Monday, April 21, 2014

NBA 2013-2014: First Round Playoff Matchups - Analysis Leveraging Performance & Power Ratings

Here is our continuation to yesterday's article that discussed final power & performance ratings by team.  In this 2nd part we are going to break down each of the first round series using our metrics & data to try and give you that critical edge when you commence your handicapping process.



First let’s dissect the Western Conference:


ACTUAL
REGRESS
VARIANCE
WEST
Wins
Wins
Wins
San Antonio
62.0
56.8
5.2
Oklahoma City
59.0
54.3
4.7
L.A. Clippers
57.0
59.4
(2.4)
Houston
54.0
48.4
5.6
Portland
54.0
43.5
10.5
Golden State
51.0
55.0
(4.0)
Memphis
50.0
48.2
1.8
Dallas
49.0
51.4
(2.4)

First Round matchups

#1 San Antonio vs. #8 Dallas: according to this matrix (and many of the other metrics & statistics throughout this analysis also support this thought) Dallas and San Antonio only had a 5.4 game difference in projected record (Spurs were 2nd in NBA while Dallas was 6th in projected record) – significantly less than the actual 13 games difference their records indicate.  With the Spurs, similar to the Heat, we do have to mentally/qualitatively adjust their statistical performance based on the fact they often rest players during the season, and typically start each year off slower than others teams since they are an aging team and winning another championship is of the utmost importance to their franchise.  Despite this smaller than expected variance the Spurs are a MUCH better team defensively, especially EFG% as they check in at #2 in the NBA while the Mavericks are just #24.  Unless Dallas significantly ups their defensive performance – which would lead to exposing some of the weaknesses of the Spurs like offensive rebounding – San Antonio should not have much of a problem disposing of Dallas – but the opening lines on these games might be a touch high even though the Spurs are much like the Patriots in the NFL – seems a good idea to either play on them or pass; fading them is very tough.

#2 Oklahoma City vs. #7 Memphis: the Grizzlies have been one of the hottest teams in the NBA since the calendar turned to 2014, and did beat the Thunder in last year’s Western Conference semifinal round – but do they have a decent shot at pulling of the upset in back to back years?  Remember Russell Westbrook was injured in the first round of the 2013 playoffs and did not play in these teams meeting last May.  Both teams overachieved record wise compared to what their regression based record suggested – especially the Thunder checking in at +4.7 despite Westbrook missing extended action this winter.  Both of these teams are strong defensively checking in at 5th & 8th respectively – but there is a big discrepancy between offensive EFG% with OKC having the edge as the 6th best in the league vs. Memphis rating 18th.  Since Memphis is one of the worst teams in the league at getting to the FT line, and because the Thunder have revenge on their minds in this series, I do not see many ways the Grizzlies take this series deep despite their recent hot run – and despite what some media analysts believe.  They just are not efficient enough offensively, and OKC’s defense is underrated.

#3 LA Clippers vs. #6 Golden State: this heated rivalry features a pair of underachieving teams as our model projected 59.4 wins for the Clippers (57 actual) & 55.0 wins for the Warriors (51 actual) which would have checked them in at 1st and 4th in the entire NBA respectively.  Many certainly do not realize how solid the Warriors play defense as they are ranked 3rd in the NBA in EFG% defensively, while the Clippers are also a solid 5th.  Both offenses also have top 8 EFG%’s, however, their defense in that metric is better than their offense – suggesting unders may be solid plays here especially with the higher totals posted & the public generally backing the over between these two teams.  The biggest edge between these teams and the possible deciding factor may be the frequency each team reaches the FT line: LAC are 2nd in the league while GS is a miserable 28th – which puts more pressure on the Warriors shooting, which often isn’t a good thing especially how perimeter based they are.  In Game One LA shot 10 more FT’s (and almost hit that well known mark of making more FT’s than your opponent attempts 23 to 25 yesterday) but still took the loss.  It’s likely the lines on these games are too inflated on the Clippers side – look for a closely contested series between two of the better teams in the NBA.

