How unusual is the familiarity enjoyed by the Golden State Warriors' stars?
Warriors stalwarts Stephen Curry and Draymond Green have been playing together since Golden State drafted Green in 2013. The next season, the Warriors added Andre Iguodala, who is back in the Bay after a two-year hiatus. And Golden State's continuity will only get stronger when Klay Thompson returns to the lineup. His partnership with Curry, interrupted by Thompson's injuries the past two seasons, dates back to 2012.
Throughout the NBA season, I answer your questions about the latest, most interesting topics in basketball. You can contact me directly at @kpelton, tweet your questions using the hashtag #peltonmailbag or email them to [email protected].
This week's mailbag kicks off by quantifying the chemistry between Curry and Green as well as other sets of teammates and comparing the Warriors to the rest of the league. I also answer your questions about the relative odds of LeBron James' current and former teams making the playoffs, whether rematches within the same week like last week's pair of showdowns between Golden State and the Phoenix Suns are different from other meetings and players matching up against namesake teams.
"After watching the almost telepathic play between Draymond Green and Steph Curry [developed over 10 seasons together], I was wondering, is there a 'Familiarity Factor' that can be quantified? Can you calculate the number of years each player has played with players on their roster? Do teams and rotations with a higher 'Familiarity Factor' lead to more plus-minus, wins, or championships?"
-- Reggio Fox, New Zealand
I decided to go with games played together as the measure. You're right to highlight Curry and Green, who have played more often than any other two current teammates. They're also part of the next two partnerships on the list, both with an injured Thompson. And all three of them recur in the top 10 with Iguodala, who is back for a second stint with the team.
Naturally, the Warriors lead the way when it comes to total games together for all the sets of teammates on their current roster -- although perhaps not as dramatically as you would guess based on the first chart.
Although the Mavericks have no pair of teammates with more than 297 games together (Dorian Finney-Smith and Dwight Powell), their 23 sets of teammates with at least 100 games together are the most of any team.
As for the second part of your question, I think it suffers from the same issue as most measures of continuity: Are teams winning because they stayed together or staying together because they win? Curry and Green undoubtedly have great chemistry together, honed over years, but they were also awfully good together back in 2014-15 when they won their first title with Green in his first season as a full-time starter.
The Warriors' score did improve this season with the return of Iguodala. Is that the reason Golden State has played so well? I think it makes Iguodala more valuable to the Warriors than any other team, but it's probably not the difference maker there.
At the other extreme, the Chicago Bulls still have played fewer games together as teammates than the Curry/Green/Thompson trio alone has. Yet that hasn't prevented them from being one of the NBA's hottest teams to start the season.
@kpelton FiveThirtyEight now has that the Cavs have better odds (29%) to make the playoffs than the Lakers (27%). Do you think that's going happen and do you think that's the first time in LeBrons career that his previous team has a better playoff likelihood than his current one?
— RMJ=H (@rmj_equals_hero) December 7, 2021
Answering the second question first: If we're limiting to projections that factor in preseason expectations, almost certainly yes.
For projections like Basketball-Reference.com's that include only data from the current season, the 2014-15 Miami Heat probably would have outpaced LeBron's Cleveland Cavaliers when the Miami Heat were 8-6 and Cleveland 5-7 in late November. But anything that factored in higher preseason expectations for the Cavaliers would still have had them as the more likely playoff team, as justified by Cleveland going 48-22 the rest of the way while Miami went 29-39.
As I noted last week, projections using data from the current season only don't fare nearly as well early in the season predicting team performance the rest of the way as those that incorporate preseason expectations. In that piece, although the Lakers had the biggest drop-off from their preseason line and the Cavaliers the second-biggest increase, the Lakers were still comfortably projected as the better team going forward. Since then, Cleveland has expanded its sample size of playing well, although the Lakers have also gotten their two most lopsided wins of the season.
Much of FiveThirtyEight's Lakers pessimism stems from the fact that their statistical preseason projection (42 wins on average) was not nearly as hopeful as the team's over/under total (52.5). My subjective take probably falls somewhere in between those two extremes, in which case the Lakers are easily better than 50-50 to make the playoffs while the Cavaliers probably still haven't gotten to that point. I'd still give the Lakers better playoff chances myself.
"I was wondering if there is a higher chance that two teams split games when they're playing back-to-back [or within a week]. I got this impression last season when there were more back-to-backs to reduce travel and just got reminded by the Warriors and Phoenix Suns splitting games back-to-back."
-- Thilo
I did find in a mailbag last season that splits of the two-game "sets" in a single city the NBA used in 2020-21 increased compared to the same sample of games with more time in between in 2019-20.
What about all series in close succession, even when the venue changes between games, as was the case with the Suns and Warriors? To study that, I looked at all rematches between two teams over the course of the season and group by the previous outcome and the time separating the matchups. Here's how much better the winning team in the first game did in the second game by time in between.
As you'd expect, the winning team has an edge in the next game because it's typically the better team. However, the size of that edge does seem to vary depending on how quickly the rematch comes. The effect is modest as long as the games aren't too far apart, but on average, the previous winner has about a 10% edge in a back-to-back setting as compared to a 13% edge at the mean time between games (about 36 days).
That effect would make splits slightly more common in home-and-home series, or sets rather than matchups separated by an extended period.
#peltonmailbag so we know there's been Boston vs Boston, and I found someone named Dallas Comegys who played 8 games vs Dallas in the '80s. How many other name vs team matchups have there been in the NBA?
— Adam Reisinger (@AdamReisinger) December 9, 2021
When LA Clippers rookie Brandon Boston Jr. checked into Wednesday's game against the Celtics, Clippers broadcaster Brian Sieman pointed out it was Boston against Boston, and I was inspired to look up whether that had happened before. (It had, once: Lawrence Boston against the Celtics in 1980.) How common is a player matching up with his namesake team?
First, the ground rules: It has to be the full name of the opposing team. For example, Antonio Davis playing against the San Antonio Spurs doesn't count. And no middle names, ruling out John Houston Stockton against the Rockets. Excluding those, I found 16 examples, led by Allan Houston against the Rockets.
Remarkably, I only found one game in NBA history in which a player appeared for his namesake team: Stan Washington's lone career game with the Bullets in 1974.
I also found just one playoff matchup: Jim Washington against the Bullets in the 1975 conference semifinals.