Stats and why you can't trust them.

Agreed on the circles.

SI have to protect their interests, you will never ever see the source code.

FM is a great game... but at times like this I am reminded of Sick as a Parrot... another great game, really makes you manage your team as opposed to simulate a footballing world. I think it has a stronger match engine than FM though it doesn't have the full 2D match view, and you have to really work to get your victories... even on easy...

The stats range in game for attributes is -127 to +128 for most attributes - no I'm not going to explain how it works, its always been the same in our games and hasn't changed.

I'm afraid that they're done in this manner for programming reasons and personally unless you're one of the people involved in making the game I wouldn't worry particularly about the mechanics behind the scenes .... after all the whole point of the game is to simulate the real-world, how the game does it isn't meant to be something you worry about ...

I am not so much worried by "how the game does it", but that the game itself is affected by a flawed process which results in the in-game stats not accurately reflecting a players true ability for a given attribute. Given that those stats represent a large part of the game I do believe that it is of some concern to the gamer.

Well, I can go no further with this it seems, so I'm done for this thread I guess. Shame though. Thanks for having a bash at getting me an answer anyway Tom.

Well the problem is that the statistics just influence the number that is used by the game. As we don't see that number, we just don't know who to select.

There are three solutions: adjust the statistics so that you see a 20 if he is actually a 20 after calculations (but this would require the database to be completely reworked, since a 20 would not necessarily be a 20) OR make the engine more attribute-heavy, which I don't anticipate SI doing. Actually nothing will probably change, but if someone wants to write a program to change the stats I would use it... if I had the game

The main problem I think is that the visible stats is completely meaningless to how good a player actually is. And if you don't use any tools, there is really no point in scouting for players based on visible stats.

If SI stick with the -128 to 127 range, it should scale it back to a 1 to 20 on the display.

Hence, from -128 to -115, the visible stat is 0..
from 100 to 113, the visible stat is 19
and from 114 to 127, the visible stat is 20...
so for each visible stat, there is a range of around 13 in RA.. (roughly 5%)

There is still some mystery as to exactly how good a player is, but the range for a given visible stat rating won't be too large.

SI really shouldn't rely on gamers to use external tools to play the game properly...

I reckon SI should keep the -128 to 127 stats hidden, and only allow external programs to see the 1-20 range. Gamers can set the band of which the player's stats are at, but the actual figure is calculated in game.

shrekiejai said:

The main problem I think is that the visible stats is completely meaningless to how good a player actually is. And if you don't use any tools, there is really no point in scouting for players based on visible stats.

Click to expand...

That's just not true.

If SI stick with the -128 to 127 range, it should scale it back to a 1 to 20 on the display.

Click to expand...

It does, but reliant on CA.

Hence, from -128 to -115, the visible stat is 0..
from 100 to 113, the visible stat is 19
and from 114 to 127, the visible stat is 20...
so for each visible stat, there is a range of around 13 in RA.. (roughly 5%)

There is still some mystery as to exactly how good a player is, but the range for a given visible stat rating won't be too large.

Click to expand...

You're forgetting that RA isn't the only deciding factor for what the player's stat will be.

SI really shouldn't rely on gamers to use external tools to play the game properly...

Click to expand...

Do they? I feel the game is fine without using any external tool. I've never found any player to play different than his stats would suggest. The only problem is that sometimes players can be very unprofessional or unambitious or something of the sort, and that will have them play too aggressive or have them just lack motivation.

I reckon SI should keep the -128 to 127 stats hidden, and only allow external programs to see the 1-20 range. Gamers can set the band of which the player's stats are at, but the actual figure is calculated in game.

Click to expand...

SI can't do anything about external tools, as they are made by other people. Except maybe FM Scout (although it's more of a personal project by one of their employees), but you do still have a choice whether to use it or not.

The Kop Kid said:

Nope. Sorry Tom but that arguement is seriously flawed. The game makes it's calculations on those -128/127 numbers which are the final result of the RA and CA.

Fredheim Holm has current ability of 180, Visible finishing stat of 19

Vagner Lover has current ability of 145, Visible finishing stat of 16

Therefore, as far as the gamer is concerned Holm is going to be the better finisher on ALL appearances.

Click to expand...

