retired_FaultyFate

Well, not exactly. For a moment, forget how exactly AI scores. Let me make an example how it might work. Suppose I score most of my goals liike this: cross to TM, he either flicks ball to fast stricker or take a header => goal. AI realises this pattern and forces my TM play deeper + controls the space where the fast ST usually drops. If I don't change anything, headers are not threat anymore (TM too far from goal), and fast ST has less chances => my goals/shots ratio will drop. Of course, it's just a simplified example, but it's pretty realistic.

On the other end, say I always play my star defender unless he is injured. AI knows he is mentally tired (I don't - feedback is bad, though I could guess he might be tired after 40+ games). AI starts to focus playing through his zone => he makes mistakes (own goals, stupid pass - whatever). Etc.


I guess there could be a way to justify my idea through stats analysis. And that brings me to back to my thought that we either need to do a lot of preliminary manager's job, or get better feedback from staff. The latter is impossible in the current version, so one option left.

Now, I am far from saying that we can prevent all those "anomalies" or even explain them. All I want to say that some of those "anomalies" could be just messages like "hey, your DCs have played 35 out 38 games and they simply sick of football, so opposition will explore it".

retired_rayrhi

Kol, I think that you may be giving the engine a little too much credit. It is not clear that the AI can expose your tactic by 'learning' to attack different 'zones' or cut out your familiar method of scoring and the like. If this is the case then you could argue that every team knows what Thierry Henry is going to do during a match (pull wide to the left, cut inside etc.) but stopping him is a completely different kettle of fish. I don't think that the AI should be able to 'nullify' your tactics because after all, most teams play the same style and have the same threats each game. If the AI does work like this then I will hold my hands up in admiration and stand corrected. I think that this is a little too complex though, good sir.

retired_pt22

It seems to me that some folks ( I don't want to accuse anybody personally ) think in the following way:

"I have a tactic that seems logical, that has proven to work. I also have better player, therefor I must win."

This is not the way RL football works. I remember few years ago, when sunderland was struggling to stay in premiership (and got kicked out eventually). They played against Arsenal in the FA cup (I'm not sure if that was the season Arsenal won the league with out a game lost). Anyways for that game Arsenal was in the great form, and Sunderland in terrible (they were pretty convincingly holding that last place in the league, determined not to give it away). The game ended with Sunderland winning 1:0. And my friend that had his money sure on Arsenal, lost the bet. That is football, an unpredictable game of skill and luck.

Another great example would be the CL Finals in the 98 (or was it 99?), that were won by Man Utd with two goals in extra time. Hell, United's entire road to glory was probably the luckiest one (although Liverpool is very close)

However I must agree that Match Engine isn't perfect. Why do players never shield the ball, instead of just hoping it will cross the line before the opposing player catches it. Why is my defender not running after an attacker. Why is my faster player loosing a race against a slower one, when he was clearly closer to the ball.

retired_liveforward

No one would be complaining about one off things that happen IRL if it happened in the game, the problem is the one off's happen way too much in FM.

retired_pt22

Well, about that we'll have to agree to disagree.

Too often I hear from guys that bet how they lost their money on sure things.

Then again, I haven't played yet FM 2007 with a top team. I'm playing with NK Zagreb in Croatian league and I'm dominating, but Croatian league isn't a top one, and my team althoug clearly better than the rest of the teams (only Dinamo managing to cope with me, esp. in the Cup where I lost to them now 2 years in a row) is far from the great team. Some folks here say that managing a top team is completely different. Perhaps it is true. If I ever take the helm of such a team, we'll see.

retired_FaultyFate

No, it's the other way around. If I am right and "anomalies" just messages about something, then it's stupid, because instead of giving a user information AI confuses and frustrates him. I am even ready to call it cheating - sending wrong info is worse than not giving any.

I gave oversimplified examples just to illustrate my idea. It could be more complex or even messy. Irl football almost every goal allowed is someone's mistake. The thing is that those mistake mostly not obvious (should have done one more step, been one step to the left, etc.) and often forced by opposition. In FM when a star player gives stupid back pass we don't know whether it's a message that he lost concentartion (something else), or simply match engine does not have unobvious mistakes in the code, so it displays stupid one.

