By Alex Koenig and Michael Owen
The Dunk is as elusive as it is tantalizing. It’s been described as the ultimate athletic feat, combining size, speed, agility and strength. And unlike hitting a hole in-one or throwing a tight spiral, most of us will never dunk – no matter how hard we try (unless it’s on a 7-foot hoop or NBA Jam). Steve Nash, a two time NBA MVP, has never dunked in a game. Mikhail Prokhorov, the new 6’7” Russian billionaire owner of the New Jersey Nets, has made the claim that he will be the first NBA owner to dunk. Ever.
When it comes to the jam, people usually fall into one of two camps:
- “Dunks are awesome!” They get the crowd going, and are high percentage shots.
- “Dunks overshadow fundamentals.” It may be true that the incentive of dunking – making Sportcenter’s top 10, stardom, etc. – might hurt the fundamentals on a certain level.
However, what is rarely addressed is whether or not dunking, as opposed to laying it up, can help your team, both in the standings and on the bottom line. Let’s address the win column first: are teams that dunk more likely to win?
Looking at data from the 08/09 and 09/10 seasons (via CBS SportsLine and 82Games), the answer appears to be yes. In 08/09, the average team dunked the ball 310.4 times in an 82-game season. Of the 13 teams above that mean, 9 of them made it to the playoffs (69.2%). When looking at the 17 teams below that average, only 7 of them made it in (41.1%). Going deeper, the mean winning percentage of those 13 teams (0.550) was nearly 9% better than that of the 18 below (0.462). These numbers are even more drastic in 09/10, with the average number of dunks per team dropping to 292.1. Out of 16 teams above the 292.1 threshold, 12 made the playoffs (75%), with only 4 of the 15 below qualifying (26.67%). The difference in winning percentage is even greater, at a 14% advantage for the dunking teams (0.569 vs. 0.421).
Conventional wisdom would assume that the reason for this disparity is because good teams not only dunk more than bad teams, they score more than other teams as well. However, when factoring in dunking as a percentage of a team’s scoring, the same differences exist: a difference of 69% vs. 41% in 08/09 and 75% vs. 26.67% in 09/10.
These examples hold under more rigorous analysis as well: using data from both seasons, dunks were a significant predictor of winning percentage (P = .033) as well as a significant predictor of whether or not teams made the playoffs  (P = .037). Each additional dunk correlates to an extra .05% in winning percentage, and a 1.008 multiplicative increase in playoff odds (although the technique used, logistic regression, tends to overestimate this figure in small sample sizes).
The reasons for this disparity, though often overlooked, are self-evident when viewed through the eyes of a fundamentals-based basketball enthusiast. Though the odd through-traffic, isolation super-dunk does occur, the majority of dunks come as a result of one of three things: fast-breaks, hitting the open man, and posting up.
Dunk totals also represent a general aggressiveness: a willingness to drive through the lane, pick up fouls, and to try for the offensive rebound, as well as athletic ability – all things that a good team should have. Dunks are attributed with a negative stigma from basketball purists because of players like Andre Iguodala: me-first wing players on bad teams whose scoring and dunking has garnered them more media attention and financial compensation than their actual basketball talents merit.
However, this is a fallacy that considers the dunk an individual achievement. Going back to the example of Steve Nash, his 09/10 Phoenix Suns, despite having only one player in the top 40 of individual dunks (Amar’e Stoudemire) and a starter who never dunked (the aforementioned Steve Nash), had the fifth most dunks of any team (380) while registering one of the highest winning percentages (0.659) and, not surprisingly, making the playoffs.
In both years, there exist outliers on either end of the spectrum. The 08/09 San Antonio Spurs dunked an abysmal 73 times, less than half of the next-closest team, but still made the playoffs and won their division. Likewise, the 09/10 Milwaukee Bucks finished dead last in dunks (136) but still qualified for the playoffs. For the Spurs, this is primarily due to their core three players (Tony Parker, Manu Ginobli and Tim Duncan) possessing all the qualities associated with dunking – aggressiveness, post-ability, basketball IQ – but not possessing the eye-popping athleticism that their younger selves once had. The 09/10 Bucks suffered from a similar situation, with their offense being dominated by guards and an elite, but finesse move-based, center, Andrew Bogut (ranked 37th amongst dunkers).
