I will give you a funny example: Do you remember Karate Kid? "Togli la cera, etc..."? And the kid learned new "skill"?
Why did I not become a karate champion when my mom forced me to make the bed?Oh,I'm so unlucky....
this cross-training is a bad thing.A user work to find on the market and plan the training to grow players with the best salary/performance ratio,and then the game oddly train useless skill that raise tha salary of these players ruining this work????
If those players are truly the best salary/performance ratio, an extra pop in an undesired skill will likely hurt less than on a guy who is already on the "wrong path". I am always impressed by the ability of users to simultaneously ignore that these changes do not occur in a vacuum and to assume that any small change to their "optimal" plan is essentially "ruining" it.
The pop in undesired skill however hurt,so you don't have any point here.A training system that lead you to losing time in the training of a player(primary skill) and raise his salary in an undesired way(because of random secondary skill) is bad
But you said yourself that you're picking players who have the best "salary/performance" ratio. That means that your random pop in rebounding on a guard will cause less of a salary increase than someone else who has a guard with a worse ratio, since that pop will increase the salary a greater amount.
I forgot who it was in another forum (one of the GMs, I believe) so I can't credit them for this insight, but it's a good perspective: each week you're getting 10% of a normal training in one of the 9 skills that aren't being trained. On average, something that takes two full training weeks to pop would, therefore, be expected to take 180 training weeks to pop. Obviously, sublevels and training speed will factor in, but to speak as if training has been destroyed by this is making a mountain out of a pebble.