CONTEXT
In the recent years, there has been much discussion (and many memes) in the community about the quality of certain newgens, most prominently about the lack of fullbacks/wingbacks with adequate crossing attributes. Some numbers like the ones provided by Leo here showed that there appears to be some kind of issue about the newgen system:
As the sample size provided there was rather low, I've decided to run my own experiments with a larger database and over a longer period of time.
TEST SETUP
My test setup started with only one league loaded, but most relevant players being added via the database settings. At the start of the game, there were 155691 players loaded.
I've simulated the game until 2073, where the number of players dropped to 126608.
I've then exported all stats (and ca/pa) of all U16 players for every 5 years, so for 2023, 2028, ..., 2073. Then, I've created a dashboard to visualize the results.
RESULTS
As this appears to be the most controversial topic, let's get started with wing/fullbacks ability to cross and dribble:
This dashboard includes the stats of all D(RL) and WB(RL) players from the save games from 2023 as starting point and 2073 as end point.
The first row shows different graphs mostly about the current ability of the players to put the rest of the data into perspective. As you can see in the graph in the middle, both CA and PA drop significantly in the first five years and only recover slightly over the years (51.4 in 2023, 41.4 in 2028 and 43.5 in 2073). The boxplot to the left and the histogram + kde on the right also highlight the difference. The newgens have a higher standard deviation and look more like a gaussian distribution, but the mean is also shifted to the left.
Now, when we take a look into the second row, the graphs show the stats for the "Crossing" attribute, we notice immidiately when looking at the box plot and the histogram that again the mean drops significantly compared to the start of the game. Visually, it's already easy to tell that the difference is even more notable then when only looking at the CA. In the middle, I've also added a second line to the graph, that calculate the "Expected Crossing" mean. This basically assumes that CA and Crossing are directly related, so that for example a 10% in CA should lead to a 10% drop in the crossing attribute. However, as you can see, the mean of the crossing stat appears to be a flat line completely unrelated to the CA.
@Kyle Brown was so kind a chatted with the QA team about that topic earlier, as you can see here:
When looking at the gathered data, this response by the QA team seems very strange. We can clearly see that newgens are not only weaker than those from the start of the game, but also that their crossing ability is even lower than is to be expected given the available CA.
Now, this gets even weirder when taking a look at the dribbling attribute:
The CA distribution obviously stays the same, but in the second row, we can see the attribute distribution of the dribbling ability. While the mean also drops slightly, it drop is way smaller then would be expected given the drop in CA. Also, compared to the crossing attribute, it actually correlated with the available CA.
I would highly urge the SI-Devs to take a deeper look into the system and check what's going on there.
I've published the dashboard here including the data here:
https://github.com/Svonn/FM-Svonnalytics-Attribute-Analysis
And I've uploaded the 2023 and 2073 save games.
I'll soon add some instructions and the required view to the repo so you can check wiht your own data.
Best regards,
Svonn
Recommended Comments
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now