Driving around the streets of Los Angeles (for the final month – more on that later) and listening to AM radio as I like to do, I can’t help but notice all the hand-wringing on sports talk radio about the Lakers and how poorly they are playing. Much of the angst centers around the fact that during the NBA off-season, the Lakers loaded their team with top talent, adding All-Stars Dwight Howard and Steve Nash to the roster.
Of course, the city of Los Angeles expected wins out of the starting gate, but that got me wondering – is that a reasonable expectation?
Thinking back on championship teams of the past – Jordan & Pippen’s Bulls, Shaq & Kobe’s Lakers, and even LeBron & D-Wade’s Heat – these teams didn’t win in the first year they were formed. It took them time to figure out their system, to play as a unit.
This made sense to me as a theory, but as I tried to think of a way to test the theory with data, no pre-existing metric seemed to apply. So, I came up with the following two new metrics (this may have been done elsewhere, I just couldn’t find it):
- Total Team Tenure – How many minutes have the players on a roster played for their current team, combined? This metric spans all seasons – if a player has been playing for the same team for the past 4 years, all of the minutes in those 4 years count to the team’s total.
- Total Newcomer Minutes – How many of a team’s minutes are going to players in their first year on the team? This metric only applies to minutes played in the current season, by definition.
Visualizing the correlation between these two metrics and winning percentage so far in the 2012-2013 season:
Findings – Total Team Tenure (TTT)
- The team with the highest TTT, the San Antonio Spurs (74K minutes combined), is also the team with the highest winning percentage so far (82%).
- The team with the 3rd lowest TTT, the Washington Wizards (6K minutes combined), has the lowest winning percentage so far (17%)
- The Lakers are underperforming for their current TTT (52K minutes combined, of which 43K are contributed by Kobe alone)
- The overall trend suggests that in general, every 1,000 minutes extra tenure affords a team a winning percentage 2 percentage points higher
The lesson for GMs: Get top talent, but let them play together for a couple years before firing the coach (the Lakers let their coach go after the fifth game of the season)
Findings – Total Newcomer Minutes (TNM)
- The best teams – the Spurs, Grizzlies, Heat & Thunder – have the fewest TNM (less than 1,000 minutes for each team as of 12/12/2012)
- The worst teams – the Raptors, Wizards & Hornets – have relative high TNM (all above 2,000 for each team as of 12/12/2012)
- The Mavs are an outlier here – although they have the highest TNM by far (over 3,500 minutes going to players in their first year on the Mavs) their winning percentage is relative high – 52%. They completely “re-tooled”, and have done well considering 11 of their 16 players are brand new to their team.
- The Lakers are pretty much right where they should be considering how many minutes are going to newcomers
Causation or mere Correlation?
This all seems to imply that keeping players on the same roster for a long time increases your chances of winning. But couldn’t we also make the reverse argument: that teams that win a lot are more likely to keep the same players, and teams that lose a lot are more likely to overhaul their rosters. I can see the sense in this, but this isn’t always true – veterans like Dirk Nowiski retire, and championship-hungry teams like the Lakers re-tool even after high winning percentage seasons.
Before you make the case that bad teams should always be dismantled, just remember that Kevin Durant only won 20 games in his first season in 2007-08 for the Seattle Supersonics. The Supersonics moved to Oklahoma City the following year and became the Thunder, and Durant won 23 games that season in spite of adding Russell Westbrook to the roster. Follow the next few years of Kevin & Russell’s career, and you’ll get my point.
My take-away – let me wax philosophical
Sorry L.A., but this is how it should be. The fact that the Lakers can’t just dominate the league like that, presto – chango, is okay. We live in a world of microwave popcorn, instant oatmeal, ATMs, and “On-Demand” everything. But, like wine, many things get better with age. By the looks of it, this applies to basketball teams as well.
Thanks for stopping by,
Ben
Good idea, but I think your TTT metric can be skewed as exemplified by the Lakers: as you say their TTT is 52K minutes combined, of which 43K are contributed by Kobe alone.
If one guy stays in town for a decade but every other guy on the team is new, the minutes accrued in those 9 seasons don’t really contribute to a team understanding.
Otoh, two freshmen Knicks that end up in (say) Miami take that rapport along with them.
So imho… and this probably isn’t a simple thing to calculate…. TTT should be the total time players have shared a court for the same team. Have fun adding that up…
Side note – I now feel the need to go and do the same thing for the English Premier League… thanks! 🙂
Hi Andi – Awesome, glad I inspired you – I’d love to see the results for EPL. From your earlier comment, by the way, I totally agree with you. This is essentially a network problem: what is the strength of the team’s network based on time played together. It’s relatively simple to add up total time played on the team, but how to calculate minutes played with each other on the court at the same time? That was how I wanted to approach the problem at the beginning. TTT is a proxy for that metric.