#4 Houston vs. #5 Portland: this is a meeting between the two most overachieving clubs in the Western Conference as the 4th seeded Rockets won 5.6 more games than my model suggests & the 5th seeded Blazers won an incredible 10.5 more games than their statistical performance suggested!  Truly amazing how many “extra” games these teams managed to win despite numerous injuries during the season to key players like Dwight Howard & Lamarcus Aldridge.  These teams are very identical in many of the statistical areas I track and use for my ratings, but like many of the matchups discussed above one team has the significant edge offensively – and that is Houston.  Not only is Houston 3rd in offensive EFG% they also get to the line the most of any team in the league – which gives them flexibility in scoring points if they are having an off night from the field.  Their only weakness is turnovers where they rate worst in the NBA – but unfortunately for Portland they rate worst in the NBA defensively in forcing turnovers – so this seems to be a bad matchup for the Blazers.  Considering how poor Portland plays in the playoffs annually, how they should have been closer to a .500 team vs. winning 50+, and how the statistics matchup between these two squads, I see a significant edge for the Rockets here – and barring injuries do not see many scenarios where they would lose this series.

Now let’s take a look at the Eastern Conference:


ACTUAL
REGRESS
VARIANCE
EAST
Wins
Wins
Wins
Indiana
56.0
47.1
8.9
Miami
54.0
56.4
(2.4)
Toronto
48.0
49.6
(1.6)
Chicago
48.0
42.3
5.7
Washington
44.0
48.7
(4.7)
Brooklyn
44.0
40.4
3.6
Charlotte
43.0
42.4
0.6
Atlanta
38.0
36.2
1.8

First Round Matchups

#1 Indiana vs. #8 Atlanta: right off the bat the number that sticks out in the entire Eastern Conference is the Pacers winning 8.9 more games than my model suggests their statistics say they earned – big red flag there, as if they needed another big red flag against them with how poor they have played down the stretch.  According to the regression model the Pacers were the 4th best team in the East, and their play over their last 30 games or so certainly supports that thesis.  The only positive for the Pacers is they are facing a Hawks team that posted a terrible 38-44 regular season mark, which should have been 36-46 according to my model – one of the worst team’s record wise to ever reach the playoffs.  This matchup will likely come down to can the Hawks score consistently enough against the #1 EFG% defense in the NBA?  The Pacers are solid across the board defensively – which may be enough to get them past the Hawks, but any further there will be significant resistance.  Expect a lot of low scoring games between these squads in a series that is likely to go deeper and be more competitive than any Pacers fan thought was possible earlier this season.

#2 Miami vs. #7 Charlotte: Miami was just 2.4 wins short of their expected total, while the Bobcats were right on their number – keep in mind like the Spurs above the Heat routinely sat players during the season to give them rest as their aging stars look to win their third straight championship.  Charlotte is one of the worst shooting teams in the NBA amongst playoff teams, while Miami is the best – having such extremes in such a critical statistic doesn’t bode well the for the Bobcats.  And that is really the determining factor in this series – not only do the Heat own the Bobcats since Lebron James came to South Beach, but they also have all the playoff experience and can score much easier than the Bobcats can.  I have been riding the Charlotte train a lot this season, and would have given them a shot against any other team in the Eastern Conference besides Miami.  The Heat, again like the Spurs, know how valuable time off can be, so look for them to make short work of Charlotte – 5 games max.