What I am pretty sure about is when we see similar "anomalies" in a few games in a row, it's not because of code-cheating, but rather because a user has not reacted to AI "message". Notice, the "anomalies" do not necessary result in goals, therefore we do not always count all of them. As a result they seem to come out of nowhere.

Not sure how well AI adapts to a user tactic, but let's for a moment assume that AI manager behave like a real manager would. It means that Manager B watched my previous game against Team A and picked something up. The next question is whether Manager B have players, tactical knowledge, etc. to replicate approach that worked for Team A, if I use the same players, tactic etc. If everything in Manager B favor, he wins, otherwise he might be able to explore some holes, but not all, or even none. I don't see anything unrealistic, and am pretty sure it is relatively easy to program.

retired_jacque90

I disagree, but just a little. There's a difference between something being uncertain and something being unpredictable (and no, I'm not playing with semantics).

If the Arsenal starting XI plays, for example, a team two or three divisions below them I can say with certainty that Arsenal would win. However, cannot predict the final score, who will score, in which minute and who will come out injured because it is not possible.

I know I might get egged out of the stage when I say this, but I think it holds some true: In football, certainty coexists with unpredictability. If certainty did not exist, in football or in FM, then we have no business even being here talking about this. Let's go home. Close down the forums.

Both in football and in FM, we do things because we expect a result. I'm not talking about scores. I'm talking about plans. I tell my fullbacks to join the attack because I want them to be up there. I tell my player in the hole to hold up the ball a little so he has more time to spot team mates. I tell my DMC to man mark someone because I want to take that someone out of the game. And so on. All this doesn't negate the existence of counters, and it doesn't mean you should just go into matches with only one plan and expect it to work. If we didn't have certainty, we wouldn't have tactical counters. They wouldn't exist. They wouldn't be called counters, at least. We'd be doing things blindly and randomly, hoping to get any result.

And this is a sound idea, despite how horrible it might feel to some. There are a lot of things in football that are done to get a determined result, and they happen. Take a look at the last 20 years of the leagues in England, Spain, Italy, etc... and see who the first five teams are. By far, it's the same names, over and over and over again. If football didn't have certainty, just randomness alone would decree that the first five would be different teams every year.

But it isn't.

No one can predict if a perennial title contender will win or lose any single game. The game is unpredictable. However, in the long run, in the majority of cases certainty does kick in. In the long run, the game is certain. The problem I'm seeing with FM is that certainty not being there in the long run, or at the very least not when a human manager is concerned.

This is not me saying that if we take over a title contender team, one of the 'big five' regardless of the league or whatever, we should win. This is me saying that all things being equal, these teams should have an easier time getting up there at the end of the season than other teams. This simply doesn't exist in FM as it is now, and it's just not realistic at all. Seems at times you struggle just as much managing AC Milan as you do managing Ternana.

If this was the case in real life football, we should almost be seeing a random champion each year. But we don't, do we? I don't think it's "challenging" to have the same difficulty to succeed managing Arsenal and managing Sheffield. I think it's borderline retarded, to be honest.

I think for some reason we're quite happy to accept that in real life Arsenal will have it much easier, in many ways, than Cardiff to win the title. But when it comes to FM, it should be even? Is that what we're saying? As soon as a human manager steps into Arsenal, the 'Arsenal aura' dissipates? They shouldn't have it easier?

I'm not disputing that smaller teams shouldn't be able to beat bigger teams. It does happen. What I'm saying is that, in FM, the difference between 'big' and 'small' teams is almost non existent the moment a human manager enters the picture. Normally, a side with the quality of any of the big five, in any league, shouldn't have to deal with half the problems that a relegation contender has to deal with. But we're not seeing that.

In fact, I'd go as far as to say that having the better players and a plan that makes some sense, whatever that plan is, should be enough. All the different teams that made it to be top 5 year after year, in leagues all over the world, in the last 20-25 years had different managers, played different tactics, in different weather conditions, facing different injuries of different players, affected by different referees, etc. But the only constant in all this is those teams always getting the best players they could and a more or less decent manager at least. That's the only common factor.

