Developing a Rating System for SOM Hoops
by Dave Little
This process entails three steps:
1) Analyzing the action deck
2) Assigning value to the events which take place
3) Analyzing the player cards - we'll do this as we go along.
Analyzing the action deck
I began this exercise by reviewing the events which occur in the action deck. The deck consists of 60 cards which are cycled through once per quarter. Within the Control section, thirty-five cards (~60%) refer the action to a player card, seven for each position. The other 25 (~40%) refer the action to a passing card. Further breakdowns are as follows:
The Fastbreak Control section also assigns shots and passes. The specific breakdown is as follows:
The resulting outcomes of fastbreaks are: Shots: RG 27%; LG 28%; RF 25%; LF 10%; C 10% Passes: RG 35%; LG 8%; RF 8% (not all fastbreaks refer to a passing card)
Defensive assignments on the fastbreak are assigned as follows:
Rebounding is handled differently by the computer than by the board game, so no breakdown of the rebound values presented by the action deck is included in this analysis. The press is rarely used (at least by me!) so press ratings are also not presented.
Assigning Value to Events
There are two fundamental units of value in analyzing events: possessions and points. I have attempted to assign each player card a value in terms of his ability to generate points and possessions for his team. In evaluating cards, I have attempted to rate each card on its own merit, without regard to the contributions of teammates, team defenses, etc.
Each player card contains 10 elements: shooting, 3-point shooting, fastbreak shooting, passing, fastbreak passing, one-on-one defense, X-column defense, fastbreak defense, shot-blocking, and rebounding. Each of these will be rated according to the value of points and possessions generated. Many players are capable of playing more than one position. It is possible, then, to have separate ratings for a player card at different positions. All ratings are based upon the assumption that a player will play 48 minutes per game. You might want to normalize these or consider minutes played as a factor when drafting.