#3 Toronto vs. #6 Brooklyn: in actuality there is only a 4 win difference between these two squads, but my regression model suggests their records were closer to 10 games difference according to their performance in my statistical package.  Like many of the other aging, veteran teams discussed above, the Nets certainly lost a TON of starter games due to injury or rest – which should be accounted for – as should their dearth of playoff experience, especially when compared to the Raptors.  That certainly slims the record margin down a lot – but statistically speaking the Raptors played better this season, and were another team we cashed with early & often this NBA year.  Almost the entire variance between these two teams projected wins was driven by their rebounding numbers – with the Raptors enjoying a clear edge, especially since the Nets lost Lopez for the season due to injury.  Despite that statistical edge for the Raptors they will have to up their game more than that if they hope to beat this veteran squad looking to make a deep run in the playoffs.  Unless Toronto absolutely dominates the glass, especially on the offensive end – which they didn’t accomplish in Game One leading to a loss – Brooklyn is definitely the favorite here.

#4 Chicago vs. #5 Washington: this matchup features teams that are almost a mirror image of each other according to the projected wins model – Chicago won 5.7 more games than projected (48 vs. 42.3) while Washington won 4.7 fewer games than projected (44 vs. 48.7) – so what does that mean?  Well I think most have recognized the Bulls have certainly overachieved this season – to my eye they struggled right after losing Rose for the 2nd straight season because of the sheer shock of another injury to their star player – but once they got over that, and adjusted their playing style, they excelled.  Washington on the other hand surprisingly won 44 games this season, and probably to the surprise of many people should have won closer to 49 according to my model – which was actually a higher projection than Indiana.  Many of the stats I track in my model are very close between these teams as their net EFG% is almost 0.00 for both teams (although Washington shoots a higher % they also allow a higher %), but the biggest difference is turnovers.  Washington, with John Wall has a big edge at the PG spot and taking care of the basketball compared to the Derrick Rose-less Bulls.  This series will likely go deep and come down to can Chicago take care of the basketball AND get to the FT line enough to support their weakness from the field.

Email me directly: boss@thesportsboss.com
Visit my website: www.thesportsboss.com
Follow me on Twitter: @SportsBoss
Follow me on Facebook: The SportsBoss
Leave comments here on the blog

COPYRIGHT: THE SPORTSBOSS, 2014

Sunday, April 20, 2014

NBA 2013-2014: Performance Ratings & Power Ratings FINAL FOR REGULAR SEASON





OFFENSE

DEFENSE

TOTAL
Team
SUM
RANK

SUM
RANK

SUM
RANK
Minnesota
36
2

53
10

89
1
Oklahoma City
48
8

43
5

91
2
L.A. Clippers
34
1

59
17

93
3
Portland
36
2

57
14

93
3
Houston
37
4

58
16

95
5
San Antonio
64
15

35
2

99
6
Memphis
58
13

44
8

102
7
Toronto
47
6

56
13

103
8
Charlotte
68
18

38
4

106
9
Chicago
76
24

33
1

109
10
Golden State
69
20

43
5

112
11
Indiana
79
26

36
3

115
12
Phoenix
47
6

68
18

115
12
Washington
73
22

43
5

116
14
Cleveland
66
17

53
10

119
15
Sacramento
49
9

72
19

121
16
Miami
65
16

57
14

122
17
Dallas
51
10

78
24

129
18
Denver
57
12

73
21

130
19
Atlanta
80
27

52
9

132
20
New Orleans
45
5

90
28

135
21
Detroit
56
11

81
25

137
22
New York
61
14

77
23

138
23
Orlando
84
29

55
12

139
24
Brooklyn
68
18

72
19

140
25
Boston
81
28

75
22

156
26
L.A. Lakers
78
25

85
26

163
27
Utah
73
22

91
29

164
28
Milwaukee
71
21

95
30

166
29
Philadelphia
103
30

88
27

191
30
               
Similar to my NFL Performance Ratings these break each team down using the metrics & statistics I have tested and found to be of the highest correlation to team success.  One aspect to keep in mind is the ratings above are best when lower compared to NFL ratings that are best when higher.  We can use these ratings numerous ways, identifying teams that may be over or under valued, or also for total players identifying which matchups could produce more or less points.  Last key item of note is these ratings weight each stat equally, where below in my projected wins & power ratings they are properly weighted to their impact based on statistical testing.  Using an equal weight for each statistic can add a lot of value to your handicapping efforts when properly applied.

