivational factor. It is a factor, of course, but it should never be the core focus of a creative industry. Indeed, in the work I've done with creatives, explicit competing for critical acclaim actually harms the performance of the company, as it sucks up resources on meeting minor demands and current fashions at the expense of genuine innovation.

Most of the time customers don't want innovation, they want their basic expectations met.

Depending on how mature the market is, competition either drives innovation or it drives focus on doing the fundamentals as well as possible.

The companies with the best customer experience ratings (and subsequently customer loyalty and brand value) are those that meet their customers' demands, not those that strive to exceed them.

[wwfan]

Most of the time customers don't want innovation, they want their basic expectations met.

Depending on how mature the market is, competition either drives innovation or it drives focus on doing the fundamentals as well as possible.

The companies with the best customer experience ratings (and subsequently customer loyalty and brand value) are those that meet their customers' demands, not those that strive to exceed them.

Although you have some good points, this is over-simplification. Take Apple under Steve Jobs, for example. He paid no attention at all to customer demands. He also focused on innovation and design when everybody else in the industry thought cheap functionality was what everybody wanted. His argument was that customers didn't know what they wanted until you showed it to them.

Ultimately, in a creative industry, you need to trust in your vision. I believe FML failed because customer expectations and demands diluted that vision, whereas FM achieves because they, in large, haven't. I don't think external competition has ever had any major implications for that vision either, partly because SI make the game they want to play and won't compromise that based on what others may or may not be doing, and partly because no competitor has ever produced a serious rival.

Where I believe SI excel is they know when to put the breaks on that vision in respect to commercial realities. To survive, they have to release games to relatively tight deadlines. They don't have the luxury of spending hours and hours getting everything just right or to alter design for critical acclaim alone. They have to be pragmatic.

In this thread, they are being condemned for not getting something perfect in the belief that lack of competition, rather than necessary pragmatism and the complexity of the task is the core issue. If lack of competition is core, then SI are either lazy/complacent or incompetent. They either know how to fix it but won't, aren't aware the problem exists (pretty inconceivable) or aren't capable of fixing it, even though it is, apparently, an easy fix. If competition has nothing to do with it, then they are either incompetent or the problem is a very complex one with no easy solution.

I think it is a big stretch to condemn the industry leader as being incompetent or inept, so have to conclude the problem is extremely complex. I don't think external competition has very much to do with anything, although I won't dismiss it entirely.

[CaptainPlanet]

I think a lot of the decisions made by Sega seem to support the theory that the goal for SI is strongly focused on sales. I think that would be more obvious to see if there was a competitor as they would be battling with a rival rather than ignoring the competition and just following their own path (ala Apple). As a result, I think the existence of competition would have an impact on FM's development.

I could be wrong, and obviously there are various degrees of balance between the two, but usually the focus is either on sales or on art.

[wwfan]

I think a lot of the decisions made by Sega seem to support the theory that the goal for SI is strongly focused on sales. I think that would be more obvious to see if there was a competitor as they would be battling with a rival rather than ignoring the competition and just following their own path (ala Apple). As a result, I think the existence of competition would have an impact on FM's development.

I could be wrong, and obviously there are various degrees of balance between the two, but usually the focus is either on sales or on art.

I think you'll find that while SEGA wants sales to increase, the trust placed in SI's processes is absolute and the interference, bar SI having to meet contractual demands, minimal to non-existent. Given the track record of SI, SEGA would be incredibly foolish to interfere with the creative process.

I'm not discounting that managers in the creative industry don't have a deep reservoir of foolishness to draw upon. However, I don't think there is any evidence of SEGA ever forcing SI down a path they aren't happy taking. SI employees, from director, creatives, testers and coders, have consistently stated that the integrity of the game is their core motivation. Either they are all lying, or you are taking two and two and making five?

I also doubt they'd throw all their proven processes in the garbage just because a rival pops up.

[CaptainPlanet]

As a consumer who disagrees with a lot of the development decisions made (specifically around new feature choices), in addition to seeing Sega's influence over things like DRM, I find it hard to believe that game integrity and ultimately quality hasn't been sacrificed to a degree to appeal to the masses to drive up sales.

To be more specific, I think the game was going through cycles of becoming more and more complex, then suddenly appeared to start being simplified, with feature development being less around the game mechanics and more around cosmetics. I can't see evidence to support that priorities haven't changed.

I accept that a lot of this is my own personal views, and I probably have a bit too much emotion around my disappointment with the development route the game has taken in recent years. So I could well be wide of the mark. But nonetheless it's how I feel.

[wwfan]

As a consumer who disagrees with a lot of the development decisions made (specifically around new feature choices), in addition to seeing Sega's influence over things like DRM, I find it hard to believe that game integrity and ultimately quality hasn't been sacrificed to a degree to appeal to the masses to drive up sales.

