Stars or Depth: NBA Team Building Philosophy

The 2024 Celtics won the NBA title convincingly. They lost only three games and were regarded (with the possible exception of Denver) as the best team in the NBA the entire season.

THe 2010s team building philosophy was defined by having “a big three”. The 2024 Celtics convincingly broke this mold. By most of the advanced metrics, their best player was Jayson Tatum. By these same metrics, Jayson Tatum was between the 10th and 15th best player in the NBA this year.

The Celtics won this year because they had the deepest team since the Steph-Durant-Draymond-Klay Warriors. But how often is this the case? Do more balanced teams tend to be better? Do teams headlined by mega superstars tend to do better?

We’ve come with the data to answer this question. First, though, we explain each of the two NBA team building philosophies.

To receive email updates when new articles are posted, use the subscription form below!

Two Competing Team Building Philosophies

There are two competing team building philosophies. The first is that you need to start with a top 5 superstar. In the current NBA, this means that you’re looking at the tier of Jokic, Giannis, SGA, Luka, and Embiid. Most of these teams try to pair their guy with a second all-star caliber player. For the five guys above, the Robin to their Batman (Batmen?) is Jamal Murray, Dame, Chet, Kyrie, and Maxey.

Each and every one of these guys has a clear #2 to play with. Teams that take the superstar route to team building try to fill in the gaps with cheap guys. This is the basis of the superstar route to team building. And lets face it: Jokic, Giannis, SGA, Luka, and Embiid all made the playoffs and had a chance to make a run.

….but none of them won the finals. In fact, Luka, Kyrie, and the Mavericks lost to the other team building philosophy: build around depth. Instead of putting all your eggs in one basket, put lots of eggs in lots of baskets. The depth philosophy is about having your 3rd, 4th, 5th, and 6th best players be better than their counterparts on the other team.

Both philosophies are equally hard to do. To go for the “Stars” philosophy, you need to get one of the top 5 players in the league. This happens only once a decade or even more rarely for most teams. However, once you get the Batman, it is usually easy to find the Robin because everyone wants to play with the best.

Building a deep team is equally hard. To do this, you not only need to convince two guys to play for you, you need to convince 6 guys to do so. The salary cap is your enemy. Egos of guys that might be stars elsewhere is your enemy. Injuries are your enemy. But, you don’t have to rely on drafting a generational talent.

Which of these philosophies works better? In the next section we look at whether or not one of these philosophies tends to produce more winners.

The Importance of The Quality of Your Best Player

To start, we looked at the relationship between a team’s best player and their overall wins. For all teams since 2000, we ranked them in terms of how good their best player was. The x-axis of the plot below (and all the plots that follow) describes the percentile of how good the team’s best player is. A value of 60 means that the team in question’s best player was in the 60th percentile of all #1 options. A value of 100 means that the team in question had the best player in the league.

The plot below shows team wins versus how good their best player is.

This plot tells us that it is important to have a good #1 on your team. Team quality goes up dramatically until about the halfway point. After this, things tend to level off. A good rule of thumb for team building might be to take all the #1s from around the league and make sure your guy is in the top half.

One other noteworthy thing is that the curve above actually decreases as we approach 100%. This means that it isn’t vital to have the #1 guy in the league. Just a top 10 guy (by advanced metrics) suffices.

While this might seem counter-intuitive, there are some explanations. First: really bad teams tend to overuse their stars. A good player on a bad team will have tons of opportunities and might look better on paper than they otherwise look. A second explanation is that putting all your eggs in one basket is a bad team building approach. This leads us to the more general consensus which is that every Batman needs a Robin…

The Importance of the Quality of Your Second Best Player

Now we want to look at how important a team’s second best player is. We did the same thing as above – measured team wins as a function of how good the team’s second best player was. The plot below shows how this data shook out.

This curve tells largely the same story as when we looked at how good the best player was. The better your #2 option is, the better your overall team quality. The main difference between this plot and the plot from the previous section is that these data don’t level off as early. The figure below shows these two plots overlaid on each other.

While the blue curve (which corresponds to how good your second best player is) doesn’t level off around 50% like the red curve does, the difference is pretty small. It is actually pretty remarkable how comparable these curves are. This tells us that it is equally important to have a good #2 option as it is to have a good #1 option.

But what about depth? Next we look at the other side of the coin: how depth leads to overall team quality.

Importance of Depth on Team Quality

Finally, we did the same analysis while looking at the relationship between a team’s fifth best player and how many wins they had. The shape of this curve tells us how important good depth is to winning.

Lo’ and behold! The curve has the same general shape as the other two. This means that the better a team’s fifth best player is, the better the team is. The only difference is that instead of plummeting near the 100th percentile, it peaks sharply.

It kind of feels obvious that better “fifth starters” leads to more wins. However, the interesting fact is in comparing this to the other two curves we’ve seen. The following Figure shows all three curves overlaid (and we’ll help interpret it after).

Just like before, each of these curves are hard to differentiate from one another. If you have a 50th percentile best or second best or fifth best player, your team is due for about 40 wins. This is obvious because 50th percentile is average and 40 wins is average.

The interesting conclusions come from looking at the tails of the distributions, in particular the right hand side of the plot above. The 90+ percentile corresponds to a team going “all in” on what strategy or another. The red curve plummets near the 100th percentile. This means that having the best player in the league by advanced metrics can actually be detrimental to your team.

This effect is present but lessened for a team’s second best player. However, the effect is reversed for fifth best players. If you have the best depth in the league as measured by how good your fifth starter is, this is really good for team building.

Think back to this year’s Celtics team. They had one of the deepest teams in recent memory (I guess, arguably, the Durant Warriors were pretty special in that regard too. At this point, though, we’re comparing to one of the best teams of all time). Not only were the Celtics deep, but they had a dominant title run. They lost only 3 games.

What Does This All Mean?

To me, the data says that both strategies are equally viable, except at the very extremes. Investing heavily in star power leads to about the same success as investing in depth. Having the 90th percentile (3rd best) star in a given year is just as good as having the third best depth in the league.

On the other hand, if you’re going to go all in on a particular strategy, it is better to go with depth. It is better to ensure that your #5 guy is elite than it is to go after the #1 player in the league.

What this really means to me, though, is that you should take the strategy that comes to you instead of forcing a specific team composition. Let the board fall to you. If you find yourself with Tatum, Brown, and White, you’re only one or two moves away from having elite depth. If you find yourself like the 2019 Lakers with nothing, then trading depth (in the form of picks) for AD to go with LeBron is the strategy available to you. These two teams won championships using opposite strategies from one another.

As a fan, this is kind of refreshing because it means that multiple strategies can succeed. In this era the greatest parity Basketball has ever seen, gone are the days of “the big 3” of the 2010s. You can build a team however you want. We should all be happy this is the result.

Metrics Used for Team Building Philosophy

We need to be able to measure how good a particular player is. The metric we use is “minutes multiplied by BPM”. This number roughly represents how many points above average the player contributed to his team over the whole season.

This is good because BPM is “translated to an average team”. This means that the data is adjusted up or down based on how good the team is. This is important because above it is important to differentiate between “the #5 guy being good made his team good” and “the team being good made the #5 guy good”. BPM being normalized helps in this way.