On the opposite end lie the 08/09 Memphis Grizzlies and 09/10 Philadelphia 76ers. These two teams combined to average 399.5 dunks (97.9 above the league average) while having a winning percentage of 0.311 (.189 below the league average). These two teams were led by Rudy Gay and Andre Iguodala, both top-10 dunkers, who will be paid a combined $25 million next year to lead sub-.500 teams.
Given that dunking does appear to correlate with winning, what can we take this to mean? The thing to avoid is seeing dunking and winning as a causal relationship (i.e. dunking more means you will win more). Instead, we should view high dunk totals as a representation of a good team (i.e. good teams will produce more dunks). Dunks are easy points, but they should not be the only points. The threat of the dunk – through posting up or through driving – is almost as important as the act of dunking itself. It opens up the playing field and spreads the defense, lending itself to open shots and open passing lanes. Another, more controversial view in light of the recent Donaghy game-fixing scandal:  perhaps the NBA promotes the advancement of “exciting” more marketable teams by fixing games in their favor. In other words, is it possible that the correlation between dunks and victories does not occur by coincidence but by design? We’d need to do a close examination of foul calls to be sure.
These conclusions support the notion that “Dunks are Awesome!” but also that, despite the complaints of the old guard, dunks have become a fundamental. As if NBA fans needed another reason to fear the upcoming season, the only team that will enter the 2010/2011 season with three players in the top 20 for dunks are, you guessed it, the Miami Heat.
Do you have a list of the total number of dunks by tem for the 2009-10 season – or is it readily available at a site?
Interesting article – have you factored out putback dunks of the analysis? Obviously there you have not only the dunk, but an offensive rebound – the team may not have scored if the player grabbed the ball and tried to go up for a layup…
That said, the logic that you need athleticism to be able to dunk (jumping ability, timing, speed on the fastbreak, etc.), teams which dunk more tend to win more would lead to the conclusion that athleticism is correlated to winning in basketball.
@ Karl Von Asselt: For 09/10 CBS sportsline had individual dunk stats which we then compiled and divided up by dunks per team. It was not completely scientific because they did not have game-by-game dunks. So, for example, if a player was traded at the trading deadline we evenly partitioned his dunks. For 08/09 we used 82games.com which had each teams dunk% (ie % of points that came from dunks). We used that to find the total points scored off dunks and then divided by 2.
@Tamid Kamil: Interesting point about putback dunks. In order to factor that in we would have to do a comprehensive study of every dunk (unless there are boxscores that note putback dunks). It might be insightful, but probably not practical considering they are relatively rare. The notion of athleticism correlating to winning is also interesting. The only problem would be how to quantify athleticism (Vertical leap? etc.?) Also, it wouldn’t account for worse basketball players like Saer Sene who were freakishly athletic.
@Alex – I assumed you were parsing play by play data which would easily allow you to isolate putback dunks.
Without having access to each player’s combine-type athletic data, it’s not feasible to draw a direct line correlation between athleticism and winning but you can/did use a proxy (dunks).
Regarding whether or not the Bucks additions will improve the amount that they dunk, Drew Gooden brings his 36 dunks in 70 games to the table, which combined with Andrew Bogut’s 61 dunks in 69 games should lead to increased totals for both this coming year (assuming they stay healthy). Rookies Larry Sanders and Darington Hobson should also contribute, if given the requisite playing time. All things considered the Bucks should move out of the dunks basement this year.
I don’t see anything useful in this article.
If you were trying to find the benefit of dunks, you should be comparing dunks to shots within 3 feet.
If you are pretending dunks mean an inside presense, then you should be comparing short shots to all shots.
What you have is too many independent variables to come to any conclusion whatsoever, much less the laughable conclusion that dunks are a fundamental.
Come on, step up your game a bit.
I happen to find it interesting that teams that dunk more frequently also win more frequently. Would I find it more interesting if we found the play-by-play data, adjusted for the confounding variables, and then actually determined whether teams should in fact dunk more or less than their current pace? Sure, and we’re working on it, but that’s not what the authors set out to do — they made clear they were going after a correlative relationship and made a couple educated guesses as to why it might exist.