Conflation. DRM has nothing to do with development. It is a production decision.

To be more specific, I think the game was going through cycles of becoming more and more complex, then suddenly appeared to start being simplified, with feature development being less around the game mechanics and more around cosmetics. I can't see evidence to support that priorities haven't changed.

Examples, please. They need to specify the dumbing down of the game, not the clarification of language or simplification of functionality. They also need to be indicative of a step away from the quest for a realistic simulation into a quest for an arcade mode.

[eastley]

This thread is useless.

Users defend this game, no matter what.

I agree with what the guy said.

I play normal tactics, set the game up how i want etc, using tools within the game. I do not use corner tactic, I only ever use regens.

I dominate after 1-2 seasons with average teams.

[CaptainPlanet]

Conflation. DRM has nothing to do with development. It is a production decision.

I didn't say it did, I said in addition to. The combination of the two (the development choices and Sega's influence over something that SI was opposed to that was purely to try and drive up revenue)

Examples, please. They need to specify the dumbing down of the game, not the clarification of language or simplification of functionality.

Simplification of functionality is simplifying no? If previously you had to press 20 keys to do something that now took only 2 keys that would be development time spent making something more simple?

They also need to be indicative of a step away from the quest for a realistic simulation into a quest for an arcade mode.

"quest for a realistic simulation" is completely different to having integrity and a quality product. You could give each player a bladder bar, have wives and girlfriends to deal with, drive the team bus and justify them all as features under the "quest for realism" umbrella. That doesn't mean you end up with a quality product.

[wwfan]

This thread is useless.

Users defend this game, no matter what.

I agree with what the guy said.

I play normal tactics, set the game up how i want etc, using tools within the game. I do not use corner tactic, I only ever use regens.

I dominate after 1-2 seasons with average teams.

See below:

- An awesome tactic built by user on this forum, then altered by me after some 40 seasons of playing that one tactic (I'm crap at tactics, needed a base to work off)

@ CaptainPlanet:

That's a "no, I don't have any examples" then?

[CaptainPlanet]

I have plenty that fit my opinion, if I didn't, I wouldn't have that opinion. But debating the validity of each one of them with you would take far more effort than I'm willing to put in at 4 in the morning

[wwfan]

I have plenty that fit my opinion, if I didn't, I wouldn't have that opinion. But debating the validity of each one of them with you would take far more effort than I'm willing to put in at 4 in the morning

Perhaps when you are well rested, then

[CaptainPlanet]

Perhaps

[HoroHoro]

SI have almost no competition. Why? Let's be frank, the game you are playing is still a business, money making.

The effort and trouble to make a complex game such as FM? Very high. The income compared to an easy to make game like FPS? FPS makes more money easier than a game like FM. Why would any business (if their interest is not for the love of the game) make such game? A game that takes more effort, hassle to make and yet gets them less money?

We should salute the Football Manager developers for sticking to it rather than switch attention to other kind of games that is so much easier to make, yet gets them more money than FM.

[MSCCG]

I always feel that if games paid more attention to psychology, the AI would be more realistic. The language being used always makes me think that the programmers are trying to simulate something they don't quite understand, and that's why gaming AI feels a bit too "robot-y" and has seemed to stagnate a bit and just lacking that final push into the territory of the Turing test.

[eastley]

And I took users advice and built my own, similar results, Just won every comp with Southampton. Only lost 1 game in 2 years.

[Ackter]

Could you post a picture of your game status screen? Would be interesting to see your time spent on it.

[Ackter]

Simplification of functionality is simplifying no? If previously you had to press 20 keys to do something that now took only 2 keys that would be development time spent making something more simple?

SI have been running usability studies for years. With a game as complex as FM they'd be daft to not try and make it as easy for new customers to get into as possible. It all goes towards making the game as good as it can be, I don't get why this is used as an example of SI being forced to dumb-down the game. Especially as the game has got more complex each iteration.

[avine]

Hopefully everyone to a degree. Surely we all want to look back on FM12 in 20 years time and think "man, this is pathetic compared to FM32"? I don't particularly care whether that game is made by SI or someone else, but the chances are the best management games will be made by SI for the foreseeable future.

Imagine FM12-FM32 hours used multiplied per release lol-i promise to my self i would not think this way. Seriously now, even economists lately tend to abandon the thought that monopolies or oligopolies are mal. As customers some people(i would think ehm...80%) are looking to stay clear of dominance from companies that are not public etc and also want confirmation for their contribution, their effect in a market. So, about FM, the one factor that is really safer than anything alse is 'value' of a product. People see 'value' in FM and they buy it every year. I think it is fair when someone thinks based on 'value' and make his decisions as a customer.

