Jun 23

NBA Champions Super Study: Why Teams Need Multiple “Standout” Players, Part 1

Editor’s Note: Over the next week, TABMathletics author Adam Dobrowolski will look at the makings of NBA champions. This includes explaining the theory of why teams need multiple “standout” players, as well as explaining the theory why those “standout” players need to have enough experience in that role to effectively compete for a championship. Furthermore, he will show each champion from three respective eras (early BAA and NBA days, the ABA v. NBA days, the post-merger days) to illustrate how most teams follow this theory. With this study, we can see each years which teams “fit” the billing of a normal NBA champion.

NBA Finals

Last week, we saw the crowning of a new NBA champion. The San Antonio Spurs won their fifth championship in a 16-year span. Despite their credentials as a potential “dynasty” in the NBA, the Spurs won this championship in unprecedented fashion. As discussed here at TABMathletics, the 2013-14 Spurs became the first team in the NBA history to win a championship despite not having a single player average at least 0.1 Win Shares per team game. So what exactly does that mean? It means that no player on that team put up enough gross statistical production to account for the team’s win at a “standout” level.

So what do we mean by producing at a “standout” level? Win Shares represent more or less what you may think. When a team wins, it is given one or three shares, depending on the sport. That total is then shared among its players, based on how that player produced statistically. For us, a player is considered a “standout” if he or she contributes significantly more than the normal or “average” player in that league. We have a way of determining the “standout” NBA players by means of Win Shares. It involves that “0.1 Win Shares per team game” distinction. This was determined by us only, so perhaps other publications have far different standards of determining which players are “standouts.” Some just look at All-Stars. Others use the eye test and other forms of qualitative analysis. We wanted a clear and decisive method for everybody else to use, so that’s why “0.1 Win Shares per team game” is in use for this study.

We came up with “at least 0.1 Win Shares per team game” as the threshold for determining “standout” players without looking too deep into the formula that makes up Win Shares. Frankly, the formula is quite complicated, but it merely spits out a singular number to illustrate a player’s quality of play. It’s not really logically sound in our opinion to use complicated means for over-simplified ends. Thus, we actually think Win Shares tell us just part of the picture. Regardless, we need Win Shares to be the base here, because there’s no other generally accepted formula that give a quantitative (re; numeric) value for a player’s quality of play. Therefore, we chose “0.1 Win Shares per team game” based on one simple fact: if a player provides exactly one-fifth of an average team’s (re: .500 record) statistical production, we will generally receive about 0.1 Win Shares per team game.

One-fifth might not seem all that hard to achieve, but we think otherwise. Let’s just say a team wants to equal production from each position. That would this player would provide ALL statistical production from his or her position. If a player is doing that, or taking on even more production where other teammates aren’t providing it, we think that player stands out.

Of course, we need to consider some mitigating context as well, so “0.1 Win Shares per team game” can’t be the only basis for determining “standout” players. Injuries first come to mind. Perhaps a player misses 20 games due to injury, but contributes statistically to roughly seven wins. In today’s NBA, this player falls short of the 8.2 Win Shares need to be a “standout” player. Yet, the player averaged more than 0.1 Win Shares per individual game. That’s where Basketball Reference’s Win Shares per 48 Minutes metric comes in handy.

According to BRef, an “average” NBA-level player contributes approximately 0.1 Win Shares per 48 Minutes. We chose 0.16 Win Shares per 48 Minutes to be the “standout” threshold. We did so by attributing 30 minutes of “standout” play per game to a standard NBA season’s worth of games. Mathematically speaking, we get this formula:

(“0.1 Win Shares per game” * 48) / (NBA games in a season * 30 minutes per game) = “Standout” production per 48 minutes
(8.2 * 48) / (82 * 30) = 0.16

Again, we had to take some liberties by attributing 30 minutes per game, but we think it’s fair. It’s widely accepted for stats to be prorated to a “per 36 minutes” level for statistical analysis, but very few players actually average 36 minutes per game. In fact, only 17 players did so in the 2013-14 season. We wanted a threshold more common for NBA players, so we chose 30 minutes per game. In total, 104 players averaged that many minutes in the 2013-14 season, which averages out to almost 3.5 players per team. It should be commonplace for the minutes per game, so the “standout” play shines. You don’t want the “standout” production to be wiped out by “standout” minutes. (Think about it like this: S / S = 1, but S / 1 = S. We want S as our dividend, not 1.)

Looking at this solely in the context of the NBA and former major professional American basketball leagues, we will now have the basis for finding “standout” players in our study. We will ONLY consider players and teams from the NBA, ABA and BAA. In the context of the current NBA, “standout” players are determined by fitting into at least one of two possible criteria:

  1. Tallying at least 8.2 Win Shares in a single season
  2. Tallying at least 0.16 Win Shares per 48 Minutes in a single season

This will be the two and only two criteria for determining “standout” players, save for one exception that we’ll get into tomorrow. Furthermore tomorrow, we’ll account for age and experience of these “standout” players.

We think a true NBA championship contender needs two “standout” players. That accounts for 40 percent of a starting lineup. This obviously significantly than 20 percent, which is why we don’t think one “standout” player is enough for a team. In the playoffs, defenses can be scheme around that one player and shut down the rest of the team. Furthermore, we have precedent on our side. We understand that some circular logic is at play, but we already pointed out last week that the 1977-78 Bullets was the only post-merger NBA champion without at least two players fitting at least one above criterion. Honestly though, this was our hypothesis long ago.

We thought this based on basketball theory. As already mentioned, teams can scheme around one “standout” player. It’s much tougher to scheme around two “standout” players. We thought this based on watching the game unfold, as somebody like LeBron James needed to go to the Miami Heat to win a championship. We thought this based on conventional wisdom, as many NBA experts debate the role of “star power” in winning championships. Just consider our “two standout player” statement as our hypothesis. Now we’re out to support that hypothesis by looking at league championship precedence for the remainder of this week.

Leave a Reply

%d bloggers like this: