What if your team had CPR for the 2012 draft?

By Stephen Shea, Ph.D.


It’s June 28, 2012. NBA teams have gone through months of reviewing game film, conducting workouts, and holding interviews in preparation of this evening’s draft.

Anthony Davis was the prize of the draft class, and it was no surprise when he went first overall to New Orleans.

Michael Kidd-Gilchrist went second to Charlotte. Washington drafted Bradley Beal.

Now, Cleveland is on the clock.

It’s not a great time to be a Cleveland fan. LeBron has taken his talents south, and the Cavs haven’t won more than 21 games in either of the previous two seasons.

However, there is some hope. Cleveland used the first overall pick a season ago on Kyrie Irving. As a 19-year-old rookie, Irving averaged 18.5 points and 5.4 assists per game. He also shot 40% on 3s.

Cleveland’s other lottery pick from a season ago, Tristan Thompson, also looks to have some potential.

In the 2012 draft, Cleveland is looking to build on its young core. The 4th overall selection presents a great opportunity to add another piece.

Cleveland’s General Manager pulls out the College Prospect Rankings (CPR), which he uses as information in the draft preparations.

According to CPR, Anthony Davis was by far the best in the class. AD’s CPR of 24 is the best since Durant. After that, the highest rating in the 2012 class is a 7.

CPR doesn’t suggest a clear 2nd choice. Instead, there is a group of 8 prospects in a second tier that rate between 4.8 and 7. Two of those 8 are off the board (Kidd-Gilchrist and Beal), leaving 6 for Grant to mull over.

CPR isn’t a perfect system. It doesn’t have the advantage of considering medical records, performance in individual workouts, interviews with the candidates, or really any form of information that isn’t recorded in the box score of a college game. Thus, Grant knows he shouldn’t just pick down the list.

Instead, Cleveland’s GM uses the list as guidance.

The top 6 in CPR among the prospects still on the board is Maurice Harkless, Will Barton, Tyler Zeller, Terrence Ross, Draymond Green, and Damian Lillard.

The 4th overall pick isn’t Cleveland’s only pick of the night. They will pick again at 24, 33 and 34. Although CPR is making a strong case for Draymond Green and Will Barton as late lottery picks, Grant knows that most (if not all) of the NBA views these guys as 2nd rounders. So, he wisely decides to consider them later.

CPR is a simple formula to dissect. It’s easy to determine how certain players generate high scores. Zeller’s high score comes in large part from dominating the glass against smaller competition in college. There is some concern that he won’t be so dominant in this regard against much bigger and more athletic pros. Zeller might be a solid choice in picks 10-20, but he’s not worth the 4th overall pick.

It’s down to Harkless, Ross and Lillard. Grant and his team of scouts ranked Lillard the highest among that group, but there’s a wild card to be considered as well.

The team also liked Dion Waiters. Waiters is a more natural fit as a shooting guard next to Irving. He worked out and interviewed well with the team. There is a part of Grant that wants to go that way with the pick. He ponders it for a moment.

Waiters has a CPR of 1.84, a number that suggests he’s a late first rounder (or worse). The numbers don’t agree with the scouts. The numbers are screaming Waiters isn’t worth the pick.

In the other direction, the scouts don’t see Harkless or Ross (or any of the others in the group of 6 as worthy of the 4th pick).

The scouts and the numbers agree on only one option, Lillard. Lillard isn’t an obvious fit next to Irving, but having two dynamic ball handlers and shooters in the backcourt isn’t the worst option.

When early in the rebuilding stages, teams can’t be too focused on positional fits. A lot can happen as players develop (or don’t) and as trade and free agent options materialize.

Grant decides to take the player that the scouts and analytics agree on. He takes Lillard.

…19 picks later…

It’s the 24th pick and, Ross, Harkless, and Zeller are gone. (Waiters is off the board too.) But as Grant predicted, Barton and Draymond are still on the board.

Barton (CPR=6.2) and Draymond (CPR=5.0) are the top two available players in CPR. (Jae Crowder is 3rd, and Grant can’t immediately dismiss him from consideration with the pick either.)

Barton and Draymond were both in the top 8 for the class in CPR. A CPR of 5.0 or greater suggests the type of potential that should be drafted in the lottery.

Grant was still wondering if he should have taken Waiters. He’d like to grab a shooting guard that fits naturally with Irving in this draft. Now, he’s thrilled to see Barton still on the board.

Grant and his team had Waiters rated above Barton, but the drop-off wasn’t huge. CPR prefers Barton by a wide margin. Putting the information from the analytics and the scouts together, there isn’t much of a difference between Waiters and Barton.

Grant happily takes Will Barton with the 24th overall pick.

9 picks pass too quickly for the busy Cleveland team. They’re up for back-to-back picks (33rd and 34th overall).

By this point in the draft, the scouts’ favorites are gone. Grant is well aware that history suggests most 2nd round picks won’t pan out. It’s the perfect time to try a new strategy. He turns to the analytics.

