For a while now, I have been intrigued with "single number" player ratings. Whether it be PER, Win Shares, Plus Minus, or Wins Produced, I think such player rating systems are fascinating. I have been doing research on all of these models for quite some time, and have devised my own player rating metric for the world to see. I will be posting the numbers on a separate area of the site, but for now, I will discuss the general components of the formula.
I believe, as do many in the advanced stats community, that the best way to quantify basketball is through possessions. Each team has a finite number of possessions per game, and whoever uses them the most effectively, will win every time. Keep in mind that my metric, just as every other metric, hardly accounts for defense. This is for two reasons.
1.) No available metric adequately measures defense.
2.) The best of them don't have data available in bulk.
It should be said that a large part of my formula was based off the Wins Produced model, made popular by David Berri. There are a few noteworthy changes I made though.
1.) Assisted shot credit is divided between the passer and scorer.
2.) Offensive rebounds are slightly devalued.
3.) Missed shots count as missed shots, not turnovers.
I included positional adjustments, to help with the fact that different types of players have different roles. The metric is also measured per 40 minutes, so playing time isn't a(n) (dis)advantage.
To qualify for the first round of data, a player needs to have played in half of the available games (20 this year), and play at least 25 minutes per game.
My first data set will be posted under the HDR Player Ratings tab, and include the top 10 players at every position, as well as the top 10 players overall. Let me know what you think.
I believe, as do many in the advanced stats community, that the best way to quantify basketball is through possessions. Each team has a finite number of possessions per game, and whoever uses them the most effectively, will win every time. Keep in mind that my metric, just as every other metric, hardly accounts for defense. This is for two reasons.
1.) No available metric adequately measures defense.
2.) The best of them don't have data available in bulk.
It should be said that a large part of my formula was based off the Wins Produced model, made popular by David Berri. There are a few noteworthy changes I made though.
1.) Assisted shot credit is divided between the passer and scorer.
2.) Offensive rebounds are slightly devalued.
3.) Missed shots count as missed shots, not turnovers.
I included positional adjustments, to help with the fact that different types of players have different roles. The metric is also measured per 40 minutes, so playing time isn't a(n) (dis)advantage.
To qualify for the first round of data, a player needs to have played in half of the available games (20 this year), and play at least 25 minutes per game.
My first data set will be posted under the HDR Player Ratings tab, and include the top 10 players at every position, as well as the top 10 players overall. Let me know what you think.
