FanPost

Are head to head regular season results predictive of success in 4/5 playoff matchups?

Andrew Richardson-USA TODAY Spor

Today we're promoting Eric's post that researches the history of 4 vs 5 matchups, and the relevance of both home-court advantage and that season's head-to-head games, when predicting the result of the playoff series. Good stuff! -- Tim

In recent discussions the idea has been thrown around that for any given playoff matchup the team with a better regular season head to head (h2h) record wins the vast majority of the time (something like 86%?) This number, however, ignores the fact that a 1/8 matchup will, presumably, follow different patterns than a 4/5 seed. For instance, in a 1/8 matchup, the outcome of the season is far less variable and, presumably, the 1 seed generally holds the regular season h2h advantage as well. The trends of the more competitive 4/5 matchups might be obscured by the other more lopsided series.

My question: Based solely on prior 4/5 playoff series, what are the Blazers chances against the Rockets? And Washington’s chances against Chicago?

To answer the question I used basketball-reference.com to collect the outcome, h2h results, total regular season wins, and playoff matchup results. They are as follows:

Year

4 (record)

4 team

4 h2h

5 (record)

5 team

5 h2h

H2H Cum

Wins comp

Series W

#game

2003

48

PHI

1

47

NOH

2

-1

1

4

6

2003

51

MIN

2

50

LAL

2

0

1

5

6

2004

42

MIA

3

41

NOH

1

2

1

4

7

2004

55

SAC

1

52

DAL

3

-2

3

4

5

2005

47

CHI

1

45

WAS

2

-1

2

5

6

2005

58

DAL

2

51

HOU

2

0

7

4

7

2006

50

CLE

1

42

WAS

3

-2

8

4

6

2006

60

DAL

3

49

MEM

1

2

11

4

4

2007

49

CHI

3

44

MIA

0

3

5

4

4

2007

52

HOU

1

51

UTA

3

-2

1

5

7

2008

45

CLE

2

43

WAS

2

0

2

4

6

2008

55

HOU

1

54

UTA

2

-1

1

5

6

2009

47

ATL

3

43

MIA

1

2

4

4

7

2009

54

POR

1

53

HOU

2

-1

1

5

6

2010

50

BOS

3

47

MIA

0

3

3

4

5

2010

53

DEN

3

53

UTA

1

2

0

5

6

2011

52

ORL

1

44

ATL

3

-2

8

5

6

2011

55

OKC

3

50

DEN

1

2

5

4

5

2012

50

ATL

1

48

BOS

2

-1

2

5

6

2012

51

MEM

1

50

LAC

2

-1

1

5

7

2013

49

BKL

1

45

CHI

3

-2

4

5

7

2013

56

LAC

3

56

MEM

1

2

0

5

6

2014

48

CHI

1

44

WAS

2

-1

4

???

???

2014

54

HOU

3

54

POR

1

2

0

???

???

Notes: I started with 2003 because that is the first year the NBA expanded the first round series to 7 games. Lockout year win totals are converted based on win%. Seeds are corrected so that team with home court advantage (HCA) is listed as 4, even if they were technically a 5 seed; this designation is more functional than the official seeding (e.g. doesn’t matter that a team is a division winner – they still didn’t have HCA). H2H Cum is the outcome of the season series with a negative value indicating the 5 seed won more games. Wins comp is the difference between the teams for regular season wins.

First, the good news for the Blazers and Wizards: The 5 seed has won half (11/22) of all 7 game 4/5 series in NBA history. This indicates that HCA generally has little or no impact on the outcome of these series. Additionally, for series which the 5 seed won, the regular season win totals were significantly closer than for series when the 4 seed won (stats people: means were 1.9 and 4.5, respectively, with a p-value of .04).

The 4/5 matchup has only yielded two sweeps since 2003, and one of those was a 60 win Dallas team against a 49 win Memphis team. That series has the biggest regular season win disparity of any 4/5 matchup since at least 1984 (second highest is the 1995 SEA/LAL series with 9), and ultimately led to changes in playoff seeding rules. In short, there has only been one true sweep in the 21 4/5 matchups since 2003 so it seems very unlikely that the Wizards or Blazers will be swept this year. That being said, no 5 seed has ever won in fewer than 6 games (8 won in 6 games, 3 in 7 games), so predicting a Blazers or Wizards win in anything less than 6 games is likely too optimistic.

And now, the bad news for Blazers fans: Of the 19 series that did not have a tied regular season h2h series, the h2h winner has prevailed has prevailed 74% of the time (14 of 19 series). That would suggest that the regular season matchup is highly predictive of first round playoff success for 4/5 matchups, and favors Washington and Houston in these series. For series in which the 4 seed has a 3-1 h2h advantage (i.e. Houston/Portland), they prevailed 4/6 times, while the 5 seed has prevailed 5/6 times when holding a 2-1 h2h series win (i.e. Chicago/Washington); again, these results suggest that, using historic results alone, Houston and Washington should be the favorites this year.

For Blazers fans who want to take an optimistic approach, there is an odd trend that may give you hope. Previously, 4/5 seeds have finished with identical records only twice: 1) Denver/Utah in 2010, 2) Clippers/Memphis in 2013. In both these cases the teams had high win totals (53 and 56, respectively) and the 4 seed held a 3-1 h2h series victory; these stats very closely match this year’s Portland/Houston series. In both cases 2010 and 2013, the 5 seed pulled off the upset in 6 games.

Interestingly, when I extended the data to include the 5 game 4/5 series, which started in 1984, very little changed. The 4 seed prevailed 45% of the time (27/60), however, the h2h regular season winner only won the playoff series 54% of the time (25/46). The NBA’s expansion of the first round series to 7 games seems to have added an advantage for teams that fared well in the regular season but has not achieved the league’s stated goal of protecting higher seeds (unstated goal: increased revenue). This trend may not be true for other series. I did not compare series length as that change was too dramatic to draw any meaningful conclusions. I will note that 50% (3/6) of the swept series prior to 2003 were won by 5 seeds.

Fun anomalies: worst 4/5 upset probably belongs to the 1989 Atlanta Hawks who won 52 regular season games and held a 6-0 h2h advantage over the 49 win Bucks; the Bucks prevailed in 5 games. A 5-seeded Utah team also lost to San Antonio in 1994 after completing a 5-0 regular season h2h sweep. The west, as suspected, has been highly dominant since 2003, with western 5 seeds averaging three more wins per year than eastern 4 seeds; west 5 seeds would have been 4 or higher in the east in all but two seasons since 2003. This trend is far older than most of us have realized – the east has not had a 4 seed with more wins than a western 4 seed since 1989(!), and since 1992 the eastern 5 seed has had the better record only once (1997).

Tl;dr: HCA is basically meaningless for 4/5 matchups, but h2h regular season records are predictive in 74% of previous series. If you’re going on historical precedent alone, pick the Wizards and Rockets this year. Blazers fans can, however, hold out hope that this series matches similar outlier series from 2010 and 2013.