retired_rayrhi

A few of us have decided to pull together and embark on a quest to truth. We aim to carry out a meticulously fashioned experiment that (hopefully) will once and for all explain what each and every one of the sliders actually does. A lot of us have conflicting views on the theories of our beloved game, and rather than us butting heads and making 'vicious' stabs in the dark (watch out your nan may be in that room!), we have decided to join in unison and combine all of our mighty brain power for the higher good. We are currently at the planning stage and are trying to figure out the best way to carry out the experiment. We are starting with mentality (the infamous!) and hope to advance from there. Prince capri (hero and mad scientist) has kindly offered to carry out the experiments with the aid of our flawless guidelines. Since we are in different time zones, I hope to take his observations when you are all tucked away in bed, and see if they match my own. Below are the guidelines I/we have come up with thus far, but they are running guidelines so feel free to disagree/provide feedback. We have agreed that the testing game should involve 2 human teams so we can keep the opposing teams' instructions consistent.

Basic Guidelines
— 4-4-2 formation (no arrows) — arrows may distort mentality and behavior
— All sliders except control slider should be set to the left i.e., slowest tempo, shortest passing, no creative freedom (you get the point)
— Match played in full 2d mode. How many games are conclusive? 10 matches with control slider to the left? 10 matches with control slider in the middle? 10 matches with control slider to the right? WE NEED MORE TESTERS I THINK!

Questionable Guidelines
I am more concerned with observational facts, so:
— Posted screenies of the slider in action (comparison screenshots)
— I am sure it is worth posting stats and results too...

Thoughts people?...

retired_acivin

The basic framework looks fine!
I think a similar tactic for two teams would work.

But I also think that it will work better if the teams are of very similar calibre.

retired_rayrhi

How about Man Utd and Chelsea? I think it is too much effort to make both teams equal in all aspects (editing etc.) and that the essence of this experiment could be lost if we do so. I think we are trying to discover the general workings of each slider, so one would expect to observe similar behaviours whether controlling Chelsea or Morecambe. I just hope that the opposing team does not outplay us, hindering our findings. For the opposition team, should we set all instructions to 1 (all the way left)?

Luketi

Sounds good :thup: but a lot of work to do .
Firstly, we need to define the samples we need to consider: stats or screenshots (or both).

Then, we need to have some rules about taking the samples (i.e. screenies), for example:

Team mentality experiment:

* When in possession.
* Without possession.
* In our half
* In opposition half.
Etc.

This is because we need to find a pattern to reach some kind of theory, and the pattern should consist of stable samples.

retired_FaultyFate

Ok, I agree that to see a pattern we don't need exact teams. It will be very interesting to see though how too teams with all sliders set to 1 will play Prince, it would be great if you could upload some .pkms.

We don't need to run 10 matches if pattern is clear. 2 or 3 is enough. Also, I would suggest doing changing one team mentality at once. That is, say, 2 games both teams with mentality = 1, then one team with 1 the other with 10, both 10, etc.

Good luck.

retired_tammynoel

mathematically and scientifically speaking - it doesnt matter what the other sliders are on...just so long as they stay the same for each match...and it would make sense to keep these the same across all the experiments so i would suggest keeping everything at normal (middle) except the slider (tactic) you are investigating...this should also reduce the interferance from the relationships of each slider...

also in order to reduce the other unknown variables...those that are in the ai but not visible or controllable via the tactics screen you will need a large sample to eliminate these ...so 2 or 3 is not enough - not even close - 10 is not enough - you will prob. need at least 100...there are formulas for working out the right sample size...although thats hard without knowning the (number of) unknowns...

also editing the teams to ensure that every player in each position matches the opposing player is a very good idea as this will eliminate further variables that will effect what you are trying to determine...it might make sense to set all the attributes of each player to 20 to eliminate the "poor decision making" and "ability" as best as possible...

conducting a proper valid statistical experiment is not easy (well - not for the feint hearted - you have to crunch an awfull lot of numbers - which means you have to get those numbers)...

if you set this up with too many unknowns your experiments wont give you any answers that can be reliably believed or interpreted...

remember this is an experiment to determine what one individual slider (tactic) does...the more of the other variables that are eliminated the easier it will be to deduce what each slider does...

retired_kellyk

I hate to say this...but you guys may want to search deep in the forums, these kind of tests were done months ago by some of the older board members when the game was initially released. Strangely enough most of them have headed off to OTF, since this place started becoming a cauldron for "tactic requests" which explains why I take such a strong stance against it.

It wouldn't hurt if you guys had a look at it..they did...quite a lot of work. I know they reached conclusions on mentality, width and tempo and then there is Abramovics thread as well on the AI...hmm there should be enough there..

From what I gathered...mentality affects positioning and orientation of play...confirmed.

Team slider mentality has an effect on other sliders..not to the detriment of individual positioning...confirmed.

Defensive Line affects closing down...w/o ball


You guys may want to have a look at it.
I

retired_rayrhi

I seem to remember it being done for 06, but not 07. Do you have any idea, Rashidi, as to the thread name so we can search and possibly save ourselves some time?

retired_rayrhi

I agree with almost all you mentioned apart from changing all players to a similar ability (editing etc.). The reason is because every team you play with should be able to abide by the principles that we discover, thus each slider system should be absolutely consistent with every team that you play with regardless of player ability/stats. Yes we know that football players are dumb, but even Earl Barrett could follow simple instructions (to a degree) and one would expect your virtual footballers in FM to be able to do the same.

retired_tammynoel

yes - i agree with that...it depends on whether you wont to determine the true meaning of the slider only - or how players of different abilities react to it...i think the first thing is to actually find out what the slider does exactly - so thats why i suggest eliminating player ability anomolies...i think that is more important...once we have that it should be (much more) easy to apply that knowledge to the level of player in the respective team...

in fact if you want to find out how much ability affects a match then you would set the sliders and adjust the players ability 20 times...

having said all that...if you play with exactly the same players each match you should be ok - it just makes it harder to determine if its the players (lack of) ability or the slider setting that is affecting the observable results...