Last year, I wrote my favorite piece ever for BSS called The Development of Anthony Davis. I enjoyed it so much that I decided to run it back this year. The original post was long and detailed. I’m not going to repeat a lot of the set up or analysis here. If you have questions, answers can probably be found at the link above.

But here’s a readers digest version of things. My original objective was to compare Anthony Davis’ development to the other great big men in the modern era. To do this, I selected a group of 11 comparable modern NBA stars and examined their WS/48 relative to their age. This actually yielded some pretty interesting results. For one, it was neat to see career trajectories mapped out. Despite a really uniform average, some players were outliers in terms of their peak. Karl Malone, for example, started off really slow but saw his peak years come between the ages of 33-36. David Robinson, on the other hand, had a much more traditional peak between the ages of 28-30, despite starting his career at 24.

As it related to AD, the results were pretty simple: he’s looked like an all-timer. Now, we have another year of data, so let’s see how AD developed between his second and third years.

New Data

I want to make a couple of points before I continue. First, I explained and defended the choice and set up in the last post. Please don’t comment on those topics without reading the first post. Second, the first post also explains some concepts like proxy variables and some of the equations. I want to keep this short so I won’t repeat that stuff here, but I do know that the proxy variable stuff was a sticking point for some last time. To be clear, I’m not saying, “As a player gets older, they get better.” The data doesn’t even show that. The idea is that age is correlated with the returns to development and experience. As a player gets older, they generally gain experience through playing in games, practice, and training sessions. The important point is merely that these two variables are some what correlated (not even perfectly correlated).

Enough of that. Now, we’re on to the fun stuff. Out of our 11 comparable players, only 3 players (Kevin Garnett, Dirk Nowitzki and Pau Gasol) were still playing. That gives us only 3 new data points to add to our data set, which isn’t much to get excited about. Having said that, our data for older players was pretty thin to begin with as not a lot of people play past 36. So we did gain a bit of robustness on the back end of our model. Obviously, that won’t matter much but more robustness is preferred to less.

Other than that, the data and methodology is the same. We have the same scatter plot as last year with those 3 new data points. Here it is.

Yup, just as hard to read this year. If you look closely, there’s some interesting stuff. Also, note that some data points cover up other ones.

If you took a look at last year’s piece, you’ll notice one difference. AD isn’t on this graph. Yeah, I’m saving that for the next section.

How Good was AD Last Year?

Obviously, this is a post about where Anthony is going, but that includes looking back. How good was Davis last year? Take a look at the graph below.

If you can’t tell what’s going on, I added AD to the data set. He’s those red squares. Everyone else is a blue diamond. Yup. Anthony Davis has exceed everyone else during his 19 – 21 year old seasons. More importantly, not many players had a better season, in terms of WS/48, than AD did last year. In fact, this data set included 181 NBA seasons, and there were only 5 seasons with a higher WS/48. Here’s a summary table of the top 10 seasons.

5 of those top 10 seasons were MVP seasons. Some guy named Michael Jordan won 2 of the other 5. Look at the age column. The youngest anyone in this group has achieved that level of success is 27 (O’Neal, Nowitzki and Garnett). AD wasn’t great last year. He was absurd. This chart actually makes the MVP conversation about AD seem silly. He already is an MVP candidate. He just needs the winning team to match.

One of you is going to come in and spoil the party so I’m going to spoil it first. Yes, we are just using WS/48, which like all advanced stats is not perfect. It certainly doesn’t pick up everything, and this isn’t the end point of any conversation. It also has a bias towards efficiency. I think that’s a good thing, but there are a few fair arguments to overvaluing efficiency. I mean a few players from our data set are left off this list despite being objectively better players (e.g. Olajuwon and Nowitzki). Still, WS/48 is a good short hand, and it is especially useful in 2015 given where the league is going.

Look, I’m not saying AD had the 6th best season of all time last year. I’m saying he was incredibly good. He’s not perfect, but given the fact that he is only 22, the ceiling can’t really be measured.

The Future

With that, let’s turn our eyes to the future. How good can Anthony Davis be? I’m not sure, but he’ll be really good. Like he has a chance to be a top 3 player of all time. I know that seems lazy, but I’m not sure what else to write. There might not be another player in modern NBA history with as much potential at 22 than Davis. Hell, LeBron James didn’t even break a .27 WS/48 until he was 27. So I’m not sure where the ceiling is, but I know it is way up there.

I want to end with a simple request. We have to stop trying to compare AD to other players. As this data shows, he’s just too different. A good 1 to 1 comparison doesn’t exist. I’m not saying comparing players is a bad thing. Actually, from the perspective of statistical analysis, it’s incredibly important. Still, this is a case where it just doesn’t work out.

Also, I suspect this will be the last Development of Anthony Davis post. The analysis may comeback, but it will be under a different name. At this point, I’m done focusing on projecting AD, and I bet in a 2 or 3 seasons we’re going to start talking about where he stands as an all time great. Now, that’s a post I’m looking forward to.