If you’d like to do the research you described, we’d be more than happy to link to it.
Sure, it’s fine to correlate dunks with wins. It’s just pretty useless.
The “educated guesses” are even worse:
“The threat of the dunk – through posting up or through driving – is almost as important as the act of dunking itself. It opens up the playing field and spreads the defense, lending itself to open shots and open passing lanes. Another, more controversial view in light of the recent Donaghy game-fixing scandal: perhaps the NBA promotes the advancement of “exciting” more marketable teams by fixing games in their favor. In other words, is it possible that the correlation between dunks and victories does not occur by coincidence but by design? We’d need to do a close examination of foul calls to be sure.”
What there follows from anything in the rest of the post?
agree with TGT that there’s basically nothing to this; a simple correlation tells us very little. other variables to consider would be block % on dunks versus layups, shooting % on dunks versus layups, and goaltending % on dunks vs. layups.
also, many advanced stats suggest that Andre Iguodala is an incredibly valuable player, and placing him in a group of pejoratively-described “me-first wing players on bad teams whose scoring and dunking has garnered them more media attention and financial compensation than their actual basketball talents merit” seems a personal attack of some sort with little analytical value. in addition it completely misrepresents Iguadala’s skillset (defensive and hardly “me-first”) and his overall value on the basketball court. (this is not to say that he’s a great offensive player or that he doesn’t take bad shots.) See Pelton:
http://www.basketballprospectus.com/article.php?articleid=649
It seems to me that you’re over-complicating a simple issue.
Better/more athletic players dunk more often.
Teams with better/more athletic players win more often.
The relationship isn’t causal. That is, dunking doesn’t necessarily represent a winning strategy per se.
It’s just symptomatic of a team staffed with the personnel to win.
Was there ever a time when Manu Ginobli and Tony Parker dunked a whole bunch for San Antonio? Manu’s glass layups and Tony’s teardrops effectively supplant dunks in their games, and I can’t think of a point in their recent history when they dunked that much more often. Timmy might have seen the rim more often back in the day, but the range on his bank shot has always been one of the more remarkable and effective parts of his game. Also worth noting: Tony Parker is not over the hill… not even close. He’s just a tiny guy with no hops.
Also, can we talk a little bit more about how a dunk is typically the highest percentage shot in a player’s repertoire? Outside of, “Dunks are easy points,” this doesn’t get mentioned at all. I’ve seen plenty of coaches get upset when their players put the ball off the glass to finish a break. If a player can dunk well, it’s a much better option when close up because it allows the player to follow through completely. Outside of other factors*, teams that shoot a high percentage generally beat teams that shoot at a lower clip, and since dunks are likely very high percentage shots, they help teams win. It’s not rocket surgery.
*E.G. Being coached by Don Nelson
Teams with better/more athletic players win more often.
This analysis is incomplete. Basketball is a game of applied statistics. Check the box scores for NBA games the team with the highest shooting percentage wins the majority of NBA games. A disparity of free throws, rebounds or turnovers might overcome a lower shooting percentage but in most cases the team that shoots a higher percentage wins. Dunks are a very high percentage shot but then again so is a lay up. The key is generating a large number of high percentage attempts, which usually happens through steals or very good field goal defense and rebounding. For Harvard students I am disappointed by your lack of depth in your analysis.
Not saying you have to necessarily be tall to dunk, but did you give any consideration to avg team height in this study? I feel like that could be a factor too (ie Steve Nash isn’t very tall and therefore can’t dunk). Does team height lead to more dunks, which in turn could lead to wins?
Interesting stuff presented here, it has my brain asking a lot of questions.
Interesting stuff presented here Teams with better/
Sure, it’s fine to correlate dunks with wins. It’s just pretty useless.
The “educated guesses” are even worse:
“The threat of the dunk – through posting up or through driving – is almost as important as the act of dunking itself. It opens up the playing field and spreads the defense, lending itself to open shots and open passing lanes. Another, more controversial view in light of the recent Donaghy game-fixing scandal: perhaps the NBA promotes the advancement of “exciting” more marketable teams by fixing games in their favor. In other words, is it possible that the correlation between dunks and victories does not occur by coincidence but by design? We’d need to do a close examination of foul calls to be sure.”