Ahh...but you're missing something important...Current ability CHANGES, AND if you have played the game for more then one game, you'll realise that players 'visible stats', as you put it, also change...>
WHEN the player sees Holm as having 19 for finishing, and Lover as 16, then at THAT instant Holm IS a better finisher then Lover...IF Lover's current ability improves, then his finishing goes up with it...And if Holms current ability decreases, again his finishing decreases.....It makes it like real life..The number between -127 and 127 that you speak of his how good naturally the person is at that particular skill (in this case finishing). The ability comes and goes during your career based on your 'current ability'...

SI simply have a formula to calculate a current finishing ability based on that finishing number, AND the current ability of a player, and the result is the number out of 20 that you see (altho Im sure that the number out of 20 is actually calculated to a number that gives much more detail then 20 within the game engine)...

BUT

The database calculates that Vagner Love dispite all VISIBLE indications actually has a better finishing stat. You guys can go on about current ability and weighing values as much as you want but in the end it all comes down to the game displaying that one player has a better finishing stat whilst in fact the other does.

Click to expand...

The FACT is that at that particular point in time Holm IS a better finisher then Love....If Love ever reached the overall ability of Holm, then he would be a better finisher, but at that point in time, Holm is better....

TRY THIS...

Find a striker with finishing 20 and a reasonably low hidden count, say a +16. Play him for 10-15 games in a team where he has scoring opportunities. Then use the in-game editor and change the hidden count to 127. He still has a visible stat of finishing 20, he still has the same potential ability. His "confounding variables" are still exactly the same. You will find a huge difference in the amount of chances he finishes. Why? Because the -128/127 count is 'God' and not the corresponding visible stat which is often at variance to it.

Click to expand...

Again, as I said the game does calculate the number to a value in greater detail then a number out of 20....So of course even if there was two players with identical current ability, and both with a visible stat of 20 for finishing, if one of them is really a BETTER natural finisher, then he will finish more...There is more discrepency in the game then simply a number from 20...

However use a bit of logic (something you seem to be lacking, have you done much/any programming in your life?) and consider making two players, with attributes of 15 for EVERYTHING, but one with a higher current ability then the other...In this case, the players should actually be THE SAME good at playing in the game. We dont see current ability in game, and it doesnt directly effect anything (except scouts and maybe player value I believe)...This is because the game uses the CALCULATED figure...It would mean that one player would have to have a higher 'natural' finishing rating, to cancel out his lower current ability, BUT the game uses the CALCULATED number ingame.

Your example of 20 defies logic, simply because, YES this is the one area the game isnt accurate, because 20 signifies as good as it gets, when really there are several levels of 20 in the game...Every other number there ISNT, and the calculated figure is used....

Basically imagine the calculated figure is out of 25, and 20 actually stands for 20+...

Rereading the thread, having ratings changed to out of 100, instead of 20 may add a little in the way of the players actualy rating reflecting their true ability BUT it will also have the problem of not being as realistic...I mean it means you have a bit of leeway in your judgement like IRL...Imagine if Wenger was given an accurate rating for every player in the world he scouted or had at Arsenal, out of 100....

The Kop Kid said:

I am not so much worried by "how the game does it", but that the game itself is affected by a flawed process which results in the in-game stats not accurately reflecting a players true ability for a given attribute. Given that those stats represent a large part of the game I do believe that it is of some concern to the gamer.

Well, I can go no further with this it seems, so I'm done for this thread I guess. Shame though. Thanks for having a bash at getting me an answer anyway Tom.

Click to expand...

The process definitely isn't 'flawed' and works as intended, the equations involved aren't purely linear and are fairly complicated - feel free to continue attempting to decode things, but imho it won't add anything to your enjoyment of the game ..

shrekiejai said:

SI really shouldn't rely on gamers to use external tools to play the game properly...

Click to expand...

Err in what way do you have to use external tools to play the game 'properly'? - these values which you're debating are internal calculations used by the game and aren't intended to be visible and indeed only are through external tools ... everyone discussing them here bar me don't know the reasons behind the scales how they're used in the game engine etc. ... simply put I'm afraid you'll have to trust me on that they make sense and work reliably for their purpose, they've been very very similarly implemented for around 8-9 years now and so bar minor tweaking I doubt we'll change the system.

I reckon SI should keep the -128 to 127 stats hidden, and only allow external programs to see the 1-20 range. Gamers can set the band of which the player's stats are at, but the actual figure is calculated in game.

