Since the Blazers season is over, I thought I'd have a little fun with some Blazers draft history data. I downloaded all of the draft data back to 1970 (For those who don't know, that the starting year for the Blazers.). Some questions I wanted to answer: - How do the Blazers' draft selections stack up against other teams'? - Who was the Blazers best value draft pick? - Who was the Blazers worst value draft pick? - Who was the best Blazers GM? Keep in mind, drafting is different than putting a good team together. Often a very good player is drafted by the Blazers, only to be quickly cut or traded away. A prime example is Anthony Mason. Drafted by the Blazers in 1988 with the 58th pick. Cut by the Blazers soon afterwards. Becomes an all-star in 2001. Good drafting. Bad cut. It's difficult to measure the quality of a draft pick. There's always the balance of short term greatness versus long term productivity. Walton is a good example. Great for a few years, but didn't have longevity. After playing with the numbers for a while, I determined Win-Share to be the best metric for measuring the value of a draft pick. Every metric has it's drawbacks, but in the end, with so much data, Win-Share does a pretty good job at stack ranking the players. Here's a look at the average career Win-Share for each draft pick over the past 50 years: What this means is that a team selecting 11th historically selects a player that will generate 29.0 Win-Share throughout their career. A team selecting 20th historically selects a player that will generate 15.1 Win-Share throughout their career. Now armed with this data, I can compare if a team historically is drafting more talent than expected, or less talent than expected. I can also gauge if a player is 'living up to expectations' of their draft slot. Over the next few posts I'll use this data to answer the questions I stated above. And yes, I'll have some graphs. Lets have some fun with these numbers!
Okay, I first compared the amount of Win-Share that each team drafted over a period of time, and divided it by the average (or expected) Win-Share. So for example, the Blazers drafted Josh McRoberts in 2007 with the 37th pick. From our chart above, we would expect the player in the 37th pick to be able to generate 6.4 Win-Share during their career. Josh McRoberts has generated 18.5 Win-Share. 18.5/6.4=2.89. That's almost 3 times the production that the average player produces when drafted with the 37th pick. So I add all of these picks up for every team, and this is the values I get for each team: Not a surprise, San Antonio is at the top of the list. They get 45% more talent out of each pick than most teams do. Portland is above average, grabbing 5% more talent than most teams do. Not bad. Here's how it breaks down by Blazer GMs: Stu Inman was the Blazers most successful GM, grabbing 83% more talent in the draft than most GMs. Chad Buchanan sucked. Nolan Smith? Ugh. Olshey's record isn't complete yet. The players he drafted are still gaining in WS. But so far he's at 87% of the expected draft talent. My guess is the players he's drafted will keep generating WS and he'll creep over that 1.0 line soon.
Now, who was the Blazers best value pick, and who was the worst? Here's an overall chart looking at draft position vs win-share. The blue line represents league average (chart from the opening post). Above the line is good. Below the line is bad. If we measure the players actual WS minus the league average WS (expected for that draft pick), here are the top best value picks by the Blazers: Here are the lowest value picks by the Blazers: LaRue Martin was the worst pick by the Blazers, followed closely by Oden. Portland's most notorious 'bad pick', Sam Bowie, is down the list at number 13. Of course Collins and Simons are still active and generating WS, so they will move down this list as their careers progress. But if their careers ended today, this is where they currently sit. Fun data to chew on. Lots of ways to spin data. But I thought this was an interesting metric. I hope the Blazers are successful at finding that diamond in the rough in this year's draft!
Thanks for the hard-core analysis. Beautiful bar chart. 1. A 50-year average doesn't disclose stat ups and downs. Each era had its own playing style. 2. The chart is Post #1 should have a rolling average. You evaluated GMs by mistakes of other GMs. No. 1 is subjective to correct for, but no. 2 is easy.
