Player Attributes are Irrelevant... I think SI are pulling the wool over our eyes.

retired_yeahurock26 · 11-30-2007, 06:22 PM

I don't quite understand what you've done... have you set the positions in the editor to different to what they are in real life? i.e. set Drogba with 1 for centre forward and 20 for central defender? So he becomes a 'natural' central defender?

A lot of the players that you mention have the stats to play in those positions, as fmadonai says, maybe you should try it with some weaker players. While I can imagine Drogba as a pretty good centre half, and Essien as a good centre forward (and even at a stretch Carvalho as a decent centre mid) it would be interesting to see what happens if you take a player who has very specialised stats (say, for example, a striker like Michael Owen, who has high finishing, composure and pace) and turn them into a position in which they are not at all suited.

retired_cglittergirl1967 · 11-30-2007, 06:24 PM

Quote:

Originally posted by james4624:
I hate to be a nerd, but a sample size of 18 matches is way too small for you to draw an accurate conclusion. To get an accurate conclusion you would have to play at least a few thousand games. If then you were still getting the same results then you would have every reason to be worried.
Having said that though I do agree with you that attributes don't seem to be as important as they should be.

Your not a nerd but this test was a full season not 18 matches as the previuos and i do think the results are statistically significant bearing in mind the actual results differ greatly from what you would expect to get. If it took 1000's of games to notice any difference it would further prove my point that playing players in correct positions doesn't matter.

retired_halfshell69 · 11-30-2007, 06:25 PM

Quote:

Originally posted by Drunk Beware:
Another thing, would it be different if you played through the games?

Good point! a previous poster showed that the team reputation was the only thing that mattered in simmed games. (i.e. he gave andorra 190CA players and they still lost 10-0 etc) if this is the case then only results where you were in control would be relevant as the others wouldn't differentiate between player attributes.

Quote:

Originally posted by james4624:
I hate to be a nerd, but a sample size of 18 matches is way too small for you to draw an accurate conclusion. To get an accurate conclusion you would have to play at least a few thousand games. If then you were still getting the same results then you would have every reason to be worried.
Having said that though I do agree with you that attributes don't seem to be as important as they should be.

I think that for an experiment as unusual as this one, 18 should be acceptable, as there's no way IRL chelsea would win many of those matches with such a backward set up.

Quote:

Originally posted by black 'n amber tiger:
Top players are going to be in the main more flexible at where they can play then say Oxford Utd players. I say do the experiement again but as Acrington Stanley!!

Not sure if this would make that much of a differnce, as at LLM level, i quite often play players out of position as they regularly have a fairly low ability in their own position so playing a CF as a CB often works out ok.

retired_ramesh · 11-30-2007, 06:26 PM

The flaw in the experiment is by setting their natural positions differently.

A crap natural CB is still going to do better as a CB compared to an excellent natural striker in the CB slot.

What you should be trying is keeping the positioning stats the same and putting them in the wrong positions.

Unless that's what you did?

retired_cglittergirl1967 · 11-30-2007, 06:26 PM

Quote:

Originally posted by Ched:
This is quite alarming, a response from SI would be highly appreciated.

Any chance of some screen shots for this? Because i dearly hope this is just a wind up, not that i'm accusing nickdundee of anything, i just hope FM08 isn't a farce.

I'm off out now but i will post some tomorrow. What screenshots would you like?

retired_halfshell69 · 11-30-2007, 06:29 PM

Quote:

Originally posted by nickdundee:
<BLOCKQUOTE>Originally posted by Ched:
This is quite alarming, a response from SI would be highly appreciated.

Any chance of some screen shots for this? Because i dearly hope this is just a wind up, not that i'm accusing nickdundee of anything, i just hope FM08 isn't a farce.

I'm off out now but i will post some tomorrow. What screenshots would you like? </BLOCKQUOTE>

Just some showing the average ratings and the positions screen showing where they played all season; not all are needed just a couple to show us what the ass man was up to.

ddong78 · 11-30-2007, 06:47 PM

Quote:

Originally posted by nickdundee:
<BLOCKQUOTE>Originally posted by fmadonai:
I concur and have done the same thing as the op, holiday and achieved a similar result.

I think you have to take into account though, if your buying fm2008 to go on holiday, then you should expect the game to take short cuts, and the individuation of stats vs the quality of team appears to be one of those short cuts.

Yes i thought of this but in my post i pointed out that in my first test i actually watched the matches and the results were the same.
</BLOCKQUOTE>

Nickdundee, this is very interesting and that's exactly why we need to be very careful.

I am afraid that you answer here is inadequate. Your previous test had other problems: too few games, possiblity of "good" club underachieving etc.

This one here, these people above say that the game probably takes a shortcut. (I don't know)

So, you can't really combine data from two tests with different problems or open issues -this does not prove a thing.
Now, I am not saying your conclusion is wrong. I am saying that it has not been proven.

I am giving you a hard time because I find this issue very interesting and hope to be checked right, I hope you understand.

retired_mixxster · 11-30-2007, 06:54 PM

Quote:

Originally posted by Lyssien:
Nickdundee, this is very interesting and that's exactly why we need to be very careful.

I am afraid that you answer here is inadequate. Your previous test had other problems: too few games, possiblity of "good" club underachieving etc.

This one here, these people above say that the game probably takes a shortcut. (I don't know)

So, you can't really combine data from two tests with different problems or open issues -this does not prove a thing.
Now, I am not saying your conclusion is wrong. I am saying that it has not been proven.

I am giving you a hard time because I find this issue very interesting and hope to be checked right, I hope you understand.

Please explain what is that he HASN'T proved?

vansmack · 11-30-2007, 07:02 PM

i was always (from fm2005 on) thinking that only CA + natural position matters, how well a player performs. now it gets proven. would like to see some screenshots aswell

ddong78 · 11-30-2007, 07:03 PM

Quote:

Originally posted by Lyvean:
Please explain what is that he HASN'T proved?

OK, assume this situation:
a) his first test's results (when he actually played the games) were distorted by some coincidental bad-form streak of the "good" team, while the "bad" team played to its strength.
b) in this second test, you can't reach players quality conclusions with holiday mode tests, because on holiday mode the game relies heavily on reputation.

If (a) and (b) are correct, then these tests prove nothing.

So, are (a) and (b) true? I don't know. Neither does Nickdundee. So proper tests have to be made.

Player Attributes are Irrelevant... I think SI are pulling the wool over our eyes.

Discuss Football Manager 2008 for PC, MAC and PSP versions of FM 2008. Click here to view our FM resources.

Player Attributes are Irrelevant... I think SI are pulling the wool over our eyes.