I've been looking into the NBA draft logic and have made some changes to it.
The first step was the logic that ranks players. This formula includes player ratings (weighted by position), team current-season success, prestige, and national player awards. I'm making the following changes to it:
- As has been pointed out a few times, the logic had been favoring big men over guards. I've made some tweaks to the player rating component to balance that out a lot more. I am also now including all ratings (except durability) instead of just a few for each position. Ratings are still weighted differently by each position, but a fuller picture of the player's talent is now used. For example a good rebounding guard will get credit for that.
- The team success component was looking simply at wins instead of the more complete view of team success that we use most places, so I've made that change. Looking only at wins would likely inflate the status of players from mid-major teams who dominate their conference. Using the full picture of success (which includes postseason) will make that factor more meaningful.
- I've bumped up the weight of prestige factor as it was very small
- I've reduced the bonus bump for winning national POY as it was ridiculously high
The second part of the change is looking at the logic for early entry decision. Currently, there are a few restrictions in place, such as a freshman/sophomore won't leave early if not projected in the 1st round. As long as those restrictions are met, the decision was coming down the player's "Greed" rating, which is a relic of a feature of the game from years ago. Some vet coaches know that you can get an idea of that rating through certain recruiting responses. I really dislike that a hidden rating (which a lot of coaches probably have no idea exists) is so important. I've decided to remove that factor completely from the decision.
The new logic gives the player odds of leaving early based on the player's class and his projected pick. The higher they're projected, the more likely they are to leave. The old logic did not differentiate between a player projected to go #1 vs. #28. To me that's a big difference, so the new logic will reflect that. Sophomores and Juniors work similarly, but the odds of leaving at a given pick will be higher for each higher class. For example, a freshman might have a 70% chance of leaving if projected at #10, whereas a soph might have an 80% change and a junior a 90% chance. So there will still be some random chance that a player will come back to school. In my testing, the breakdown of draft players by class is in the same ballpark as it is now.
The end result of all of this is that the draft should make a lot more sense in how players are ranked and which players enter early. I believe that it will also help mid-majors hang onto players longer, but of course it still depends on how good the player is and how much success the school has had.