# Genie : Of course it depends on managing style and random. That's why it makes estimation "only" with 85% accuracy. But this value means that it will assess correctly 17 of 20 players.
And the main advantage of statistical model is that it's possible to assess its accuracy.
I think he answered your question with this part. With tutoring being one of the 'managing style factors' (as of course one cannot predict if or with whom a manager will tutor his player). Furthermore I am not sure, but I expect that he has checked every attribute that might affect the players development to see if there is any significant influence on if the player will reach his PA or not.
One concern I do share with you is the usefulness. I usually tend to improve my training facilities and coaches as soon as my funds and the board allow me to. Therefore my young players nearly always reach their PA.. But, of course, I can imagine situations in which one is simply incapable of having a world class training system. And in those situations it would be very useful to see if your new signing will actually reach his PA.
I have to say that I am quite impressed with the maths and statistics behind Genie Scout I am a first years econometrician student and I haven't had any statistical subjects yet, but I am anxious to use the model in FM.
A question I wanted to ask you for some time is how you manage to predict the 'future attributes' of players (on which the potential rating seems to be based) and what the accuracy of this prediction is. Also, it usually seems like when a player joins my club, his potential rating increases over time (with over 5 percent point with one player). Is this because I use a different training schedule?