Click to expand...

Anyone can view any stat with an external program, we could convert the stats to a 1-20 scale in saving so external editors would see that value but it'd slow down the game save and as (proportionately) few people use these editors that wouldn't make sense imho.

Some past save game editors have displayed the 1-20 range (roughly) incidentally - they've never got the equation 100% right but some have got pretty close to it.

mlsmanager2002 said:

Well the problem is that the statistics just influence the number that is used by the game. As we don't see that number, we just don't know who to select.

There are three solutions: adjust the statistics so that you see a 20 if he is actually a 20 after calculations (but this would require the database to be completely reworked, since a 20 would not necessarily be a 20) OR make the engine more attribute-heavy, which I don't anticipate SI doing. Actually nothing will probably change, but if someone wants to write a program to change the stats I would use it... if I had the game

Click to expand...

Your comment confuses me a little statto as you always tend to be an advocate of not seeing any attributes so the fact that their internal range differs slightly to the visible range should please you surely (as it adds a little 'fuzz' to things and you never quite know which 15 rated player is slightly better than another 15 rated player ..).

Incidentally the game also uses players personalities and mental state to slightly alter these stats in matches - this isn't visible apart from through analysis of performance and commentary (ie. when a player is very motivated or wound up) ... this is realistic as players do play better or worse when compared to their ability at times imho ...

Hence you might have a 17 year old kid and Alan Shearer as options for taking a crucial penalty in a Cup Final, despite them having the same Penalty stat its very very likely that Shearer will cope with the pressure much better than a young kid (although not 100% the case) because of his experience and personality ...

see this is what im on about, i personally enjoy the mystery of slightly different players with the same attributes.

TROD.

I agree 'fuzz' is good. In the game I programmed I used something similar.

However, when a 20-skilled shooter is not as good at shooting as a 17-skilled shooter, purely based on statistics (that may or may not turn out to be true, as was the point of this thread) that is where problems arise in my book.

Perhaps you could institute some kind of scouting feature that says 'he's got good shooting/tackling skill but he may not perform consistently' to remedy this problem?

I was going to leave this be, but as we actually have one of SI here, what the hell...

Marc Vaughan said:

The process definitely isn't 'flawed' and works as intended, the equations involved aren't purely linear and are fairly complicated - feel free to continue attempting to decode things, but imho it won't add anything to your enjoyment of the game ..

Click to expand...

I think that you missunderstand me. O.K, as an example let's take the player who is IMO probably the most extreme example of excellence from the game's history. Maxim Tsigalko in 01/02. Granted, the guy had a finishing stat of 20, but his overall stats weren't that great. There were other players in the game who also had finishing 20 and had better 'supporting stats' such as positioning but who couldn't hit the side of a barn, yet this guy could bag 80 goals in a season. There was absolutely no way whatsoever to tell that from the presented information that this guy was a football god, and I played the game for an age before discovering him by pure luck.

The same problem still exists. There are players with, for example, finishing stats of 20 who simply cannot finish, whilst others with a lower finishing stat can. All hidden stats and mental stats aside, if the game has a stat of 'finishing 20' displayed for a player and the player cannot finish, then that stat if pointless. When the data (stats) presented does not translate accurately into performance, then yes, I'd call that a flaw.

Understand Mark, I'm not trying to 'decodethings' as some kind of maths excercise, neither am I trashing your product, having bought (note, not downloaded) the game for a few years now. Unlike most though your company do seem to actually take on board what gamers say so I have felt that this topic was worth persisting with.

see its a difference of opinion, sure the stats range from more than 1-20 but it depends on what you want, on your basis you want everything out in the open so its more of a 1-50 thing, so Tsigalko would have finishing 50 and defoe only 45. However, im of the opinion that you cannot see everything in football, not every nook and cranny, so i think the mysticism of two different 20's is actually better for the game.

TROD.

mlsmanager2002 said:

I agree 'fuzz' is good. In the game I programmed I used something similar.

However, when a 20-skilled shooter is not as good at shooting as a 17-skilled shooter, purely based on statistics (that may or may not turn out to be true, as was the point of this thread) that is where problems arise in my book.

Perhaps you could institute some kind of scouting feature that says 'he's got good shooting/tackling skill but he may not perform consistently' to remedy this problem?

Click to expand...

