clock menu more-arrow no yes mobile

Filed under:

From Past to Future: Part One

Getty Images

It's been said over and over again - basketball is different from all the other sports. Because only 5 players are allowed on the court for each team at any time, one person makes such a big difference between winning and losing. Compare this to football, soccer, baseball, rugby, etc where a single man has a harder time affecting the game like in basketball. It's why the idea of an on-court superstar (or something that approximates it) is such a huge boost to a team's title chances and one that weighs heavily on front offices' mind.

For any team that's fortunate enough to win a #1 pick in a draft, the odds of you getting that much coveted on-court superstar just went way up. As I've detailed in my previous post here, historically speaking, the #1 pick usually produced a quasi-star or a superstar, based on my very basic Weighted Win Share Score. In fact, if you extend the WWSS right up until 1982, the average comes out at 1.15 - that's near the "superstar" level we set at 1.44. Of course, one thing that wasn't included in the post was that, of the past 15 #1 picks, only 5 stayed with their team past their rookie extension (or in the case of Big Dog, after his 1st contract). In fact, from 2005 to 1982, only 8 out of 28 #1 picks stayed with their team past their 1st contract OR rookie extension (in the era of rookie scale contracts).

This piece by Sean Higkin on the dilemma of teams after drafting a quasi-star~superstar is very, very good. His idea applies even more so to players drafted #1. Typically, players drafted #1 are given the most resources and attention in terms of development, nurturing and guidance, so the idea of losing them is hard on the team. Imagine working so hard to build your own house - you buy the wood, the nails, the concrete, the tiles, and sacrifice your time, building your house arduously with your own blood and sweat, only to see it destroyed by a hurricane (no insult intended) or stolen by a bank (foreclosure). Sucks doesn't it?

It is why it's imperative that the moment the New Orleans Hornets won the lottery back in May 31, plans should have been set and made on being contenders and making our franchise as marketable a place as possible (who are we kidding when we say we want someone who doesn't really search that? Even just a tinge of marketability helps a ton). Marketability aside, this post would like to look back on the teams of the past few #1 picks, what they did wrong, what they did right, how they constructed their team, what happened in the front office, sideline changes, etc. Now, my real concrete memory of the NBA outside of the MJ/Pip/Rod-Bulls and the Stockton/Malone Jazz started in 2002, when Yao was the #1 pick. Back then I was an 11 year old who cared nothing about everything on the back stage. It wasn't until 2003, when LBJ was the #1 pick, that I started caring about front offices, coaching, all trades (even the small ones). And that was almost 9 YEARS ago so I'll be depending on websites to not only help me relive my memories but also on information on team transformations BEFORE 2002. Moreso, I'll be depending on you guys to help me understand how the teams past 2002 constructed their teams.

Comparison Test

To start off our analysis, I began by comparing the past 26 #1 picks (from DRose to James Worthy) and tried to find out who were the players who are similar to each other (Note, I excluded Greg Oden since he only had 2 years worth of data on him). Of course, different teams build around their #1 picks in different ways depending on what type of a player he is. Was he a big man? A faceup guy? A post up guy? was he a shooter? was he a slasher? was he a scorer? or was he everything in between?

Those are questions that I tried to answer. You may think it's easy because it's very apparent from how they played (or their reputations). It's easy to see that Robinson, Olajuwon, Ewing, Shaq, Duncan, Yao and Howard go well together as those "post" players. DRose and Iverson are perimeter gang busters and LBJ is everything in between. But I am not concerned with how their careers evolved - I am only interested in the 1st few years to look at how the team was able to convince the player to stay (or leave for that matter). So the 2012 LBJ wasn't exactly the same as the 2006 LBJ. Nor is the 2012 Duncan similar to 2000 Duncan, nor 2009 Yao similar to 2002 Yao, nor *you get the point*. So I only considered the first 4 years of the players in analyzing, which players were most similar DURING the time when teams were just starting to develop and build around them. Sounds logical right?

Plus if you know me, I never like using one tool alone (eye test and stat analysis). As much as possible, I want to mix in both into my studies because I believe each tool catches something the other can't.

Now that we've established which data to use, the question then becomes, how can we measure the similarity of 2 players that spans almost 3 decades? Initially, what I wanted to do was use their shot chart as data. Problem is that, hoopdata only started posting shot locations in 2006/2007. That leaves me with just 2 players with a full set of 4 years I could use (DRose and Bargnani). So that was out of the question.

Eventually, I decided on these 22 statistics as measuring tools on how similar two players are (courtesy of Basketball Reference) - Age, Experience Prior to the NBA (college or overseas), Games Played, Games Started, FGA/36, FG%, 3PTA/36, 3PT%, FTA/36, FT%, PER, ORB%, DRB%, TRB%, AST%, STL%, BLK%, TOV%, USG, ORTG, WS and WS/48. I excluded DRTG cause I really didn't like it as a statistic (LOL).

The rationale: I consider Age and Exp as the 2 most important aspects in comparing similarities. My rationale is that they are the 2 factors that meant (without certainty of course) "readiness". When a 4 year collegian enters the league, he is usually branded with a certain "readiness" factor around him.