[x42bn6]

You see, at first glance, here I have to believe you, as I don't have the knowledge to argue the point. My only sources of reference are Paul Collyer, who has told the forums how difficult it is, and my own friends, who are successful software designers/coders, who also say it would be horribly difficult.

However, I can extrapolate. You are suggesting that it should be theoretically relatively simple to program a robust squad building / transfer AI that extends over 20 odd seasons. You've offered a coupe of theoretical solutions. However, you have forgotten practice. For this to be programmed, it needs to be tested. For it to be tested, you need to have constant simulations of 20 seasons or more, some being played by users, all being evaluated by testers, that find and feed relevant data back to the coders in order for them to correct code. Not looking so easy now, is it?

Soak testing is already performed today. I'd argue it's no more difficult.

If you pick Markov decision processes, you can even derive analytical results out of it to reduce the amount of testing.

If the squad building is a separate module, it can even be balanced separately, to ensure the previous and new solutions have the same probability distributions.

The hope is that the code has been written in that way, of course, because if, say, the new squad building "alpha" AI was defective, it would fail while testing only that bit of code.

You are also failing to take into account balance problems. If the AI is slightly too aggressive in perfecting its squads, the transfer market gets flooded with money, destabilising the gameworld economy. Not aggressive enough, then the transfer market is totally stagnant, which also destabilises the gameworld economy. Get the balance wrong and the game because totally unplayable long-term. Consequently, major risk taking in this area should be avoided.

No. Major risk-taking simply needs more care.

All you need to do is make sure that given the input probability distribution (which can be easily achieved via AOP, i.e. AspectC++), this squad building module produces a similar probability distribution. If the software is sufficiently tested (which it will be, if SI have very good developers), then this shouldn't be incredibly difficult and reduces the risks substantially. Soak/UAT testing is then relatively agnostic to whether it's a surgical upgrade of a specific module or a one-line bugfix.

[QUOTE=wwfan;7973819]And even with the above, why should they let their competitors know which process they are using? Even if it is an industry standard? If there are more than one, you aren't going to tell potential competitors which one you know works best for football management simulations.

Because it would only be a tiny bit of what is important about the AI. "Goal-based AI" is vague enough, for example.

Even details like the game generating 45 minutes of the match in 2 seconds tells us a lot of things - i.e. that there is a timeline generated by the code and it is regenerated whenever a significant event happens. It also confirms that there is a random seed involved when the match starts because it helps do the replays and you can jump forwards and backwards between events. It would also explain why it's difficult to generate cross-game results (i.e. relegation battle => goal scored in another key match) because it's much more difficult to process matches in parallel, reducing the level of threading (as if it were one match per thread, there would be cross-thread dependencies). This hasn't killed SI. Something vague like "goal-based AI" won't either.

Now this sums up my real issue with you. Although I might accept your knowledge of coding as being robust and thus your observations legitimate, the above is massively problematic. I don't have to rely on others to tell me this, because it is my area of expertise. Your definition of stakeholders, although stabbing in the right direction, is pretty poor. However, your understanding of organizational motivation, hinted at again here and detailed in your other posts, is non-existent. Unfortunately for you, this undermines any faith I might have in anything else you are saying.

I'm happy to defer all of this motivational stuff to you. All I am trying to get across is that the concerns of a monopoly are different to the concerns of a company with an equal competitor, and one of those is losing business to the other. How you manifest that concern to your employees can be done in any sensible method, but if anything, it reduces the level of complacency because nobody can afford to sit on their laurels.

[wwfan]

Soak testing is already performed today. I'd argue it's no more difficult.

If you pick Markov decision processes, you can even derive analytical results out of it to reduce the amount of testing.

If the squad building is a separate module, it can even be balanced separately, to ensure the previous and new solutions have the same probability distributions.

The hope is that the code has been written in that way, of course, because if, say, the new squad building "alpha" AI was defective, it would fail while testing only that bit of code.

FM is certainly modular.

Out of interest, have you had any practical experience in AI coding at any complex level or is all your knowledge theoretical?

I'm certainly interested in knowing quite how useful it would be to run an extensive soak of a 20 year simulation across 20 odd leagues. I can understand that this would pick up trends, but would it pick up the reasons for when the trend started to veer way from an acceptable parameter. Would it be possible to fine tune the testing around the soak, or would you still need individual testers casting individual eyes over specific transfers? Basically, will a soak/UAT be enough in itself?

[Dave C]

I don't think I've said it's easy to solve, and mantralux has given a suggestion that sounds easy but of course will not be difficult to implement in practice.

Mantralux's suggestion is entirely unworkable and would move us from one undesirable state of affairs, to another.
The two players per position isn't how real-life sides often work, where smaller teams with smaller squads provide cover through 'utility' players that cover multiple positions. So you actually risk financially unsustainable models.

