Preliminary Investigation into Player Training: Age vs. Career Longevity
We have long wondered about the dreaded career path that plagues some of our players and not others. I for one am getting used to having my star youngsters finding themselves on less than ideal career paths. Likewise, we all know that young players and 6/6 players fetch better money on the market, meaning PPM managers value these players more. However, is it justified? Let's have a look.
To investigate this I looked at a few scenarios. These scenarios are not designed to expose the truth immediately, but rather give a high-level glimpse into the relationship between some of the variables at play. In each scenario I compared two players. These two players shared some commonalities but also had some differences. Both players were trained using the same coaches at the same facility level removing any impact by staff and facilities. I am comparing the training results for one attribute and one attribute only, the same for both players chosen such that the quality is very similar. The table below summarizes the four scenarios and how each investigates something a little different. (Note: C/L = Career Longevity.)
|C/L Path||Different by One Season||Unknown||Different by Three Seasons||First Season at Current C/L|
So first things first, I did not list the scenarios in any particular order either than what I investigated first. With that in mind, let's get started:
Scenario 1: Players of the same age and C/L. Their C/L paths are different by one season. They are nearing the end of their current C/L rating.
|Oldest 6/6 Age||18||17|
|Attribute Quality Compared||88||86|
|Average Training (Month)||1.09||1.10|
|Standard Deviation in Training (Month)||0.04||0.04|
The above table is in place to quickly compare the two players. In this scenario the players are nearly identical with the exception of 1B dropping to 5/6 a season before player 1A. However, the end result is two players that have nearly identical training. Sitting back and looking at this scenario tells me a something. Considering these players are on very similar C/L paths, I am not surprised their training is so similar. Thus, a difference of a year in C/L path between two players won't really impact their training when they're on the same level of age (and C/L).
Summary: Different C/L path, when on the order of a season, doesn't impact training provided age and current C/L are equal.
Scenario 2: Players of the same age and C/L. Their C/L path is unknown (both 16 6/6).
|Oldest 6/6 Age||?||?|
|Attribute Quality Compared||86||86|
|Average Training (Month)||1.46||1.50|
|Standard Deviation in Training (Month)||0.06|| 0.06
In this scenario two players that in all appearances look the same were compared. However, when looking at their training there is a noticeable difference. This leads one to believe that something not visible on the player profile page is causing this difference. There has been discussion and it has since been confirmed that there are some hidden values we users cannot see. These manifest themselves in the form of fractional C/L values or "sub-levels" as they've been referred to. The principle is such that each player has a C/L value and we only see the rounded C/L value. Using a comparative training technique it is likely possible to determine if there is a difference in this invisible fractional C/L value between two players. Here, I would predict that player 2B will have a longer C/L path than player 2A since his fractional C/L appears to be higher resulting in improved training rates.
Summary: It may be possible to predict players with better C/L paths based on comparative training. This provides clear evidence for "sub-level" C/L (or fractional C/L) discussed in a previous PPM Magazine article and the existence of which is also confirmed by the developers.
Scenario 3: Players of the same age and different C/L. Their C/L paths are different by many seasons.
|Oldest 6/6 Age||15||18|
|Attribute Quality Compared||91||90|
|Average Training (Month)||1.04||1.19|
|Standard Devtiaion in Training (Month)||0.04|| 0.02
In this scenario we have two significantly different players. While both were pulled at 6/6, one dropped early and became 5/6 at age 16 while the other stayed at 6/6 through age 18 and dropped to 5/6 at age 19. They have a different C/L in addition to their largely difference C/L path. Here we can see in contrast to scenario 1 a difference in C/L path impacts training but it is not so simple. Granted, the C/L of the two players are different, let's ignore for a second what we see in the player profile page as C/L and consider fractional C/L. Player 3A had 4 seasons at 5/6 followed by 3 at 4/6 and is in his first season of 3/6. Player 3B had 4 seasons at 6/6 followed by 3 seasons at 5/6 and is in his second season at 4/6. Thus, the rate at which they're dropping C/L is very much the same, but their initial C/L is what differed. Let's look at it this way: player 3A and 3B's fractional C/L slopes are the same, but they have very different intercepts. Thus, while they're dropping fractional C/L at the same rate, the large difference in initial fractional C/L (remember both were 6/6 when pulled!) is what snowballed into this massive gap in training we see now.
Summary: While a change in C/L and a dramatic change in C/L path impacts training as we'd expect. Given the similarities in the rate at which C/L changed between these two players indicates the initial fractional C/L a player has at "birth" largely impacts his training down the road.
Scenario 4: Players of the different age but at the same C/L and both in their first season at this C/L.
|Oldest 6/6 Age||18||15|
|Attribute Quality Compared||84||84|
|Average Training (Month)||1.21||1.33|
|Standard Deviation in Training (Month)||0.04|| 0.06
Both players are in their first season at 5/6. 4A is on the longest career path for a player pulled at 6/6 while player 4B is on the shortest career path for a player pulled at 6/6. What is seen in this scenario is a drastically different training rate and perhaps the most surprising of them all. If all that dictated was the fractional C/L of a player, you would expect both these two players to train nearly identically since they're in their first season at 5/6 (4B will have a guaranteed 4 seasons at 5/6 since he dropped to 5/6 at age 16 while 4A will have 3-4 seasons at 5/6 since he was 18 and 6/6). As they're in their first season at 5/6, this would mean that their fractional C/Ls are nearly identical. However, we see a distinct difference in training with the younger player 4B training better than the older player 4A. Thus, age impacts training not just C/L and the C/L path (fractional C/L).
Summary: Age impacts training, plain and simple. If you see a 19yo freshly 5/6 C/L 500 OR on the market and a similar player who just dropped at age 16, the 16 year old will be the better long-term player even though the first player is on a better C/L path.
What I've discussed above is merely a quick snapshot into the partially-unknown world of player training. There are some things we can influence to improve training (staff, facilities etc.) however there are some things we cannot. While we may target keeping players that of desirable C/L path and while we may hunt for 6/6 players on the market, there is an element of unknown which I call fractional C/L (others call sub-levels) and can also be referred to as C/L path that we cannot see. However, looking at the above quick investigation and keeping this in mind, we can glimpse into this unknown world that the developers have only recently admitted exists.
I have every intention of doing a full-scale analysis on this data however I'll admit times are very busy right now. It is due to this unknown publishing date that I desired to get this brief high-level overview out to you to brush up on your knowledge as the hockey season transition is just around the corner. Until then, good luck in the new soccer season and the upcoming hockey playoffs.