The reason that some players might excel despite having slightly lesser stats is normally due to one of several factors (ditto vice versa - a player with great stats under-performing):

* He's being played in a manner which simply suits his strong points.
* He has mental flaws/strengths which affect his game. For instance a player with slightly worse stats might have a very strong personality:
eg.
Pressure
Unaffected by pressure during big-matches he might out-perform a player who looks better than him on paper who can't handle such pressure

Professionalism
A professional player might perform much better than an unhappy unprofessional player who lacks motivation to try

Lack of work-rate
Some players with huge talent (ie. Le Tissier in his prime) might still drift in and out of games because they almost lack work-rate to persevere with things.

Lack of concentration
Some players might lack focus and concentration thus making silly mistakes more often than other players, this is often the case with youngsters.

etc.

(there are loads of these things kicking around in the game, each complicated by reliance on other factors in the game - simply put try and think how people react to things irl, thats what we're trying to model)

* Weaknesses offsetting skills
A player with 20 finishing for instance won't automatically be able to finish all chances, if he has a poor 'off the ball' attribute then its likely he'll find himself under pressure for a lot of his shots and so might end up scoring from fewer chances than a player with low finishing but high off the ball.
Similarly you'll sometimes find during matches that a player simply under-performs because he's good - sounds strange? ... ok I'll try and explain ...

Everton are playing against Arsenal, in a freak happening most of the Arsenal strikers are injured so they're playing Henry and a 19yr reserve player up front.

Now Everton are well aware of the threat from Henry and so man-mark him, the 19yr reserve on the otherhand is almost dismissed as a threat and attention is more heavily placed on keeping Henry under control - in these circumstances its possible that Henry will stay quiet the entire match but the 19yr might find himself with space to play ...

Hope this helps,

Marc

Perhaps you could institute some kind of scouting feature that says 'he's got good shooting/tackling skill but he may not perform consistently' to remedy this problem?

Click to expand...

Scouts and Coaches do mention this sort of thing during their reports if they're good enough ... similarly with positive aspects of a player.

(ditto with physio's regarding injury prone players etc.)

Marc Vaughan said:

* Weaknesses offsetting skills
A player with 20 finishing for instance won't automatically be able to finish all chances, if he has a poor 'off the ball' attribute then its likely he'll find himself under pressure for a lot of his shots and so might end up scoring from fewer chances than a player with low finishing but high off the ball.

Marc

Click to expand...

While we have you here..

One of the things which has frustrated me a bit regarding the finishing (stats aside) is the number of 1 on 1's with the keeper that are missed. My striker is often in a mile of space with only the keeper to beat, only for him to smack it either straight at the keeper or put it wide on the majority of occasions. This seems to happen regardless of the striker used.
In real football the chances of a top class striker scoring in this situation are extremely high, whereas in the game it's probably only around 25 percent.
Have you guys or any of your testers noticed this and will there be any changes for the 2006 version for this or any other unrealistic in-game occurences that you (SI) may have picked up on?

Also, if you have the time, which of the new features in 2006 do the lads at SI believe will have the biggest impact on the gameplay?

The Kop Kid said:

While we have you here..

One of the things which has frustrated me a bit regarding the finishing (stats aside) is the number of 1 on 1's with the keeper that are missed. My striker is often in a mile of space with only the keeper to beat, only for him to smack it either straight at the keeper or put it wide on the majority of occasions. This seems to happen regardless of the striker used.
In real football the chances of a top class striker scoring in this situation are extremely high, whereas in the game it's probably only around 25 percent.
Have you guys or any of your testers noticed this and will there be any changes for the 2006 version for this or any other unrealistic in-game occurences that you (SI) may have picked up on?

Also, if you have the time, which of the new features in 2006 do the lads at SI believe will have the biggest impact on the gameplay?

Click to expand...

Actually if you check the real-life stats the exact amount of 1-on-1's with a keeper scored is a paltry 30%, this is roughly what it is in the game (can vary up and down from this a small percentage depending on luck and short-term trends) ....

Even for very good strikers the ratio doesn't really rise much above around 35-40% during the career of a very good striker ...

Think about it, a player only requires one poor game where he misses 4-5 chances to undo his average, every player even the best have barren spells ...

Im away for 8 days and immediatly interesting discussions come up I bring the quality of posts down, as you can see with this one)