The Boston Celtics' Biggest Challenge

| by David Berri

Tball is an attorney practicing in relative obscurity in New Hampshire.  Beyond a demanding home life, an active practice, and occasional rambling on this site, he devotes an obscene amount of time volunteering at the Boys & Girls Club of Greater Nashua (online donations are accepted).  Likes include pop economics, behavioral analytics, and audio nonfiction.  Dislikes include irrational, close-minded, egocentric, sports journalism posers and light beer.  And please, if you have a criticism voice it – it may be helpful if I volunteer to do this again.

Dave Berri posited the other day that Jeff Ma’s intuition over statistics argument has led to Ma’s Smackdown success this year, but is hardly a formula for predicting a Celtics like upset in the future. Similarly, cherry picking a third of the season when a team played well and using that select performance to predict future performance — while ignoring two thirds of the season — is untenable except for rationalizing activity after the fact. I took that as a challenge to find a rational mathematical means of predicting the C’s advancing to the finals (*** Spoiler Alert *** I failed).

These Celtics have been compared to the Houston Rockets who won the title in 1995. That team rose from a six seed to defeat four very strong teams on their way to a championship. That year, The Dream was unavailable for a chunk of the season, but strong and healthy for the playoffs. In addition, Clyde Drexler was also unavailable for much of that regular season (because he was still a member of the Portland TrailBlazers). 

Similarly, Kevin Garnett and Paul Pierce were unhealthy for a chunk of the second half of the 2009-10 season. Furthermore, Doc Rivers has indicated he worked hard to rest his players and manage their health for the playoffs. Although a coach can manage minutes, we expect that once a player steps on the court they play to their ability. So while the Celtics point differential for the regular season may have been skewed by injuries, regular season WP48 should be relatively reliable.

So this line of thinking led me to the question, “How good would the Celtics have been this year if their players had been available for 82 games averaging the same minutes they are playing in the playoffs [using regular season AdjP48]?”  To answer that question, I give you Table 1:

Just to be clear… Table 1 looks at how many minutes each player on the Celtics logged in the playoffs.  It then allocates the players across the five positions. With positions assigned, a player’s WP48 is calcualted utilizing each player’s regular season ADJ P48 [to see how this is done, please look HERE].

After all of this we see that the Celtics project to be about a 60-win team if they could have maintained the health of this roster and played this rotation for 82 games. Of course, the playoffs are different than the regular season because of rest between games. Added minutes are also assigned to starters (normally 2-6/gm dependent on age and position) and the bench gets shortened, a process that hopefully eliminates the least productive players from the rotation. For an apples-to-apples comparison, I’ve completed the same exercise for the Cleveland Cavaliers, Orlando Magic, and LA Lakers. Again for each team, I have asked the following: “How would this team have performed this year if every playoff player was available for 82 games averaging the minutes they have averaged in these playoffs [using regular season AdjP48].

Pull up a chair, here is the analysis of the LA Lakers….

the Orlando Magic…

 and the Cleveland Cavaliers.

Using this system, you would not have predicted any of Boston’s upsets. So – as noted above – this approach didn’t quite work out as hoped. That being said, the difference between the Celtics and Magic is not quite as great as the season numbers (i.e efficiency differential and wins) would suggest. In other words, the Celtics certainly had a reasonable chance of upsetting Orlando.

Here are a couple other thoughts appeared while looking at these numbers. Shortening the bench should get your best players more playing time in the playoffs. However, there were exceptions to this rule in the playoffs.  For example, Varejao and Moon lost PT in Cleveland; although this change in minutes would not lead us to think Cleveland was not going to defeat Boston. In fact, no matter how you twist and cherry pick the numbers, these numbers cannot be arranged to predict a Cleveland loss in the playoffs, at any level. 

I also didn’t find much fault in LeBron’s play in the playoffs. For the sample size and higher level of competition, his playoff numbers do not seem to deviate significantly with his MVP performance in the regular season.  For example, on the days in which LeBron had more than 48 hours to rest (whether it was the elbow or something else), he posted a 0.614 WP48 [a mark estimated from LeBron’s Win Score].  Such a performance is beyond what we saw in the regular season. However, on the days in which LeBron had only 48 hours to rest, he only posted a 0.271 WP48.  While still very good, that LeBron would be produce about 10.5 fewer wins across a season. This would obviously make the Cavs on those days inferior to the Celtics. 

Turning to the NBA Finals, the above tables show the Lakers’ rotation to be worth 60.7 wins over the regular season and the Celtics’ rotation to be worth 59.8 wins. How close is that .9 wins?  About 2 extra minutes of KG in place of Davis is worth .9 wins or, if KG had played his average 34 minutes in the one playoff game he missed — replacing 34 minutes of Big Baby/‘Sheed — you’d have Celtics team expected to total 60.7 wins. In other words, very little separates the talent of these two teams (remember, defined according to the minutes the coaches assigned to allocated this talent). So if this method of predicting series outcomes were useful (huge if), I’d still have to say the NBA Finals is a coin flip.

- Tball

Editors Note: Tball did submit this before the NBA Finals began. But the WoW editor was a bit slow.  The minutes for the Lakers and Celtics have been updated for the first three games of the NBA Finals.