A different metric (applied to PGs)

Abstract

I made a new metric that compares players by position by taking that their per minute pace adjusted stats, averages them, and then compares each player to the average and then sums each players difference from the average for all relevant categories. I used two different methods for weighting shooting percentages. The first does no weighting and the second weights by pace adjusted per minute shots attempted.  With no  weight on shooting %s I found the top 5 pgs to be Chris Paul, Jose Calderon, Jason Kidd, Chauncey Billups, and Baron Davis. When I weighted shooting %s the top 5become Chris Paul, Jason Kidd, Baron Davis, Jose Calderon, and TJ Ford.

Why?

What motivates an idiot to try and undertake making yet another basketball metric? I was recently in a bit of a discussion with Stumptown Doug about his post PG Grading and this finally got me off my duff to try my hand at what I think might be an interesting metric for ranking players relative to their position. In Stumptown Doug's analysis scoring pgs will have an immediate helping hand as points are weighted much greater than anything else. For example a 10 apg (assists per game) 10 ppg (point per game) pg would be the same as a 15 ppg 5 ppg per game pg, but I think the 10apg pg is more valuable and any good metric should reflect this. The first that might be done is to just multiply each assist by 2, which is reasonable, but then all stats would need to be weighted by how many points they cause the team to score; I dont know how to do this for things like steals, blocks and rebounds. Instead I just wanted to weight all categories equally by seeing how close each pg is the the average for every category and then sum how far they are from the average in all relevant categories: 2 point field goal percentage, 3 point field goal percentage, free throw percentage, rebounds, assists, steals, turnovers, and points. In one instance I tried to weight the shooting categories by shots attempted so that good shooters who shot a lot were rewarded more and bad shooter shooters who shoot a lot were further penalized for it.

Method

First all raw, non percentage stats, were scaled by minutes and then by pace. I wanted to do this so big minute guys were on the same footage as not so big minute guys, it also adjusted for slow paced teams as scoring 20 on the Suns is much harder than scoring 20 with the Blazers and this needs to be accounted for. Then for each category I computed the average and then divided each players adjusted stat by the average and subtracted 1. So if the player was better than average in that category the player received a positive number if worse they received a negative number with the opposite being true for turnovers. I then summed each players difference from the average over all categories. I did this for no weighting of shooting categories by shots attempted and with weighting by shots attempted.

Results

 weighted shooting stats (top 25)

 Chris Paul     2.77
 Jason Kidd     1.69
 Baron Davis     1.31
 Jose Calderon     1.13
 T.J. Ford     1.02
 Chauncey Billups     1.00
 Brandon Roy     0.90
 Brevin Knight     0.69
 Gilbert Arenas     0.67
 Andre Miller     0.61
 Eddie Gill     0.47
 Deron Williams     0.41
 Devin Harris     0.40
 Bobby Jackson     0.40
 Jameer Nelson     0.39
 Tony Parker     0.38
 Jamaal Tinsley     0.34
 Anthony Carter     0.34
 Steve Nash     0.32
 Rafer Alston     0.22
 Mo Williams     0.22
 Sam Cassell     0.20
 Mike Bibby     0.16
 Dan Dickau     0.14
 Darrell Armstrong     0.13

no weight on shooting stats (top 25)

   Chris Paul    3.20
 Jose Calderon     1.86
 Jason Kidd     1.84
 Chauncey Billups     1.46
 Baron Davis     1.38
 T.J. Ford     1.16
 Brandon Roy     1.08
 Steve Nash     1.06
 Eddie Gill     0.93
 Jameer Nelson     0.82
 Deron Williams     0.81
 Mo Williams     0.69
 Devin Harris     0.67
 Gilbert Arenas     0.64
 Bobby Jackson     0.57
 Derek Fisher     0.51
 Anthony Carter     0.49
 Tony Parker     0.38
 Dan Dickau     0.32
 Mike Bibby     0.30
 Mike Wilks     0.30
 Anthony Johnson     0.29
 Jason Hart     0.27
 Sam Cassell     0.27
 Jason Williams     0.24

Discussion

It is hard to find a metric where Chris Paul is not awesome. Also Deron Williams and Steve Nash, elite pgs, did not fare to well under this analysis.

Wrap It Up!

This was a labor of love and sometimes hate as I am not a big spreadsheet guy. If you have any questions or concerns I would love to hear them. If you want a copy of my spreadsheet email me at jonestr at unm dot edu. Please let me know what you think of this and if you would like to see anymore of this for another position. Thanks for reading this. Also I got all of my stats from player stats and team stats.