The implications of a player’s College Win Score

This weekend I took some old calculations I had sitting around, regarding about 107 past NBA player’s College Win Score per 48 and their Career NBA Win Score per 48 and ran a regression analysis to try to predict the offensive upside/downside of each of the draftee’s from last week’s confusing NBA Draft (how many “proxy picks” can you have in one round?).  In this post I discuss what I found and some implications I think I can make, and in the next post I will state the high and low range for the career numbers of each player chosen in this month’s draft.

The Equation

Here is the result I got.

College to NBA Win Score Equation:
Y= 2.45 + .43X – 0.16, standard deviation: (+/- 1.84)
Where Y= the player’s career NBA Win Score average, and X= the player’s career NCAA Win Score average.

Here are the links to my self-made data lists in case you wanted to run with your own analysis and call mine “dumb”.  (List One; List Two; List Three; List Four).  They’re kind of mishmash lists, because I would start one list and then I would put it to bed and two months later start another.  In List Four I actually evaluated an entire first round (The Chris Webber Draft), and the third list features a lot of stars from the 1990s.

I was a bit limited by a couple of things.  One, for some odd reason, Basketball-Reference does not list complete statistics for any college player from the 2000s.  Second, if you go back to far into the 80s, the college statistics are bafflingly unreliable (for some player’s seasons, the minutes aren’t available).  So the sweet spot for college statistics are players who played most of their college ball in the mid 1980s to mid 2000s.

After I ran those numbers, then last night I double checked my work by randomly selecting players from the “Ws” in the’s player’s listings.  I got virtually the same result.

There is a basic implication from my equation that I should have seen intuitively but did not.  The production numbers for big men fall further than they do for the wing and backcourt players. 

This only makes sense.  There are simply not that many adequate big men.  Certainly not enough to spread throughout the college ranks.  Therefore a decent big man should be expected to dominate to a greater extent.

The second implication comes indirectly from Professor Berri’s post of a few days ago.  In it he posted the Win Score per 40 averages for every player chosen over the past 15 or so seasons and came up with this (I translate them to Win Score per 48 by dividing by .83):

PG:   8.91
SG:   10.12
SF:   11.99
PF:  15.17
C:    14.84

Now, run those averages through my formula, and compare the average draftee’s production to the NBA positional average from last season and you get:

SG: (-0.15)
SF: (-0.15)
PG: (-0.77)
PF: (-1.48)
C: (-3.63)

This does not figure in defense, but it comports exactly with our “Wins Above Bench” value stratification.  Meaning, there are a lot of decent win producers at the wing positions, less so at point guard, less than that at power forward, and the center position is a desert.   That’s exactly what we found.  (Others who have made Wins Above Replacement calculations have failed to take into account the radically different values at the different positions and I think that’s a mistake).

Before my final implication, a quick comment about defense.  To the extent my “defensive win score” produces valid results, or that it is a valid concept at all, I have tended to find that defensive value in basketball (meaning the ability to hinder your opponent’s productivity) is almost identical to defensive value in baseball.  Meaning, it is important to be sure, but it rarely turns a dud into a stud.  When it comes to win power that really separates the great win producers from the merely common win producers, that kind of win power comes mainly from personal statistics, not from oppressing opponent statistics. 

I’ll have more.  I have to cut this short.  Tonight or tomorrow I’ll discuss implications for this season’s rookies, and I’ll try to work defense into the equation, using the statistics compiled from the one season when I specifically separated offensive wins and defensive wins.


One Response to “The implications of a player’s College Win Score”

  1. Jerbil Says:

    “Y= 2.45 + .43X – 0.16” -Is this a misprint? Why the 2 constant terms?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: