By: Louie Vicchiollo
With the beginning of free agency less than 24 hours away, NBA fans can start to get excited. Hell, they should already be excited; Chris Paul is a Rocket, Phil Jackson isn’t a GM anymore, rumors continue to fly, and most importantly, Jimmy Butler is a Timberwolf.
Yep, it still feels really good to be able to say that. Just checking.
July 1st is arguably one of the most exciting, fun, nerve wracking days of the entire year when it comes to the NBA, especially with this year’s underwhelming playoffs. Free agency begins, players begin to get snatched up, and every fan can wonder what moves their team will make before all of the big name players are gone. The beginning, the big name, superstar, rumor filled players are the most exciting, but that does not mean the end of free agency doesn’t matter. You could be one Andre Igoudala away from making the playoffs, or one Paul Millsap away from being a contender. Now, I’m not saying Ben McLemore matters just as much in free agency as Blake Griffin, but that isn’t to say he won’t have a significant impact on the team that signs him.
Teams often dramatically change during free agency, and this is what I want to dive a little bit deeper into in this piece: what effects do major changes to a roster have on winning? Roster continuity is an amazing stat provided by Basketball-Reference.com (which is where all the statistics in this piece come from) that details what percentage of minutes played by a team come from the players on the previous year’s team.
A huge part of winning basketball games is having good team chemistry, which can only be developed the more you play together. This is why we see teams like the 2010-11 Miami Heat start 9-7 when they added LeBron and Bosh or the 2014-15 Cleveland Cavaliers starting 5-7 with the return of LeBron and addition of Kevin Love; it takes time to get used to playing together, that is just the nature of basketball. Draft picks, free agency, retirement, injuries and more can affect team chemistry and roster continuity in major ways.
So what effect does this roster continuity have on winning historically? Since the 2000-01 season, the average roster has had a roster continuity value of 66%, meaning that 66% of the minutes played that year were played by the players on the same team’s previous year’s roster. Check out the graph below detailing every team since the 2000-01 season and how their roster continuity and win percentage relate.
The trend is pretty clear: the higher the roster continuity %, the more likely you are to win. It also makes sense that the average team with a roster continuity rating of 66% should win 50% of its games, which the graph indicates. Now of course, this graph isn’t perfect. There isn’t a perfect relationship between roster continuity and win percentage, but it is very interesting to look at. Let’s look at a few of the outliers, indicated by the green data points and labeled.
Starting with the most fun one, the 2015-16 Warriors. Ah, yes, the 73 win team. While it should never be understated how difficult it is to win 73 games in a season, when you look at roster continuity it makes sense why they won as much as they did. They had 95% rating for their roster continuity that year, only adding Ian Clark, Anderson Varejao, Jason Thompson, and Kevin Looney’s minutes to the previous year’s roster. It probably didn’t hurt that the Splash Brothers hit a combined 678 threes on the way to Steph’s second, and unanimous, MVP. Give us a reigning champion, the reigning (and about to be back to back) MVP, and a 95% roster continuity rating and watch the wins pour in. Like I said, while 73 wins is amazing, it isn’t too surprising when looking at the 2015-16 Warriors.
The 2006-07 Seattle SuperSonics (Marked OKC on graph) should have been really good if we are going off the roster continuity prediction, so what happened? They should have won 49 games but they only won 31?! Injuries. Ray Allen only played 55 games, and Rashad Lewis only played 60. It should also be noted this isn’t come in and shoot four threes Ray Allen that we saw late in his career, this is 26 points per game, four and a half rebounds and assists per game, shooting 44 percent from the field Ray Allen. Losing your top two players to injuries won’t help you win games.
Now, the champion 2007-08 Boston Celtics. With a roster continuity rating of 50%, how did they win more than even half their games, let alone a championship?!? Remember that guy who we just talked about on the Seattle SuperSonics who got hurt when they were supposed to be really good? He joined the Celtics. Yeah, roster continuity can work the other way too. With the addition of Ray Allen and, sigh, Kevin Garnett, two superstars at the time to the already very strong core of Rondo and Pierce, the Celtics were able to overcome the low roster continuity rating with their superstar talent. ANYTHING IS POSSIBLE!!!!
