Welcome to the debut of Stat-town Royalty.
I'm not a professional statistician, and I don't advocate that statistics are the answer to everything. This isn't stats vs. the eye test. Both of these methods of observation are fraught with subjectivity and inaccuracies, as both are--at some point--interpreted by human perception, and that perception introduces biases. So, while this (and hopefully these) article is called Stat-town Royalty, it only uses statistics as a tool to find the truth--not as the end-all-be-all answer to all of our questions and curiosities. This article (and any more to come) is inspired by the excellent discussion that takes place here at Sactown Royalty. So please, keep it coming! Now lets dig into them stats.
The assist is an elementary basketball statistic. In my 20 plus years of being a basketball fan, and therefore looking at basketball statistics and box scores, the assist has always been a core stat. When people list off a players box score, its typically in a Points/Assists/Rebounds format. I guess you could call it one of the 'Big 3' basketball statistics. Well, the basketball stats world is a changin' and as it changes the assist is becoming more and more irrelevant. Thanks to things like Synergy Sports and NBA.com's stats page utilizing SportVU data, we get to see more usefull information about passing that assists used to imply. Assists, at least traditionally, imply good passing. Whether its individual passing (like Chris Paul or Rajon Rondo) or team passing (like the Spurs or the Kings in the good old days), the assist was how we gauged the best passers. There were additional implications as well, such as unselfishness. A player whom averaged a low number of assists and a high number of field goal attempts was undoubtedly a selfish, ball-hogging black hole on offense. Conversely, a high assist player always had great court vision, was excellent in the half-court offense, was unselfish, and made his teammates better. Are those supposed axioms always true? Can you think of a time when maybe your eyes told you a player moved the ball well and was unselfish--yet the statistics said he was a ball-hog?
The Assists impact on team performance
This is the question that inspired this article: Do teams that generate more assists win more basketball games? It's a question that in one way or another has been asked by commentators here on Sactown Royalty. Well we all know that when the Kings were riding a wave of success during the dawn of the century, they were playing a style of basketball that not only generated copious amounts of assists, but also captivated the world of basketball fans with its fluid passing and unselfish team-first play and spawned a generation of Kings fans. Now here's the problem; I'd love to see another team like that. Skilled passing is fun to watch. But I also like to watch winning basketball, and the statistics say that a team does not have to generate assists to generate wins. Have a scatter chart friends:
What you're looking at is a scatter plot representation of a data set. The data set is a team's assist per field goal made and their wins for the 2013-2014 NBA season. Assists per field goal made gives us a pace adjusted way of seeing how often a team's made shot was assisted. For example, the Kings made 3026 field goals this year and had 1547 assists giving them an assist to field goal ratio of 51%. From the plot above you can see that this is well below average, as the average is around 58% (if you're curious about the actual data, you can find a link to the data set I used at the bottom of this article). However, what we want to know is if there is a correlation between assists per field goal made and wins. Perhaps you see one on the plot, perhaps you don't. Either way, statisticians have a way of objectively determining whether there is a correlation between two sets of data--the correlation coefficient. Its a fairly complex formula to find this value, but luckily spreadsheets can do it for us. The r value (correlation coefficient) will fall between -1 and 1, with numbers closer to 0 being the least correlated. For these data sets, the r value is around 0.14, which is considered to be a weak correlation. So, we have determined that in the sample size that we have (the 2013-2014 season), having an offense that generated more assists per shot did not inherently mean your team won more games.
What does this mean for the Kings?
In Akis's recent article about the Kings improvements despite the same win total as last year, he mainly focused on what has improved, leaving out some of the things that declined. As pointed out in the comments of that article, assists declined. Dramatically. However, now we know that a decline in assists doesn't mean a whole lot in terms of our team getting better or worse. This would also be the case if their assists went way up--although that would certainly be more fun to watch. The bottom line is that unlike rebounds, which are a zero-sum statistic, having more assists doesn't have much of an impact on winning. Even the correlation value for a simple thing like total rebounds and wins comes back with a (0.49) moderate to strong correlation. I'd love to see the Kings move back to a motion offense with lots of pick and rolls and back cuts and shooters running without the ball, but I also would like a team that wins games and makes it into the playoffs. If I had to choose between having an aestetically pleasing team and a winning team; I'm choosing the winning team every time. Yet, if I had to chose between two winning teams; one ISO oriented and one motion oriented on offense, I'm going to pick the motion offense every time.
Thank you for reading, and until next time!
References, Links, and Data:
- Thanks to Basketball Reference for the raw stats.
- My Spreadsheets used for this article can be found here
- For those interested, here is some info on correlation coefficients