Yep. Excellent points. When I tried to break down the numbers by eras, I found that one player could throw off the entire average. For example, Kobe Bryant was drafted with the 13th pick and he generated a 172.7 career WS. Any player or GM measured against that number during that era pails. The more I widened the timeline, the smoother the data got. I find it interesting that over a 50 year span, the number 3 pick (52.5 WS) historically out performs the number 2 pick (45.4 WS). Is it because GMs are swinging for that home run at number 2, and leaving a solid lower-ceiling player at number 3? Another way I tried to look at it was actual WS drafted vs what was still on the board (mistakes by other GMs). But when I measured this way, so many GMs all miss big on a few players. For example, Artis Gilmore was drafted 117th, and generated 189.7 WS. Yes, pick 117th. 116 picks were 'wrong' that year. Wow. Just one pick like that and it throws a lot of the metrics way off. All the GMs except one look bad that year. Looking at these historical numbers, it's amazing how many good players were missed by virtually every team at least once. - 2008, #35, DeAndre Jordan, 88.6 WS - 2007, #48, Marc Gasol, 82.5 WS - 2006, #47, Paul Milsap, 91.7 WS - 2002, #35, Carlos Boozer, 80.3 WS - 1999, #52, Manu Ginobili, 106.4 WS - 1998, #32, Rashard Lewis, 90.9 WS - 1989, #36, Clifford Robinson, 89.7 WS - 1988, #52, Anthony Mason, 88.3 WS - 1986, #49, Jeff Hornacek, 108.9 WS - 1979, #65, Bill Laimbeer, 105.6 WS - 1978, #36, Maurice Cheeks, 103.5 WS - 1971, #117, Artis Gilmore, 107.4 WS - 1970, #122, Dan Issel, 82.3 WS There is definitely a science to drafting, and some teams do it better than others (looking at you SA). But with that said, every GM screws up. Really good players slip thru virtually every year. Do you go for the low ceiling, solid talent, or do you go for the home run and hope you can get a kid out of HS like Kobe? Even bad GMs can swing and hit a home run occasionally. But does that make up for passing up on solid talent for 5 years in a row looking for that one home run? Tough call.
Whichever reason, you need to compensate by rolling that average. During the ABA years, a player would hint which league he would join, then the other league would draft him lower than his talent deserved. Why waste a pick. Everyone knew about Gilmore. NBA scouts salivated that he might be another Jabbar. (Turned out to be less quick but more muscular.) There were a number of other ABAers not on your list, which covers only the top 60 picks. Jim Chones, the best center coming out because junior Walton chose to stay his senior year, dumped his teammates late in his senior season and signed with the ABA, ruining Marquette that season. Walton criticized him. The NBA still drafted such players, in case they switched leagues after the ABA contract. Later, high schoolers were automatically discounted in the draft. At the draft, Rashard Lewis cried at the thought of being a Sonic. Houston had promised him and then they passed him over. The callow kid couldn't get time from the Sonics' 2nd-year, 2nd-round Laker pick Ruben Patterson, who was a man compared to Lewis.
Your ambition is exceeded only by your execution! One addition I would make is the highly un-quantifiable player development... Take SA, for example. Have they drafted spectacularly well over the past 20 years, or have they done the best job of developing their young players? Or is it somewhere in between? (Or both?) I think the only way to partially answer that would be to look at draft picks that were jettisoned early on and players brought in, e.g., how their win shares fared for SA vs elsewhere.
I'm kind of surprised how bad Zach (Collins) is. Is this partly a function of how many games he's missed?
Yes. This is ranking him as if his career ended today. For a number 10 draft choice he hasn’t played a lot of games, and the games he has played he hasn’t measured up to what other number 10 draft picks have done. Let’s hope he gets healthy and starts to get some solid games under his belt and moves up on this list.
I once (before the 2015 draft, when we had #23) tried to work out "types" of players who were lower-draft steals. Then I applied it to the 2015 draft. How'd I do?
All things considered, Stu Inman got us Drexler (#14), Porter (#24), and Kersey (#46). He also drafted Fat Lever at #11. That's pretty impressive. Yes, he drafted Bowie over Jordan. Bad call there. But Bowie wasn't a total bust. He ended up playing a respectable 10 years in the league averaging 10.9 ppg, 7.5 rpg, and 1.8 bpg. And he netted the Blazers Buck Williams.
I noticed even the #5 pick did better than #2. Also I wonder how the Olshey rating would look if you switched Lillard to Buchanan era?
Good call. I'm impressed that you picked that out. Yes, if you move Lillard over to the Buchanan timeline, Buchanan jumps up to 1.28. Very solid. Keep in mind that Meyers Leonard (meh) was also drafted in 2012, along with Will Barton (solid).
Yes, I was surprised when I calculated the 50 year average. Although the Blazers drafted Damian Lillard with the #6 pick, historically the #6 pick hasn't performed well. Even The #9 pick outperforms the #6 pick by quite a bit. When I look at this graph, a few data points surprised me. * Walton isn't labeled, but he's the #1 pick at just under 60 WS. Although he had a couple of spectacular years (MVP), he ended up just below average career WS for a #1 selection. His injuries of course cut his career too short. * Same with Roy. A couple of really good years, but in the end he ended his career with about average WS for a #6 pick. * And Rudy Fernandez ended up right about average career WS for his #24 pick. His was due to him going back to Europe early cutting short his NBA career.
I think this is great analysis @42N8Bounce, but we can look a bit deeper than just 3 vs 2 and you can break this down into tiers when determining pick value. Would be interested in looking at standard deviations and variance, but I don't know if there is much statistical significance between 52.5 and 45.4. I would break it down to: #1 #2-5 #6-13 #14- 24 #25-35
