Is there a bias towards offensive players in today's NBA?
Sports are for the fans. The fans who feel great pride in their cities teams, who form attachments to players and coaches. The fans talk as if they are part of the team, as if the team is part of them. But more than that, the fans pay for tickets, for merchandise - the fans pay for the league and it’s players.
Fans - well most fans - like scoring. They like watching their favourite player go out and drop thirty points, they like Vince Carter like high-flying dunks, they like Dame Lilliard logo threes.
This isn’t by accident.
For the last couple years, the leagues have been slowly shifting its focus and rules to favour high producing offences - not just a trend in the NBA, but rather in almost all modern sports leagues. Whilst there is clear evidence of this from as early as the 2000s, it can be argued to have started as early as the 1980s with the rise of popular offensively gifted figures such as Michael Jordan.
Some rule changes and their advantages to offensive play were explicit, such as legalising zone defenses and introducing defensive three-second violations after the 2000-01 season in an effort to "move away from iso-heavy and stagnant NBA offenses" and unclog the paint, and the removal of hand checking after the 2002-03 season, "freeing up the guards to roam and penetrate without the fear of being manhandled."
Whilst others were more subtle, often never really being rules but rather directions from the league on how games should now be officiated. This is an issue with the perceived lack of calls in the NBA for travelling and carrying, as well many offensive players being allowed to use ‘non-basketball moves’ to draw foul calls.
This idea of the game changing to favour offensive play isn’t just fan speculation either, many prominent NBA figures have spoken about it in recent years.
“I think that the NBA has done a great job of changing with the times and it gets better and better.”
These sentiments are shared by recently retired Kyle Korver, a 17-year NBA vet, when he stated in the same article: “I think the game is much more fun and open and free than when I first came into the league. Most of the rule changes have been so we can score more points and that’s made the game more fun with higher scoring and a lot more actions. It’s more of an offensive game.”
However, it’s pretty easy to see how two players most renowned for their offensive skills - Boone and Korver - would appreciate the way in which the modern NBA appears to be heading.
What about the defensive specialists whose lockdown skills allowed them to act as anchors for their team defensively?
Players like Dennis Rodman and Ben Wallace, who combined for fourteen all-defensive teams (12 first-team, two second-team) and six DPOY (Defensive Player of the Year) awards in their combined thirty seasons - as well as nine rebounding titles and one blocks title - would be heavily criticised for their lack of offense in today’s NBA.
The duo who only had one season averaging more than ten PPG - Rodman, 11.6 in 1987-88 - would be called ‘low-scoring’, critiqued for their lack of floor spacing, and their questionable free-throw shooting - 58.4% for Rodman, 41.4% for Wallace - would probably have them removed from the court during crunch time as if their defensive skills didn’t more than make up for their offensive weaknesses.
Furthermore, this bias towards offensive players - and particularly scoring - can be seen in the awards given at the end of each season.
The clearest example of this is the MVP (Most Valuable Player) award, which has now basically become the ‘Offensive Player of the Year’ award - similar to how the NFL MVP is now basically a Quarterback award.
In fact, since 1990, only two NBA MVPs have averaged less than 23 PPG in the season they won - Steve Nash in 2004-05 (15.5 PPG) & 2005-06 (18.8 PPG). Whereas this was relatively common in all the decades before the emergence of Jordan, and the focus on scoring by many fans and the media - 1950s (two times), 1960s (six times), 1970s (two times), and 1980s (two times).
So, how does this apparent favouritism towards offensive production affect the MIP (Most Improved Player) award?
When assessing a player's improvements, arguably the easiest change to spot is a player’s PPG, as this has a clear numerical value that can be seen to have a direct correlation to winning - such as the old saying goes, 'the easiest way to win is to score more points than your opponents.'
However, scoring isn’t the sole factor a player can improve in - even before delving into more advanced analytics, the box score offers a key list of other aspects a player can help their team in.
Some of the stats - rebounds and assists - can be clearly understood in how they would help a team win a game through offense; you need the ball to score, so creating more chances for your team can help you win offensively. Inversely, limiting turnovers can give your team more chances to score.
But even for the basic level defensive stats - steals and blocks - not only does reducing your opponents scoring chances help you control the score of the game, but the old adage ‘defense wins games’ is particularly true in a game like basketball, where a defensive stop can be the key to getting an uncontested fastbreak scoring opportunity.
It is with this idea that overvaluing offensive stats can lead to disadvantaging a player with significant defensive improvements from winning MIP that I have analysed the data of the 2010s MIP winners, and a select group of players with significant defensively improved seasons during the same period, to attempt to see whether the idea of an offensive bias can be backed up with statistical production.
Following will be a short selection of tables presenting the data of the most convincing cases, but there is also a more in-depth version here, including analysing more defensive seasons, for anyone that is interested - a glossary of key terms is available on the first few slides for anyone unaware of certain basketball terms.
Furthermore, PPG - and the other box score stats - isn’t always the most accurate way to assess a player's offensive improvements. Whilst they give a good indication of the the players production, they fail to give an indication of their shooting accuracy.
