[Discussion] (option to) randomise PA for database players on a normal distribution

[Gangor]

One problem with having such a well scouted database of players is that players generally turn out quite similarly over many games. This is managed to some extent by having random ranges for some players but often results are quite predictable. The same lower league gem from Norway will become amazing every time you buy him. Wouldn't it be great to be able to discover that one in a million 200 PA player currently playing in the National League South, or perhaps the Kazakhstani Messi?

I propose that in order to create this additional variety, PA should be randomised based on a normal distribution. I suggest that the scouted PA become the mean and a standard deviation be set to (PA-CA)/3. 

This would mean that while the PA for most players will be fairly close to scouted PA, it's likely there would be a certain number of gems to be found. Some duds would be generated too. 

Example - 100 CA, 130 (scouted) PA player. There would be a 68% chance for the PA to be between 120 and 140, 95% chance of it being between 110 and 150 and 99.7% chance of it being between 110 and 160. There would also be a 0.15% chance of a higher PA.

Example 2 - 50 CA, 110 (scouted) PA. 68% chance of PA being between 90 and 130, 95% between 70 and 150 and 99.7% between 50 and 170. There would be a 0.15% chance of a PA over 170.

I think this would add a nice bit of spice to early game scouting and perhaps change the destinies of some lower league teams. The reason I put "option to" in brackets is that I'm unsure what the processing overhead would be. It's possible this would make loading up a new game take longer (worst case) so some people with lower end PCs may prefer to do without it.

[Preveza]

Interesting philosophy, and it is an area SI will have to evaluate in time.

[Neil Brock]

Reviewed so locked.