Draft Projecting at the All-Star Break

Lots of draft talk lately. A depressing reminder that this season is turning out to be much like the last decade of Wolves seasons than what many including myself expected. That said, old habits die hard and I am eager to start hitting the draft pipe again.

I have a fun new projection toy that I really like.

The model was built using NCAA to NBA data beginning with the 2002/03 college season to find variables and parameters that best predict "Wins generated across rookie contract" (i.e. Win Shares production during a player’s first four years). The method is nice and unbiased and the out of sample predictions are pretty good. There will always be misses, but I feel pretty comfortable outsourcing my draft opinions to the model.

For those who can interpret here is the model’s standardized Beta coefficients:

Intercept 0.27
Pnts 0.90
Position -0.52
2pnt att -0.57
3pnt att -0.53
Asts 0.46
Orbs 0.35
Age -0.30
SOS -0.07
Tos -0.22
Stls 0.18
Blks 0.25
FTA -0.09
Total Min -0.10
Weight 0.21
Height 0.36
pnt*pos -0.26
2att*pos 0.17
3att*pos 0.37
ast*pos 0.08
orb*pos -0.06
age*pos 0.10
SOS*pos 0.17
blk*pos -0.11
fta*pos 0.14
wgt*hgt -0.18
wgt*pos 0.20
hgt*pos 0.24
h*w*pos -0.09

The college season isn't over yet, but most players have either established or failed to establish themselves by this point. It is time to start taking projections of college numbers seriously, so I wanted to put the numbers for some of the more popular (and less popular but underrated) college prospects into context.

Below I am going to post the position by position draft projections generated by my model for the last 3 seasons (the 3 out of sample years for the model) and plug this years prospects into where they fit next to recent years' prospects.

"Wins" is the number of wins each player is expected to generate across a four-year rookie contract. The yellow background indicates the level at which I recommend drafting a player (based on average production at different slots). The darkest yellow is a #1 pick, next is a top 5, next is top 10, next is 1st round, while red backgrounds I recommend not drafting at all.

First up is point guards:

Phbkqxu_medium

Assuming Smart actually is a point guard, this is looking like a strong class. Nobody who can hold a candle to Irving, but that is true at every position over the past 4 draft classes. Smart has a definite shot at being a star. MCW looks like a top pick and both Burke and Wolters deserve a paycheck.

Gybtsc5_medium

Smart's classification matters here, but even without him this is a solid class. McLemore isn't deserving of a #1 pick and barely even a top 5, but he is absolutely a legitimate prospect. Oladipo is absolutely worth consideration, and Shabazz would look okay if it wasn't for the fact that he is going to go well above where he should.

L8srpgs_medium

All aboard the Porter wagon? DX has LeBryan Nash going at #25. I think some team may be disappointed.

Bugpwlm_medium

Especially given the PF strength in the league right now, this is not the position to target in 2013. That said.. Muscala should be a steal as a backup 4 in the early second or late 1st round. Bennett is worth a gamble if he slips, but he probably won't.

At3zr8q_medium

I was wrong on Drummond, Sanders, Whiteside, and Cousins, maybe I will be wrong on Len or Austin! This model seems to have struggled the most at the center position in the last few years, but I still don't think we should get excited about relatively unproductive college bigs.

So... there you have it. Looks like a solid draft especially on the wings where the Wolves need the most help.







X
Log In Sign Up

forgot?
Log In Sign Up

Forgot password?

We'll email you a reset link.

If you signed up using a 3rd party account like Facebook or Twitter, please login with it instead.

Forgot password?

Try another email?

Almost done,

Join Canis Hoopus

You must be a member of Canis Hoopus to participate.

We have our own Community Guidelines at Canis Hoopus. You should read them.

Join Canis Hoopus

You must be a member of Canis Hoopus to participate.

We have our own Community Guidelines at Canis Hoopus. You should read them.

Spinner.vc97ec6e

Authenticating

Great!

Choose an available username to complete sign up.

In order to provide our users with a better overall experience, we ask for more information from Facebook when using it to login so that we can learn more about our audience and provide you with the best possible experience. We do not store specific user data and the sharing of it is not required to login with Facebook.

tracking_pixel_9347_tracker