Plus, it's a bodge-job solution to a tiny fraction of the squad management issue. Which makes it pretty pointless.

Depending on how "legacy" the codebase is, easy solutions might simply have complex implementations. Markov decision processes, the suggestion I put out, are very easy to actually do but integrating that with the legacy codebase could be difficult. Imagine, for example, if squad building wasn't a module in itself, but a criss-cross of dependencies.

I'd like to hear from developers why it's so hard. You can find algorithms for Markov decision processes quite easily. Does the codebase not use anything similar? If so, why is it hard to adjust?

Firstly, it's bound to be a criss-cross of dependencies because it involves a squad module, a tactics/ME module, and a finance module. The finance/contracting module is one of the big underlying problems for squad building AI in FM.

Secondly, you've talked about solving this with Markov Chains quite a bit, and said it's an easy solution. However, while that sounds brilliant and well-informed, you have not actually suggested how/why that's a simple solution. So could you elaborate on how Markov Chains would solve the problem? I'm far from convinced.

I'm happy to defer all of this motivational stuff to you. All I am trying to get across is that the concerns of a monopoly are different to the concerns of a company with an equal competitor, and one of those is losing business to the other. How you manifest that concern to your employees can be done in any sensible method, but if anything, it reduces the level of complacency because nobody can afford to sit on their laurels.

Actually, competition breeds a sense of having to play safe, because a broken product will cost you more than a stable one. When you have less economic pressure on you, you can be more creative without the same level of risk attached. You've decided what you think makes theoretical sense without really thinking it through. Intense competition makes for an unimaginative workforce. If you think "Wow, this has to be right or I'll get fired", you focus on not doing anything wrong, not on being innovative.

[wwfan]

@ x42bn6: In conversing with DaveC, you'll be talking to somebody who is much more knowledgeable about squad building and the financial model than I am.

Convince him, and you'll have convinced me.

[pigfacemonkeyman]

Mantralux's suggestion is entirely unworkable and would move us from one undesirable state of affairs, to another.
The two players per position isn't how real-life sides often work, where smaller teams with smaller squads provide cover through 'utility' players that cover multiple positions. So you actually risk financially unsustainable models.

Plus, it's a bodge-job solution to a tiny fraction of the squad management issue. Which makes it pretty pointless.




Firstly, it's bound to be a criss-cross of dependencies because it involves a squad module, a tactics/ME module, and a finance module. The finance/contracting module is one of the big underlying problems for squad building AI in FM.

Secondly, you've talked about solving this with Markov Chains quite a bit, and said it's an easy solution. However, while that sounds brilliant and well-informed, you have not actually suggested how/why that's a simple solution. So could you elaborate on how Markov Chains would solve the problem? I'm far from convinced.




Actually, competition breeds a sense of having to play safe, because a broken product will cost you more than a stable one. When you have less economic pressure on you, you can be more creative without the same level of risk attached. You've decided what you think makes theoretical sense without really thinking it through. Intense competition makes for an unimaginative workforce. If you think "Wow, this has to be right or I'll get fired", you focus on not doing anything wrong, not on being innovative.

Yes possibly, but isn't this what some people are looking for from SI - forget about creating new bells & whistles and concentrate on improving the basics.

[milnerpoint]

Yes possibly, but isn't this what some people are looking for from SI - forget about creating new bells & whistles and concentrate on improving the basics.

What if to improve the basics from what they are now they need to be innovative and take a bit of a risk?

[pigfacemonkeyman]

What if to improve the basics from what they are now they need to be innovative and take a bit of a risk?

Yes possibly, but isn't this what some people are looking for from SI - forget about creating new bells & whistles and concentrate on improving the basics.

They probably would, maybe you missed this bit ^

[Ackter]

The game can't be seen to be standing still, and no new features will make it seem to a lot of people that it is. It has to be a mixture between adding new features and refining old ones.

[milnerpoint]

They probably would, maybe you missed this bit ^

You highlighted this bit and gave a response based on that,

If you think "Wow, this has to be right or I'll get fired", you focus on not doing anything wrong, not on being innovative.

Yes possibly, but isn't this what some people are looking for from SI - forget about creating new bells & whistles and concentrate on improving the basics.

My point is, being innovative does not always lead to new features or as you say "bells and whistles", being innovative could be exactly what the AI needs to improve, putting someone under huge pressure to get something right may infact lead to the basics not improving, because improving said basics could mean taking a risk.

[Sikker]

Sorry guys, but isn't this forum about Football Manager not management theory?
And I must emphasis theory as I've read several of these posts. No doubt you've read a book or two on the subject, but I doubt you know the culture and structure of SI to actually use your theory on a viable analysis of SI.
Ooops, guess I fell in the theory discussion pot myself