Moreover, I think the answer to the "divine trick" problem is that the game engine is always based upon a comparison between offensive and defensive players. The difference between, say, 25 and 26 in JS is going to cost salary but is going to have almost no effect on gameplay unless you are trying to cope with defenders of comparable skill. The real answer is that training somebody to be vastly superior to an opponent simply isn't worth that much. So, sure, you can add a single-skilled player cheaply who will beat an opponent in a single skill but lose in most other skills. But that's rarely a useful thing to do as a general plan, because I doubt that 5 players who are better in just one facet of the game than their opponents would be a very successful team. Look at how those USA teams did that picked 12 scorers (or 11 scorers and a token college guy to sit on the bench).
Monoskilled players are basically useful against other monoskilled players. Whether they are valuable, therefore, really depends upon your opponents. But it's almost certainly the case that if you train a player in a skill where they already dominate their opponents, you will be training them in a way that adds salary faster than you are adding effectiveness. As in real life, players and teams generally gain more rapidly by working on their weaknesses than by working on their strengths.
Well for my experience this is true, not sure 100% but let me say this is true, in the 2 positions hardest to train, SF and PF but for example in the position of center, I've a big doubt that a player with lvl 12 in ID, IS and reb + lvl 7 in all the other skills can stop a monster with lvl 20 in IS, ID and reb but with lvl 1 in all the other skills.
Simply I think that a lvl 14 in the 3 skills can be "competitive" against a lvl 20 in the same skills with half/week salary
1990-2022 Stalinorgel - https://www.youtube.com/watch?v=pV-Xppl6h8Et