Surely someone who only scores ~3.5 more PPG - such as Green and Anunoby - but on significantly improved shooting, is not only a more improved offensive player, but also more beneficial for their team, than someone who scores more but on worse splits.
But it is true, numbers can lie - at least solely using box score numbers can be misleading - deeper analysis must be had to better see a player’s true impact.
As said before, a player's scoring is not a total representation of their offensive ability - the same goes for steals and blocks not being the end all for a player's defensive ability.
This is why various analysts over the years have created various advanced metrics that compile data from numerous aspects of offensive and defensive data, in order to better show a players total impact on both the offensive and defensive ends of the court.
For the case of judging a players total improvements, I believe WS and VORP to offer the most complete understanding as they present how a player’s skill directly impacts their teams winning (WS), and how they compare to an ‘average level’ player of the same position (VORP).
Whilst the data presented shows an improved analysis of a players defensive improvements compared to the 2010s MIP average, there is still a further level we can go in our defensive analysis - specific advanced defensive metrics.
The study of these data sets allows for a more in depth understanding of a player’s actual defensive ability, as a lot of great defense can be played without producing an actual ‘stat’ - such as with contesting shots, a player can have a lot to do with an opponent missing a shot without actually blocking it.
However, we are at somewhat of a disadvantage for no other reason than these stats not being tracked earlier. DVR and ‘Hustle stats’ - deflections, charges, shots contested, and loose balls - were not officially recorded until the 2013-14 and 2016-17 seasons respectively, meaning data for their improvements cannot be analysed until the following seasons.
Despite this limitation, the data offered is still of such good value that I believe it is worth including the amount given, as it still offers room for a direct comparison between the defensive seasons during the times since they were collected, and the MIP seasons of those seasons.
Through the data presented, clear cases can be made for all four of the defensively improved players to be considered equally - or greater - improved than the average MIPs from the 2010s.
In the case of George, the data shows that he made significant progress across the board during the 2018-19 season, with only five categories - BPG, 3PT%, and STL%, BLK%, and deflections - not showing any improvement.
George’s main area of improvement above the MIP average was in the advanced metrics, where the only three categories his improvements didn’t outperform the MIP average - OWS, WS, and ORtg - were all still significantly improved from the season before.
Whilst George’s overall improvements do not outright outperform the MIP average to the point where it is clearly due to an offensive bias that he didn’t win, there is still sufficient evidence for an apparent bias in the voting patterns for two reasons.
Firstly, whilst not winning the award outright cannot be said to be solely due to an offensive stat production bias, the fact that George only received three voting points (T-11th) suggests his defensive improvements were overlooked by the voting committee - for reference, the winner, Pascal Siakam, had 469.
Secondly, George actually won MIP in 2012-13, a season in which he had better offensive improvements than this season, but weaker overall. So, the favouring of George’s more offensively improved season, and his surprising lack of voting points this season, suggest that there is at least some level of an offensive bias in regards to MIP voting.
For VanVleet, the case for an offensive bias affecting his chances of winning MIP are similar to George’s.
Although he had smaller levels of improvement within his shooting splits compared to the MIP average, they were all still positive changes.
VanVleet’s advanced metric improvements also showed some weakness compared to the MIP average, although there was only one category he actually regressed in - ORtg. For the defensive aspects, VanVleet showed significant improvement compared to the MIPs.
This trend continued when looking at the advanced defensive metrics, where there was only category VanVleet was reasonably under the MIP average in - shots contested.
From looking at the data presented, it appears that VanVleet’s main factor weakening his case for MIP was - similar to George - his limited shooting splits improvements.
All other data presented suggests that VanVleet should have been seriously in the consideration for the award that season, but instead only received six voting points (10th) compared to the winner’s - Brandon Ingram - 326.
But it is the cases of Green and Anunoby that present arguably the most convincing cases for an offensive bias.
For both, the only traditional box score stat they were not at least equal to the MIP average increase in was PPG, but even then they both improved in that category.
In fact, this small reduction in their scoring improvements compared to the MIP could arguably be negated by their significant shooting splits improvements.
Anunoby showed large improvements across the board compared to the MIP average, suggesting that his relatively smaller scoring improvement should still be considered a net positive, as he is producing it on a significantly more accurate basis, and so being more beneficial from a team perspective.
Green showed similar improvements in his shooting splits, with his massive eFG% increase showing how drastically his overall shooting production was. However, it is worth noting that Green suffered greatly in FT% that season, especially important as none of the 2010s MIPs had a negative value in that category.
Similar large improvements compared to the MIP average can be seen in both the advanced metrics and advanced defensive metrics of Green and Anunoby, once again highlighting how their overall game improved significantly those seasons.
This strong case presented for the reasoning Green and Anunoby failed to win MIP awards is furthered still in the fact that neither of them received a single vote for the award, with twenty-five and fifteen players receiving at least one vote in their respective seasons.
Whilst there appears no definitive way to answer whether not an offensive bias exists without directly interviewing various NBA team members (coaches, GMs, etc), or the voting media members, hopefully this analysis has shone light on the situation and the apparent bias towards offensive players, and particularly offensive improvements in regard to MIP voting patterns.
(Publication date: May 30th)