Chandaman

Just a few things.

kolobok:

See, I wouldn't be so sure. I experienced the usual kind of anomalies no matter where I played, although I couldn't tell you if it's better or worse depending on the league. As far as I know, it happens in the leagues I usually start in (or end up in), which are normally Argentina, Spain, Italy, England, France and Holland.


Luketi:

Right, of course. I don't think anyone is arguing to get the same result over and over again after each reload. Even if it's just an experiment. What your experiment suggests is that the statistical spread is quite off.

If we assume that better tactics and better players will tend to triumph over worse tactics and poorer players (not an outlandish assumption, I don't think), then you would expect results in that vein. When you run your experiment, at the end of the run you should see a majority of wins for the team with better tactics and better players. The bigger the sample size, of course, the better. But I think we can agree on this point.

However, you ran your experiment and you got a spread all over the statistical board. This could mean two things, if our assumption of tactics and players is correct:

- The two teams were close to each other in terms of tactical and player quality, therefore the experiment just happened to get the two teams that would yield results wildly one way or the other.

or

- Out initial assumption was wrong, and tactics and players do not affect the outcome of a game as much as we think they do. The inherent uncertainty of the system and the huge number of variables that affect things for 90 minutes are a bigger deciding factor than the initial experimental conditions of tactics and players.

I guess that's what we're trying to assess here. I think a better experiment would be to pick up a team and make them go on a run of (x) matches against three other teams. These three other teams would be picked at arbitrary 'quality points' within the same league, just to observe how the statistical spread changes.

For example, we pick ManU, and we set it up on a friendly that we save beforehand against a relegation candidate, then a middle of the table team, and finally another title challenger. A priori I would assume that, given all initial conditions being equal and we eliminate as many variables as we can, the final results should look something like this:

ManU vs Relegation Candidate - (x) games played:
70% ManU victories - 20% draws - 10% Relegation Candidate victories.

ManU vs Middle of the table team - (x) games played:
60% ManU victories - 20% draws - 20% Middle of the table team victories

ManU vs Title Contender - (x) games played:
40% ManU victories - 20% draws - 40% Title Contender victories.

(don't go by those numbers, that's just to illustrate how I think the spread should logically look)

Maybe it's worth a shot. Just to see the numbers, if nothing else. Naturally, the more games you run, the better.


arteta:

Yup, pretty much. Bears repeating. In fact I'd say feedback is more important than consistency, in some situations.

wwfan

That might have had something to do with it. I am winning but I become more offensive? Surely, grabbing the third goal is best done on the counter. If Derby become more aggressive in trying to get back into the match, leaving the wings exposed isn't going to do you any favours. Especially when you have it all to play for and they don't and can throw caution to the winds. Plus, you were unlucky. However, I can guarantee that had you made the right half-time teamtalk and the right tactical swithc (deeper, tighter, less forward runs, mnore timne-wasting) you would have comfortably seen out the game from a 2-0 half-time lead. </BLOCKQUOTE>

I appreciate your reply, but all you are doing is emphasising that the AI has the upper hand at all times. You claim that I had my wings exposed by having my full backs get forward. I have been doing this all season. Why aren't I overpowering THEIR wings? Why should I have to react to the AI? I'm the dominant team and I'm playing at home. You have no proof to claim that Derby 'threw caution to the wind' just like there is no way you can 'guarantee' that had I selected the correct team-talk (how do you know I selected the wrong one anyway?) and changed my system that I would have comfortably seen out the game. I am not sure you can predict varying outcomes unless you have a time-machine. Even then, if perhaps you are correct, then how the bloody hell is this somehow linked to conceding 3 penalties? And why oh why did this happen on the very last day of the season when I needed a win for promotion? Furthermore, why did I lose a 2 goal lead to be dumped out of the play-offs against Leicester? I looked through my matches, and of all the games I had lost (only 5 in total), not once did a team come from behind to get a win. </BLOCKQUOTE>

Arguably, the AI does have the advantage sometimes because it makes the right decisions when we don't. But equally, we have the advantage when we do and it doesn't. I have lost count of the times I have seen the AI continue to come at me as I ripped them apart on the counter, or not go defensive enough to combat my home system. However, I have also seen comebacks because I screwed up. My perspective is always 'I screwed up' if I lose/draw form a winning position. Blaming the AI doesn't even come into it. Sure, the AI is not perfect but it is a too easy a scapegoat when things don't go the way you expect.

