The Miller Metric
Posted by Andy Larsen on March 30, 2009

BasketballDan commented on my Ronnie Brewer post last week, saying : “Larry Miller had a formula for rating the players by adding points, assists, blocks, rebounds subtracting turnovers and all. I don’t remember the exact formula (apparently). I would like to see how the team shakes out using that formula. Are you strong enough in stats to answer that? If you are I would be highly impressed.”
We, here at Jazzbots, rarely like to disappoint the reading audience. So, here’s the formula:
Miller Metric = Points + Rebounds + Steals + Blocked Shots + Assists – Turnovers – Shot Attempts – Personal Fouls.
If you’d like, you can divide it by Minutes Played, that’s usually how its presented in the paper. Here, though, I’m going to compute it per 48 minutes, because it allows us to compare it to other well known statistics.
What’s really nice is somebody else has computed these already for every player on the Jazz. Those nice people are at the Wages of Wins Journal, a statistically oriented NBA blog. Here are the results:
Position Adjusted Miller Metric (PAMM) = Miller Metric per 48 minutes – Average per 48 minute Miller Metric at Position Played
Win Score = PTS + REB + STL + ½*BLK + ½*AST – TO – FGA – ½*FTA – ½*PF
Position Adjusted Win Score (PAWS) = Win Score per 48 minutes – Average per 48 minute Win Score at Position Played
|
Utah |
Position |
Rank PAMM |
Miller Metric |
PAMM |
Rank PAWS |
WS48 |
PAWS |
|
Carlos Boozer |
PF |
1 |
19.6 |
6.6 |
1 |
16.6 |
6.2 |
|
Deron Williams |
PG |
2 |
18.0 |
6.4 |
5 |
8.7 |
2.2 |
|
Andrei Kirilenko |
PF |
3 |
18.9 |
6.0 |
6 |
12.6 |
2.1 |
|
Paul Millsap |
PF |
4 |
17.4 |
4.5 |
2 |
14.6 |
4.1 |
|
Ronnie Brewer |
SG |
5 |
12.8 |
3.3 |
3 |
8.8 |
2.7 |
|
Mehmet Okur |
C |
6 |
16.8 |
3.1 |
4 |
13.6 |
2.2 |
|
Kosta Koufos |
C |
7 |
12.2 |
-1.5 |
8 |
10.9 |
-0.5 |
|
Brevin Knight |
PG |
8 |
9.4 |
-2.2 |
11 |
5.1 |
-1.4 |
|
Kyle Korver |
SF |
9 |
8.3 |
-2.4 |
9 |
6.9 |
-1.0 |
|
Matt Harpring |
SF |
10 |
8.2 |
-2.5 |
7 |
7.7 |
-0.1 |
|
Kyrylo Fesenko |
C |
11 |
9.4 |
-4.3 |
12 |
8.9 |
-2.5 |
|
C.J. Miles |
SG |
12 |
5.2 |
-4.3 |
10 |
4.9 |
-1.2 |
|
Ronnie Price |
PG |
13 |
6.2 |
-5.4 |
13 |
3.5 |
-3.0 |
|
Morris Almond |
SG |
14 |
2.8 |
-6.7 |
14 |
2.3 |
-3.8 |
|
Jarron Collins |
C |
15 |
2.3 |
-11.4 |
15 |
3.4 |
-7.9 |
Both rating systems get the same top 6. I actually agree with Miller more than Win Score, I’m not sure I see the justification for docking assists but not rebounds. For the statistically minded, the correlation between the two scores is 0.92.
Now, I don’t think that the Miller Metric is more accurate or useful than something like PER, which actually runs a regression to see the true value of each statistic down to the decimal. But as a quick and dirty measurement of player value, it is surprisingly worthwhile because it recognizes the negative value of missed shots.
2 Comments
stats teacher on April 03, 2009 said:
Andy,
You should thank your mathematics teacher! Good job with the numbers!
Teach
Leave Your Comment
How do I get an avatar for my comments?
It is easy and completely free. Just go to Gravatar.com and sign-up with your email address. Gravatar is globally recognized, so your avatar will show on every Gravatar-enabled website or blog.
Fan Flickr Stream
Recent Comments
Bob on November 7, 2009 said:
"The players do park in the lot on the SW side of the Arena. The door they..."
Danny Hansen on November 7, 2009 said:
"Awesome article. As a Jazz fan I have always wanted Memo taking..."
Annie on November 7, 2009 said:
"Caleb I’m thinking the players park in the player’s parking..."
Annie on November 7, 2009 said:
"Very nice Trevor. I was just mentioning the same thing the other night..."
Popular Tags
Jazz Polls
Loading ...
Recent Readers
Monthly Archives
- November 2009 (23)
- October 2009 (75)
- September 2009 (69)
- August 2009 (35)
- July 2009 (31)
- June 2009 (69)
- May 2009 (31)
- April 2009 (66)
- March 2009 (59)
- February 2009 (68)
- January 2009 (64)
- December 2008 (68)





vs 






























BasketballDan on April 01, 2009 said:
Wow. I’m am most highly impressed. Tons of good info there to digest, ponder, and compare. I don’t think anyone should be surprised that 3 out of the top 4 rated players are PF’s in a system that directs its offense thru that position first. On the other hand all three of those players have been known to play a variety of positions at times. Thanks for the info on the formula, the Stat Junkie Web site, and the stats themselves.