| Forum » General » Training new table-October 2011 - | Date | |
|---|---|---|
|
Username
715 msgs.
MVP of the game
|
thanks  |
21/10/2011 19:16 |
| - Div/Gr | ||
|
Username
5205 msgs.
Golden Ball
|
I highly recommend ignoring this table. It doesn't tell you any useful information except which positions are trained and approximately how much overall average will change. However, it is so approximate as to be useless. If your high-progressing player has high stats, they will improve far less (as much as 1/4) the value that low stats players will improve, and if your players have some high and other not as high stats, then the improvement is something in between. That means you can't use these numbers to do anything useful, especially since stats are "the only thing that matters in the actual game™." Until we can nail down EXACTLY how stats improve with each training in a lab environment (i.e. someone with a school charts the training amounts of a really-low-stats-high-progression player in each position for each training that affects that position), and then figure out how they drop as the stats improve, it will be impossible to use these numbers to reverse engineer progression, even to an approximate number. This quest could take all season, so don't hold your breath. |
23/10/2011 00:18 |
| - Div/Gr | ||
|
Username
3363 msgs.
Best scorer
|
^^ Exactly right^^ Its a complex non-linear reverse exponential formula used that will take time to figure out. Progression % from players in schools are exact. This is then known along with motivation, coach % and training % (once this is accurately obtained from the training table or from the developers) The player is then promoted into the youth team and then trained a particular suitable training like "shots on goal" for a CF - the formula can then be obtained but only for the first starting point on the reverse exponential curve. The subsequent curve gradient variable can only then be obtained from multiple points on that curve by the same player as he progresses. So a fictitious CF arrives at my Netherlands school with 87% progression (I wish! ) My 14 yr old CF in the above example trains "Shots on goal" Progression: 87.00 % (known) Motivation: 107 % (known) Youth Coach: 99% (known) Training: 47% - from the table at the start of this thread - this is by no means a guaranteed number as @Cello has stated. So he trains 0.85 for finishing and 0.38 for technique for example with an overall average increase of 0.45 We now have a starting point to calculate his training from progression or viceversa. The very next day he trains the same training and: he trains 0.84 for finishing and 0.37 for technique for example with an overall average increase of 0.446 As you can see his training will start slowing immediately he starts his journey along his forecast curve. Hopefully then the clever mathematician managers amongst us can give us the final formula. Feel free to pick apart my post - thats just the way I see it at the moment. Oh and "reverse exponential" is just my name for the forecast/progression curve - I'm sure there is a proper mathematical term for it  |
23/10/2011 14:22 |
| - Div/Gr | ||
|
Username
5205 msgs.
Golden Ball
|
logarithmic? | 23/10/2011 14:52 |
| - Div/Gr | ||
|
59 msgs.
Rookie
|
I would expect the curve of improvement verses practice to have an inverse exponential shape. Whether that proves to be the case is another matter. In any case a small tweak in any parameters in the equation will give significant differences in how the curve would look. So as it seems the training values are still quite fluid, any attempt to work out progression currently would probably be a waste of time. |
23/10/2011 16:02 |
| - Div/Gr | ||
|
301 msgs.
First-team player
|
why cannot i schedule more than 7 days of training?... it's a bug? | 23/10/2011 20:00 |
| - Div/Gr | ||
|
Username
5205 msgs.
Golden Ball
|
yes, it's a bug | 23/10/2011 20:22 |
| - Div/Gr | ||
|
Username
715 msgs.
MVP of the game
|
There has to be some sort of formula of a few different variables. Unless they've introduced some sort of random variable. I'm not even going to take a crack at it. I had to choose between a Physics and Math degree during my Sr. Year of University (I got kicked out of Engineering and had a bunch of Math and Physics courses stored up). I chose Physics for a reason Have fun with it (seriously!). They should offer a GB prize to whoever can figure it out  I hope the developers just give us the formula. |
23/10/2011 20:55 |
| - Div/Gr | ||
|
Username
5569 msgs.
Golden Ball
|
If you click on "FORECAST" and pay your 10GB, then plot the curve on paper. Then draw tangential lines on each part of the curve, taking the slope of each line, you can get a good idea of the slope at any given point. Now, the derivative of those slopes, decreasing over time with respect to average, gives you the rate of change of that slope over time. Plugging that into a base number, then multiplying it out should give you an equation to work with. If you take the integral of that equation, you can figure out the total amount trained as an area of the graph as a whole. Then you will know what they saw when they opened the suitcase in Pulp Fiction. |
23/10/2011 23:46 |
| - Div/Gr | ||
|
Username
3363 msgs.
Best scorer
|
nobody important said: I would expect the curve of improvement verses practice to have an inverse exponential shape. Whether that proves to be the case is another matter. In any case a small tweak in any parameters in the equation will give significant differences in how the curve would look. So as it seems the training values are still quite fluid, any attempt to work out progression currently would probably be a waste of time. Inverse Exponential - exactly right! Tweaking is done! Time to turn that name around mister and crack the formula |
23/10/2011 23:47 |
| - Div/Gr | ||
2
