If the Miami Heat end up "hot" going into the playoffs, that won't be a reason for their potential success.
As we gear up for the stretch run of the 2012-13 NBA season, don't fall into an all-too-common trap. Thinking a team has an advantage because they've been "hot" going into the postseason is only going to steer you down an unfortunate path.
It happens every year.
Some team starts beating its opponents with alarming frequency as the 82-game stretch draws to a close. Then, the media and fans in general start overvaluing its playoff chances. Surely this momentum is going to carry over past the regular season.
Well, it doesn't.
The playoffs, as Kobe Bryant might say, are the same beast but a different animal. Rotations thin out, the physicality of the contests increases and the intensity is amped up at least a few notches.
Teams like the 2011-12 San Antonio Spurs have bucked the trend and carried a lengthy streak into the postseason, but they won on their own merits. A nonphysical entity like momentum can't be credited for their playoff successes.
Tim Duncan would probably shake his head at you for discrediting his accomplishments.
Let's take a look at how playoff teams over the last 10 years have fared. The following graph shows a team's winning percentage over the last 10 games of the regular season on the x-axis and the number of playoff wins on the y-axis:
Obviously, the data points with 16 playoff wins are the ones who won championships dating back to the 2003 NBA Finals. As you might notice, their winning percentages leading into the playoffs aren't exactly clumped together toward the right half of the graph.
There's a slight positive correlation present here among all 160 teams, but the R^2 value, which represents how well the best-fit line approximates the data, is only 0.03895. A strong trend will typically have a value upward of 0.6, so this one falls well short.
Essentially, there's little to no correlation.
Additionally, it should logically follow that teams with a higher winning percentage over their past 10 games will have more success in the playoffs. It's just that those teams also tend to have higher percentages over the course of a season. They aren't necessarily "hot."
So what exactly is this hotness that we like to talk about? Shouldn't it be relative to a team's typical performance?
To answer that question, let me turn to the Socratic method and pose one more inquiry. Which hypothetical team is hotter: Team A, which goes 9-1 (90 percent) over the last 10 games to finish 65-17 (79.3 percent), or Team B, which goes 9-1 (90 percent) over the last 10 games to finish 41-41 (50 percent)?
Each team finished the season with an identical winning percentage over the last 10 games, but Team B should be considered much hotter because it drastically outperformed the expectations levied upon it by its performance prior to the "hotness."
With this theory in mind, I looked at the "heat" of a team, which is defined quite simply as the difference between a team's winning percentage over the last 10 games and the overall winning percentage. Using Teams A and B once more, the former would have a "heat" of 10.7, while the latter would check in with a "heat" of 40.
Team B is clearly the hotter squad leading into the playoffs. So does that mean it's bound to have more success?
In order to determine that answer, let's look at one more graph. While this y-axis still represents the number of playoff victories, the x-axis now shows the "heat" of a team:
Interestingly enough, seven of the 10 championship-winning teams over the past 10 years have had negative "heat" values, meaning that they were actually on a cold streak.
The worst of the bunch were the 2009-10 Los Angeles Lakers, who stumbled into the postseason with a 4-6 record to finish the year at 57-25. Conversely, the team with the greatest "heat" value was the 2003-04 Detroit Pistons squad, one that finished the season with a 54-28 record after going 8-2 down the stretch.
Once more, there's little to no correlation shown by this data. R^2 is a staggeringly low 0.01518, which indicates that the data points are more chaotic than anything. If there is a trend, it's a negative one, suggesting that teams actually benefit from cooling off at the end of the regular season before starting the most crucial part of the year.
To simplify this data, here's one more chart. This time, the axes are reversed so that playoff wins appears on the x-axis and "heat" on the y-axis. Additionally, there are fewer data points because the average "heat" for each number of playoff wins was found:
There's a slightly stronger negative trend here, though it's still a statistically insignificant one. We're also dealing with an admittedly small sample size.
For example, only one team has won exactly eight games over the past 10 years: the 2002-03 Detroit Pistons.
We can attempt to solve this problem by grouping together teams by the number of series they won. Essentially, we're meshing together teams with zero-to-three wins (x = 0), four-to-seven wins (x = 1), eight-to-11 wins (x = 2), 12-to-15 wins (x = 3) and 16 wins (x = 4).
Let's see what that does:
Now we have something to work with, as R^2 has jumped up to a statistically significant 0.85133. There is now a strong negative trend.
That's saying that teams advancing further into the playoffs tend to be cold at the end of the regular season. Don't make the mistake of implying causation from correlation, though. This is simply saying a correlation exists. Nothing more and nothing less.
Important to note is that there's no evidence whatsoever that being "hot" going into the playoffs leads to an increased level of success. If anything, the opposite is true.
So next time you're discussing what will unfold during the most dramatic part of the season, don't be that guy or girl who says, "Well, [insert team here] got hot at the end of the regular season, so their momentum is going to carry them through the playoffs."
Instead, be the one who calls out that person on their bluff. The end of the regular season shouldn't matter much to a contender, unless they're fighting for a certain seed.
Where they register on the "hotness" thermometer should be irrelevant.