I used FGA, 3PTA and FTA seperately because I didn't want 3 PT shooters to be mixed in with high FG shooters (eFG%) or mixing 3 PT shooters, or high 2 PT FG% scorers with FT drawing. At least with those 3, we could somehow get an idea of the players style of play. Shot location would have been better, but alas, it is not available.

PER was used because it's a single digit metric that is best used in comparing high usage players. I used the pace adjusted metrics to cover for the cross-era comparison (remember, the league was fast back in late 1980s, super slow in the 90s and then slowly found a middle ground in the 2000s). Finally, I used ORTG, WS and WS/48 because they paint a good picture of how good offensively the player was.

Also, lockout seasons of 2011 (for Derrick Rose) and 1999 were extrapolated to 82 games to be "fair".

Similarity Matrix

How do we say that the players' first 4 years were similar? I used percentage change (a very basic concept). So what I did was find the minimum between the 2 stats of 2 players, use that as my denominator and calculate for %change. Basically what happens is: (max(x,y)-min(x,y))/min(x,y).

And then I assigned weights to all 22 statistics, multiplied them by (1-%change) to get a picture of how different one player is from the other. I gave age and exp a weight of 10, and the other 20 a weight of 5. So the total should be 120 (2x10 + 20x5). We standardize by simply dividing the total "score" by 120 and then multiplying by 100. What we get is a "percentage" of how similar the 2 players are.

Now understand that a lot of things are left of from this analysis - for one, we don't know WHERE those FGA's came from (it's why I wanted shot location instead). Shaq's FGA will come very differently from someone like Yao or Duncan. Similarly, Lebron's shots will come differently from Glenn Robinson's. So that's one flaw of the similarity "score". Another thing is whether those shots came from the low post, high post, slashing, etc. but there is no sort of data to track those things. What we do have are these data. So, here is what I call my "Similarity Matrix". It basically shows how similar 2 players are.

Where does Anthony Davis fit into all of this?

First off, we need to look at which players are closest to Anthony Davis. For one thing, if Anthony Davis averages anywhere near his lone season in college, then the following players will be the people that will be the most similar to Anthony Davis:

Yao, Brand, Duncan, Shaq, Robinson, Ewing, Olajuwon. Ok that was obvious. What do all these players have in common? With the exception of Shaq, all 6 players are players who averaged over or near 50% from the field (Brand is the only one who shot below 50% at 49.5%), shot FTs at or near 70% (with Olajuwon and Duncan averaging 66.4% and 68.1% respectively), PERs of over 20, with rebounding numbers close to 11/22/17, Block percentages of 4%, turnover percentages around 12%, Usage rates of around 25%, and WS/48 close to .200 and played (and started) almost 90% of all possible games in their first 4 seasons (approximately 328 games).

For comparison, Anthony Davis shot almost 63% from the field, 71% from the FT line, a PER of 35.1, rebounding numbers of 11/24/18, block percentage of 13.7%, turnover percentages around 8.6%, usage rate of 18.8% and a WS/48 of 11.8. Assuming there is some sort of adjustment period because of the HUGE talent upgrade that Davis will face on a night-to-night basis, compared to his college season, then the numbers above, aren't really "out of the question".

So yes, the idea that Anthony Davis is the best prospect since LeBron and the best big prospect since Duncan holds some water. Why no Dwight Howard you may ask? Well it was only in his 5th season when Dwight Howard truly broke out as a superstar. Compare this to the other 7 players. Robinson, Olajuwon, Duncan, Ewing, Shaq and Brand were explosive right from the onset. it was only Yao who struggled from this group and he exploded onto the scene in his 3rd season.

Something that you'll notice out of that group is their age - none of them came into the league as young as Davis. The youngest who came in was Shaq, and he played 3 seasons of college basketball. So comparing Anthony Davis, who won't turn 20 until March NEXT YEAR will be kind of like comparing apples to oranges. You can't really compare which fruit is more "yummy" because both have different tastes (so it would be unfair not only because "yummy" - ness is a very subjective idea but also because they taste differently). You can however compare which fruit was "healthier" or "sweeter" or "more sour".

So, what I'll do is to make 3 separate posts detailing how those 7 teams evolved through the rookies first 4 seasons. This is so that we'll get a good idea of what strategies failed and what strategies were successful. We'll first talk about Yao, Brand, Duncan and Shaq. Followed by Robinson, Olajuwon and Ewing. The last post will focus more on the following 4 players - who are the 4 franchise cornerstones drafted since I started following the NBA intently - Chris Paul, Dwight Howard, Lebron James and Dwayne Wade. All 4 players are players I consider to be the superstars from the past 10 drafts (I'd love to include Roy, but that would just be depressing). All 4 are two way stars that can completely change the complexion of a team (no matter how bad that team is on paper) and who can affect the flow of a game all on their own. I didn't include Kevin Durant because he just finished his 5th season.

There you have it. Expect 3 blog posts in the coming weeks that will focus on those 11 players.