Detroit Pistons are the NBA’s most bizzaro team

Nov 1, 2014; Auburn Hills, MI, USA; Detroit Pistons owner Tom Gores speaks to the media in a press conference before the game against the Brooklyn Nets at The Palace of Auburn Hills. Mandatory Credit: Tim Fuller-USA TODAY Sports
Nov 1, 2014; Auburn Hills, MI, USA; Detroit Pistons owner Tom Gores speaks to the media in a press conference before the game against the Brooklyn Nets at The Palace of Auburn Hills. Mandatory Credit: Tim Fuller-USA TODAY Sports /
facebooktwitterreddit

The Pistons are off to a very strange start this season. They have taken down several strong teams while struggling in what should have been cakewalks. Of course, that happens to some degree for all teams every season, so I decided to look into just how strange the Pistons’ play really has been.

In general, the best guess at teams’ odds of winning any given game are the betting markets. We can say that with confidence because if there were any significantly better method, whoever had access to it would be raking in a lot of money–and taking it from Vegas.

This chart shows the outcomes of every NBA game this season against their projected outcomes. Since it shows every game from both teams’ perspectives, each game is counted twice and the overall chart displays rotational symmetry. Obviously, there have been many surprising outcomes. But, as is to be expected, there is significant correlation between projections and results. If that wasn’t the case, every better could easily beat the system simply by picking the underdog consistently.

nba results v lines
nba results v lines /

The Pistons’ chart, on the other hand, looks almost the opposite. Points in the top right are the games in which they were the underdogs and did in fact lose. Points in the lower left (I know, there are none) are games in which they were the favorites and proceeded to win. The bottom right are games where they upset another team and the top left are games in which the Pistons were upset by another. So the positive-positive and negative-negative quadrants are the ones for which Vegas picked the correct winner and the other two quadrants are games Vegas got wrong.

pistons results v lines
pistons results v lines /

The Warriors game (in which the Pistons were projected to lose by 12.5 and lost by 14) was really the only one that pulls Detroit far from almost perfectly flipping the script in every game. Without that game, the Pistons’ results look as opposite to projections as any team is aligned to them.

pistons results v lines no gs
pistons results v lines no gs /

Of course, every team’s results look weirder if you remove the most reasonable ones from the sample, so let’s see if the Pistons stand out not only from the league as a whole but from other individual teams who also have the potential for fluky results from a small sample.

You may note that each of the charts above include a trend line, the best linear approximation of the empirical data represented by the points. Those lines have two key variables. The higher the correlation coefficient (it is mathematically impossible to be higher than 1 or lower than -1), the more consistently being favored by a lot of points corresponds to winning by a lot of points and vice versa. The closer the slope is to 1 (not many will be higher, but they can be), the closer Vegas is to nailing the actual scoring margins.

TeamProjections vs. Reality SlopeProjections vs Reality CorrelationPercentage of Games Picked Correctly
Cleveland Cavaliers-0.413-0.26373%
Detroit Pistons-0.502-0.23720%
Philadelphia 76ers-0.485-0.210100%
Minnesota Timberwolves-0.496-0.19250%
Golden State Warriors-0.422-0.161100%
Houston Rockets-0.144-0.07942%
Boston Celtics-0.004-0.00236%
Atlanta Hawks0.0210.00957%
Portland Trail Blazers0.1970.06654%
Utah Jazz0.330.11460%
Chicago Bulls0.4250.13860%
Dallas Mavericks0.2740.15558%
Denver Nuggets0.4540.19250%
Brooklyn Nets0.3540.21775%
New Orleans Pelicans0.3620.26183%
Indiana Pacers0.6710.28158%
Miami Heat0.8210.28670%
Milwaukee Bucks0.8210.29073%
New York Knicks0.7160.32467%
Los Angeles Lakers0.7310.35564%
Charlotte Hornets0.9430.41883%
Orlando Magic0.5810.43775%
Toronto Raptors0.9050.50162%
Memphis Grizzlies1.6280.54683%
Washington Wizards1.8480.56578%
Phoenix Suns1.4470.59264%
Los Angeles Clippers0.8990.59980%
Oklahoma City Thunder1.7740.62473%
San Antonio Spurs1.1540.68091%
Sacramento Kings1.2880.70983%
NBA Overall0.7660.41667%
Pistons (omitting GS game)-2.071-0.79611%

The most important value there is the correlation. While the Cavaliers edge out the Pistons to date this season on that count, it is almost entirely because of their 30-point blowout over the Grizzlies. A bit like the opposite of the Golden State game for Detroit. Apart from that, they actually look a lot like a highly predictable team (as is borne out visually and in Vegas typically picking the correct favorite in their games).

cavs results v lines
cavs results v lines /

The Warriors and Sixers are highly predictably blowing everyone out and being blown out respectively. But when teams are doing so, the degree to which they do is relatively erratic. So their margins have been unpredictable even though the games themselves have been highly predictable.

dubs results v lines
dubs results v lines /
sixers results v lines
sixers results v lines /

Really, the Timberwolves and Rockets are the only other teams that are off to bizzaro starts. And neither of them to nearly the degree of the Pistons. Meanwhile, the Kings have been the most predictable team thus far and the Celtics the most random.

kings results v lines
kings results v lines /
celtics results v lines
celtics results v lines /

The Pistons have played a relatively tough schedule thus far and performed very admirably against it. While there is no real reason for their current trend of playing well against good teams and poorly against bad ones to continue, let’s all hope that it doesn’t. Because the slate should mostly get easier from here on. If the Pistons keep playing antithetical to projections, that will probably result in about the most frustrating possible .500 record.

*note: a previous version of this article had slight errors in the tabulated data for Boston, Brooklyn, and New Orleans and misnamed the Hawks as the owners of the most random results to date