I'll go with Yoda's format...
(seemingly) easy:
1. Modify LI offence effectiveness
2. Modify 2-3 zone effectiveness
3. Modify effects of Game Shape - Keep it random but a little less... Include long term effects. For example, if a player plays (and trains - see below) more than 120mins a week for several (4+?) weeks, his GS is severely affected and may take longer to normalise. Or if a player plays excellent (definition of excellent needed) minutes for several weeks, his GS can improve even further than proficient.
(seemingly) difficult:
1. Modify training process - There is no reason that a player should only get "training time" by playing in games. Perhaps each player has a default 30mins (for example) training time and in-game time is added to those minutes. I actually play basketball and the times I improve the most are during actual training. It's a time when I can try new things and repeat skills without risking the W.
2. An individualised and deeper approach to tactics. For example, selecting two tactics for a game, a primary and a secondary. Or an option to choose if opponent plays x defensive tactic, we play y offensive tactic. Difficult, I know, but it would be good.
3. New tactics that give more benefits to particular stats.
4. Modifying the GE so that there is an advantage to scouting opponents and learning tactical advantages. Base tactics would need to be less effective and specialised tactics more effective (unless opposing team's tactics trump them). For example, RNG vs 2-3 would be more effective than it already is. LI vs 1-3-1 would be more effective.
5. Somehow make effective the tactics that are seldom used (princeton, LP).