Yes, 3 penalties were unlucky. However, PaulC has explicitly stated that anxiety and nevousness are coded into the engine. Needing to win, conceding a goal and having no covering full-backs is likely (based on PaulC's analysis) to bring about desperate defending with a one goal lead in a must win match.

arteta is god

...and you know what, mate, if this is in fact true, then brilliant. Unfortunately, and obviously, there is no way to prove that this indeed was the case. If this was the reason then I could absolutely take it on the chin. I just don't know though. Once again, we are dealing in theory. In all honesty, I find it highly doubtful that a. I would lose after winning 2-0, and b. I would lose to 3 penalties.

wwfan

True, but it is equally true you can't prove the engine cheats. You offer me one comeback during a season as proof of cheating, I return with PaulC's statement that anxiety is coded to happen in such situations (and would be more likely with a pressuring teamtalk and attacking formation). If, as in the early stages of FM06, comebacks were happening en masse (and SI did admit a problem with player conditioning was leading to this) then your point would stand.

As it is, the only 'proof' we have been offered is PaulC's (the engine's lead programmer) analysis of why comebacks may happen in certain situations. He is either telling the truth (in which case you are at fault with your tactical planning) or lying (in which case you are right in you assumption of the engine cheating). Having followed his advice, I don't see many comebacks, even in big matches, so I assume he is telling the truth.

Millie

With regards to "nervous" players, I notice it a lot with youngsters. I've been playing with Arsenal and I get it a lot with Ryan Garry, Justin Hoyte and Djourou (though, their lack of talent may also explain their idiocy )

The "no pressure" or "faith" team talks tend to calm them down, but obviously by half-time it may be too late.

arteta is god

You are still shying away from the ridiculous events that occurred. The fact is, I am dealing in fact, and you are still dealing in theory. Even Paul C's explanation is not sufficient to explain it. Why am I dealing in fact? Well, because the stats don't lie. It's commonly agreed now that the AI enjoys a higher goals/shots on target ratio. To me, this is enough to prove that the game over-compensates. This is where my argument ends. Hammer, me, and many others have shown numerous screenshots. Even Millie admits that it happens. I am not trying to prove anything other than this. To me, this is proof enough. To you optimists, it's down to human error. This is where we disagree and probably always will. When formulate it like this:

Tactical Ineptitude is directly correlated with the AI enjoying fewer shots on target but with more goals than the human user

it sounds ever more ridiculous. And this is actually the argument against mine, Hammer's and co.

My argument goes along the lines of:
The AI has a significantly higher goals/shots on target ratio than the human user. We have numerous postings of evidence to prove this.[/B]

Your argument fails because there is no feedback system and no fact. It is all theory. You can't really hide from the fact that these anomalies occur. Hammer, Chandaman, myself, and even Millie will tell you this to be true.

Chandaman

Simple question to illustrate why the lack of feedback is a killer.

Picture this scenario. Happens I don't even know how many gobs of times in every league in the world, in any team.

Two teams meet and play somewhere in the middle of the season. The teams are irrelevant for the purpose of the example. So are the tactics used by both. First half ends 0-0. And unremarkable game.

I'm managing one of the teams and, at halftime, I do two things:

- Using the team talk, I try to encourage them to go try for a goal more ("encourage team", "one more goal you can do it", "where's the passion lads?"... whichever form it takes).

- I switch to a more attacking formation.

Second half starts. Ten or fifteen minutes later, my team scores.

The question: Did we score because I gave that team talk? Or because I switched to a more attacking formation? Both? Or is it neither and we just happened to score?