Finally, this is a Wolves blog, so let’s talk about the 2003-04 Timberwolves. Similar to the 2007-08 Celtics, the Timberwolves added some really good talent this year in Latrell Sprewell, Sam Cassell and Fred Hoiberg (yes, I’m letting my Iowa State bias come into play here). Sam Cassell finished the year with 12.1 win shares, Sprewell with 5.7, and “The Mayor” with 6.1. Much like the 2015-16 Warriors, it probably didn’t hurt to have your best player playing in MVP form. This is yet another example of how talent can outplay the low roster continuity when you have a high level of talent.
So, now that we have looked at the individual seasons of every team since the 2000-01 season, you might be thinking “sure, low roster continuity can be outplayed by talent for a season” or “just because you have high roster continuity for a season doesn’t mean you will be good” but that doesn’t mean roster continuity means anything over longer periods of time (which, of course, is what roster continuity is measuring, continuity over longer periods of time) so how about we look at the averages for every franchise since 2000-01? Check out the graph below.
Hmm… what a surprise, the higher the roster continuity rating, the more likely you are to win games even over multiple seasons. This shouldn’t be a surprise, but I do want to mention a few quick things before we leave this graph for the internet archive to swallow up. Miami, Dallas, and San Antonio all have win percentages way higher than what their roster continuity would suggest, and Minnesota, Washington and Oklahoma City all are significantly below what we would expect. What could be the cause of this? I don’t have an exact answer, but one thing that I think is interesting is the front office of each of these teams. The teams above the prediction curve all have front offices that are widely regarded as great, and the teams below… well… but the front office is already indirectly involved in the continuity ratings because they are the decision makers in what players are on the teams, so is there another factor that changes this dramatically?
One thing that isn’t factored into roster continuity is coaches, and they play a big, big role in it. Like I mentioned before, roster continuity is calculated by the percent of minutes played by players from the previous year’s team on the current year’s team; a change in coaching can drastically change who plays, how much they play, and how often they play. Even if you only bring in three new players, a new coach could play those three players much more than the previous coach would, or vice versa.
Long story short, coaches affect minutes a lot (who woulda guessed?)
Let’s look at another graph, this time how the number of coaches for every team since 2000-01 has affected the win percentage since that season.
Another clear indicator of how our response variable (win %) changes with a new explanatory variable (number of coaches from 2000-17). You probably didn’t need the graph to guess, but the more coaching changes you make, the less likely you are to win. This is another factor that can be forgotten when thinking about team chemistry, not only do the player on the court have to gel together, but the coaches and players have to make a good connection both on and off the court.
There is one thing that needs to be said not only about coaching changes, but about this whole roster continuity concept in general: try to be like the Spurs. One coach since 2000 and an average roster continuity rating of 78% is a formula for a really good team for a really long time. Granted, they have had some okay draft picks and their coach is decent, but regardless this is just another thing the other teams in the league should take note of that the Spurs do and follow suit.
Yes, free agency is exciting. Really, REALLY exciting, and it should be! Your favorite team, the hometown team that you know and love could land a superstar, or add “that one piece we need to get to the next level”, but that doesn’t mean it is all sunshine and rainbows. The more you change your roster, the less likely you are to win, and that why you pay the front office the big bucks to know when to keep someone, or when to let them go. I don’t think anyone can argue the sheer awesomeness of the Jimmy Butler trade (exception: Bulls fans), there is no denying that and I won’t even try to. But I will also be the first to say that attempting to keep your roster continuity rating high will almost always help, and it should be done. With all of that said, I have a prediction for you, Timberwolves fans.
We will be just fine. No, better than fine! We will be good, potentially really good.
Remember that guy who got hurt for the 2006-07 Seattle SuperSonics, then joined the championship 2007-08 Boston Celtics? Or how about the guy who took the 2003-04 Minnesota Timberwolves and their 27% roster continuity rating and gave them 58 wins? Jimmy has the talent to do what they did. Be excited Timberwolves fans, Jimmy Butler is a top 15 player who can help the Wolves reach that next level, who can help bring the Wolves to the playoffs, who can help bring Minnesota a championship.