Historically, the odds of a 2nd round pick making a huge impact are slim. But, historically, teams didn’t use analytics.

Among the available prospects, the top two in CPR are Draymond Green (CPR=5.0) and Jae Crowder (CPR=4.7). In fact, these players were 8th and 10th overall in the class.

In this draft and the past, Cleveland has assembled some promising young guards: Irving, Lillard and Barton. They’ve also got a young big in Thompson. Crowder and Green are both forwards. They fill an obvious hole in the current young Cleveland core.

There are some questions about Green and Crowder. They were both impressive college players, but they fall into the “tweener” category. They don’t fit the mold of any traditional NBA position, profiling somewhere between the traditional power forward and small forward. They are smaller than traditional power forwards, but don’t quite have the game best suited for traditional perimeter players.

Fortunately, Grant understands that “tweener” can mean “versatile.” He understands that positional versatility on defense allows his team to switch screens (which is a major part of the modern NBA offense.)

Grant goes with CPR and uses the 33rd and 34th overall selections on Draymond Green and Jae Crowder.

Grant didn’t make any of his choices based on CPR alone. He never went against the scouts. He used the analytics to supplement all of the other information he had. In total he was able to grab 4 guys that scouts liked and 4 guys that the analytics liked.

All 4 players were in the top 10 in CPR for the draft class.

Cleveland welcomes their new additions: Damian Lillard, Will Barton, Draymond Green, and Jae Crowder.

Only time will tell if these were the right selections.

Ok, that’s not exactly how things went.

Cleveland did draft Crowder with the 34th overall selection. Actually, it’s more accurate to say they drafted Crowder for Dallas with the 34th overall selection. Crowder was packaged with their 24th (Jared Cunningham) and 33rd (Bernard James) in a trade for Zeller (and Kelenna Azubuike).

Zeller had a high CPR rating, but his rating was below that of Barton, who Cleveland could have drafted with the 24th pick. CPR didn’t like this trade.

The biggest mistake for Cleveland was the Waiters pick. CPR preferred every player drafted after Waiters and before Fab Melo, who was a terrible pick at 22. The team could have gone big with Drummond, or found a wing with length in Harrison Barnes or Terrence Ross. The combination of the scouts and the analytics suggested Lillard. Did Cleveland go a different way because they were concerned about pairing Lillard and Irving?

CPR doesn’t come close to telling a team everything they need to know about a prospect. As far as analytics go, CPR is just scraping the surface of possible contributions to a team’s draft preparations. CPR doesn’t make the decision, it’s just another piece of information for the decision makers of the organization to consider.

That said, if Cleveland’s GM had turned the draft over to CPR after the Waiters pick, the Cavs would have drafted Barton, Draymond and Crowder 24th, 33rd, and 34th overall.

The table below displays the CPR scores for all college players drafted in 2012.


Author’s Note: For this article, CPR is as calculated in Basketball Analytics: Spatial Tracking. The simple formula presented there does not adjust for injury-shortened seasons. Khris Middleton tore his meniscus in his junior-season opener. He missed significant time due to that injury and struggled to find his form when he played. As a result, he registered a CPR of 0.74. After his sophomore season, Middleton had a CPR of 3.05, which would have ranked 21st in the 2012 draft class. That score would have rated him above guards like Waiters, John Jenkins, Marquis Teague, and Austin Rivers.


2012 Draft CPR

1Anthony Davis24.1
15Maurice Harkless7.0
40Will Barton6.2
17Tyler Zeller6.1
3Bradley Beal5.9
2Michael Kidd-Gilchrist5.3
8Terrence Ross5.0
35Draymond Green5.0
6Damian Lillard4.8
34Jae Crowder4.7
25Tony Wroten4.3
11Meyers Leonard4.2
38Quincy Miller4.1
19Andrew Nicholson3.8
16Royce White3.7
21Jared Sullinger3.7
5Thomas Robinson3.6
9Andre Drummond3.6
12Jeremy Lamb3.6
36Orlando Johnson3.1
58Robbie Hummel2.9
24Jared Cunningham2.9
7Harrison Barnes2.8
18Terrence Jones2.7
59Marcus Denmon2.7
50Kyle O'Quinn2.6
28Perry Jones2.6
10Austin Rivers2.5
47Kevin Murphy2.4
14John Henson2.3
29Marquis Teague2.3
13Kendall Marshall2.2
45Justin Hamilton2.1
31Jeffrey Taylor2.0
23John Jenkins1.9
4Dion Waiters1.8
22Fab Melo1.8
27Arnett Moultrie1.8
51Kris Joseph1.7
42Doron Lamb1.6
41Tyshawn Taylor1.5
55Darius Johnson-Odom1.5
44Kim English1.5
46Darius Miller1.1
43Mike Scott1.1
33Bernard James1.1
37Quincy Acy1.1
60Robert Sacre1.0
26Miles Plumlee0.8
30Festus Ezeli0.8
39Khris Middleton0.7

Leave a Reply

Your email address will not be published.