Can we point at one thing and say we scored because of that, with certainty? Can we really talk about cause and effect, and say without error "Yes, we adjusted to A, so B happened?"

arteta is god

And since this thread is aptly named total lack of logic, I have a good one to report:

After the shock of losing to 3 penalties and being dumped out of the play-off semi's by Leicester, after another remarkable 3-2 comeback, I recovered and began my next attempt at promotion glory. After bringing in Sahar, Samba, and Matty Taylor, I have a squad brimming with Premiership quality.

I had a great pre-season, banging in 20 goals in 5 games. My first league games were:

Birmingham (A)
Leeds (A)
Wolves (A)
Burnley (H)

Three away games! That's strange in itself. Anyway, with my slightly updated and more attacking 4-4-2, I managed this:

Birmingham (A) W 3-0
Leeds (A) W 5-0!
Wolves (A) W 3-1 (I'm getting in to this)
Burnley (H) Uh oh!

Burnley were 23rd in the division when they came to the Stadium of Light. The league position is not too relevant after only 3 games, but the fact that they have conceded 4 to Swansea, 4 to Bristol City, and 3 to Gillingham, IS.

Walk in the park, right? Well that's what I thought. I know I sound like a broken record by now, but I can not restrain myself from sharing my pain. I know the game by now, so I checked a few things: Their morale ranged from low to really low. They had scored only once in those 3 games they lost. 2 of the games they lost at home. The odds were 1-6 for me to win. In other words, there is absolutely no reason for me not to attack them. I should not have to worry about their tactics at ALL.

The game begins:
The match starts as you would expect. Burnley play a 4-4-2 and are backs to the wall. Somehow, they hold out for the first half. The possession is a staggering 69% v 31% in my favour. Mark Noble, my maestro, had touched the ball 78 times already. He's running the show playing a 9. He gets man of the match most matches. I have had 14 shots on target and their keeper is playing a 9. Burnley have had 0 shots on goal. At this point, you would have to say that I have my tactics spot on. Surely there is nothing else I can be doing? Maybe bump a few mentalities up, play a bit wider? It's not necessary though because I like the way we are playing and I am sure we will get a goal in the second half.

The second half kicks off. It's basically the same as the first half. Burnley change to the infamous 5-4-1 wall formation in the 56th minute and it 'feels' inevitable that I am not going to score. Suddenly, the highlights flick to my DM Gavin McCann bringing down a Burnley player. It''s the 82nd minute. 'That looked harsh' mentioned the commentary. No danger though, it's a free-kick from about 30 yards out. Up steps Joey Gudjhonssen. Need I tell you what happens?

Full time
Sunderland 0 Burnley 1
Possession 65% v 35%

Shots/on target. Wait for it!...
Sunderland — a whopping 28 shots, and a staggering 16 on target. Burnley keeper with a man of the match 10.

Burnley — shots on goal: 1, shots on target: 1, goals: 1.

It's not that I lost the match (inconsistency, complacency could be issues to name but two), but the manner in which I did is entirely gut-wrenching. If you were to see my team, you would wonder why I am not in the premiership. It's packed full of stars. Experienced, technical, determined professionals. After my first three games, you would be insulting me to blame my tactics. Clearly it is not that. I personally am certain that my team is SO bloody good, that the AI is no longer capable of matching me. If I am supposed to win every game, then just let me bloody do it. I'm not even going to go in to the next issue, but I have restarted that match three times since and have tied every game 1-1. Weird huh? One of the matches they equalised through a penalty, and in another they equalised from a corner. Another preset 'banana skin'? We'll never know. And I don't have the will to keep re-testing...

I have screenshots of the whole thing if anyone would like to see.

Hammer1000

I feel your pain Arteta mate, sadly some other idiot is going to blame your tactic, team talk or just the fact that you did'nt create enough positive energy!.

Millie

Hammer, are you actually going to contribute to these debates, or are you going to do the intellectual equivalent of hiding behind the big kid and shaking your fist going "Yeah! ner ner!"