In real life basketball you have players that have breakout years and some that improve their play immensely during the year or regress. With Rating Fluctuation players ratings will change per month instead of per year. If a guy is playing better than expected then his rating could go from 70 to a 72 or 73. If he's underachieving it could go down to a 68 or 67. A player wouldn't regularly regress or progress by 3 or 4 points each month, as 3 would be the max. Along with that, if a player is rated 70 and is playing at a 67ovr level of play, then he would regress to that. His stats wouldn't go down because of the rating change because his rating would affect the level he is playing at. A player would have to play consistantly terrible all year to lose by 10 points or more, which would be extremely rare. This would not affect morale issues (such as a player being down 6 points because of him being unhappy. If a player is regularly a 68 and is down 6 points to a 62 and plays at a 59ovr level, then his regular rating would go to a 65 and he would be down 6 points to a 59. The computer would do all this mathematically.
-Most Improved Player award could then be given to the player that gained the most on his ovr during that year.