Because, after all, if football is unpredictable, and smaller teams should be able to beat big teams as often as it happens in FM, and nothing is certain, etc, etc... how do we explain the same big names always being up there, year after year in real life? Hm?

retired_rayrhi

The problem is, Chand, is the bias favouring the AI. Take for example, current form and morale. In FM, these concepts are vital components when it concerns the human. We all know how difficult it is to get out of a slump. I'll give you an example: Last night I came up against a despondent Boro side. 8 games into the season they were rock bottom, winless, and the team was full of red morale arrows. I was 3rd at the time and unbeaten. I was flying high with confidence so I saw no reason other than to attack them even though I was away. I lost 3-0. The next match I played against Villa who were second from bottom, winless, and again, suffering from very low morale. This time I was home. I lost 2-1. It's amazing to me that two teams won their first games of the season against me. In FM, every time you play a team, all other factors that contribute to the performance of the human are not considered with regards to the AI. It seems that every time you play a new match, you are doing so against the base level of that team, meaning: even if Arsenal are having a pants season against the rest of the league, you can be sure as anything that they will put 4 past you at the Emirates. I genuinely think that this could be one of the reasons the AI has the upper hand. It's not worth going into a game with the mindset that you are going to destroy a team because they are low on form on morale, because that team are still going to play just as well against you regardless of outside factors.

retired_jacque90

Yes, that's something I noticed as well. It's easy to notice if you know where to look. Like in the example I posted not too long ago of my Inter side and Lazio. They lost 4-0 on sunday, then met me that wednesday and slotted a few past me as if nothing had happened. And of course, whenever it's me that is trashed 4-0 I have to work some kind of miracle to keep morale up. They seem unaffected.

Is this where part of the overcompensation is coming from? That we always seem to meet AI teams either unaffected by morale, or at the top of their game?

If so... I don't know if I'd call it cheating, but it hits the post.

retired_rayrhi

A few people have mentioned that they feel the game will start throwing the numerous already spoken of 'anomalies' at you if your tactic is flawed. This would be fine if there was no observable match engine, but...

if what people are saying is correct, then the match engine is useless. If every time I set up a tactic that is flawed and I am punished by late free-kicks, odd goals, and 1 shot 1 goal syndrome, then all I am doing is, like I already said, being punished and that's it. It's not possible to eliminate bad habits if we are not witnessing the true consequences of them. It's not enough to simply claim that a tactic is flawed and that's why you are losing at home all the time. When the AI scores a goal, it just 'does.' The goals don't seem to be linked to any particular passage of play or any flaw in your system, it just 'scores' and that's it. Thus, there is no learning curve, no system set in place in order to learn from your mistakes.

I recently attempted a 4-3-1-2 with Newcastle and my brother did with Everton and we failed miserably. Our settings were implemented logically and with 'football' in mind. We never got through the season because we were doing so poorly. I'm talking almost relegation poor. So, we started a new game with the same teams and we both now are playing a basic 4-4-2 set up 'logically' and once again with 'football' in mind. We are both battling for Europe now (4th and 5th respectively) and it's been an enjoyable season. My question is this: if my footballing theory is the same from tactic to tactic then why does one work and one doesn't? You would think that I would be able to discover the answers by watching the match, but I can't. If a tactic is good, I win, if it's bad, I lose. There's no rhyme or reason, it just 'is' or it 'isn't.' I know that a 4-3-1-2 is susceptible down the wings, but I never saw the AI exposing this. They just 'scored.' I think that this is the reason why a lot of people feel that the game is just a rock, paper, scissors calculation. It's because there is no clear pattern or observable truth that one can apply to another given situation. The only indication we have at the minute as to whether a tactic is good is when we win. If our tactic is bad, we lose. If it's okay, we draw. You've all been in a situation, I'm sure, where you have painstakingly and proudly set up a defensive 0-0 bore draw tactic for the upcoming Man Utd away match, tweaking each slider logically, paying attention to the most minute detail, only for it to be annihilated after just 2 minutes by a Paul Scholes rocket and then going on to lose 3-0. Your heart sinks and you wonder how and why. All that time for nothing. We all saw Reading do it to Man Utd over the weekend, but why can't we do it? We think we know how, but we don't. The game does not think like we do. That's why we need a significantly improved manual, but I digress. In its current state, the game has no need for an observable match-engine. Until it begins to reflect our tactical set-up and starts to communicate to us in such a way that we can 'react' to what we are seeing, the engine will always just be a luxury gimmick and we may as well revert back to the old commentary screen, because the game at the minute is not ready for a 2d-match engine I feel.