For this final version let’s add a matrix that shows where each team ranked in the above non-weighted performance ratings throughout different points of the season:










ASB




19-Dec
26-Dec
2-Jan
9-Jan
15-Jan
23-Jan
27-Jan
2-Feb
9-Feb
13-Feb
5-Mar
20-Mar
FINAL
Atlanta
12
12
9
16
15
10
14
15
17
17
17
21
20
Boston
25
26
26
27
28
27
27
27
27
26
26
26
26
Brooklyn
23
24
24
21
24
22
22
22
22
22
23
23
25
Charlotte
7
5
9
9
10
9
10
10
11
12
10
11
9
Chicago
16
15
11
13
10
12
13
11
10
8
9
11
10
Cleveland
26
22
22
20
21
19
18
20
21
21
15
13
15
Dallas
21
20
20
24
22
23
23
20
20
18
22
20
18
Denver
10
12
17
10
9
13
11
14
14
15
19
18
19
Detroit
13
10
16
22
19
16
20
18
15
13
16
16
22
Golden State
18
19
12
13
12
13
15
12
12
10
10
8
11
Houston
9
9
8
7
8
7
7
6
4
6
3
5
5
Indiana
1
1
2
2
2
3
5
7
7
7
6
8
12
LA Clippers
8
7
4
5
6
6
5
5
5
5
3
2
3
LA Lakers
27
27
28
26
26
26
26
26
26
26
28
27
27
Memphis
19
18
13
11
12
10
9
9
9
10
12
8
7
Miami
16
11
14
13
18
19
17
19
22
22
19
19
17
Milwaukee
29
29
29
30
30
30
30
30
30
29
29
29
29
Minnesota
2
1
1
1
1
1
1
1
1
1
1
1
1
New Orleans
11
17
15
19
19
21
18
17
17
16
21
22
21
New York
22
24
25
24
23
25
24
24
24
24
25
24
23
Okla City
4
3
7
5
3
4
4
2
3
4
2
3
2
Orlando
25
23
23
23
25
24
25
25
25
25
24
24
24
Philadelphia
28
27
27
28
27
28
29
29
29
30
30
30
30
Phoenix
20
20
18
18
16
13
12
13
13
14
13
13
12
Portland
7
6
5
4
3
4
3
4
2
2
5
3
3
Sacramento
14
12
18
11
12
17
21
23
19
20
18
13
16
San Antonio
7
3
6
8
7
8
8
8
8
9
8
6
6
Toronto
4
8
3
3
3
2
2
2
5
2
6
6
8
Utah
30
30
30
29
29
29
28
28
28
28
27
28
28
Washington
17
15
21
16
17
18
16
16
15
18
13
16
14

Next up let’s take a closer look at the above matrices, and provide some color behind a few teams that “stand out” for seemingly rating higher or lower than their final record.

Minnesota: the Timberwolves have been ranked #1 for 12 of the 13 editions we published including each of the last 12 since Christmas.  Why did a mere 40-42 team rate so highly?  The answer is very simple – Minnesota was very solid in all categories that comprise this analysis besides the biggest one on each side of the ball: Effective FG%.  The Wolves were ranked 21st offensively & 27th defensively in that area, which is the most heavily weighted statistic when using linear regression as I do below.

Portland & Houston: these teams are matching up in the opening round of the Western Conference playoffs, and were ranked tied for 3rd & 5th respectively in these ratings – very solid obviously.  Both teams had Top 4 offenses and mid ranked defenses – look for a lot of points, with turnovers (especially for Houston both sides of the floor) & offensive rebounding being key drivers to the winner of this series.

Western Conference: of the 15 teams in the conference 7 rated #1 thru #7 here, and just 2 rated amongst the bottom 9 teams – shows how strong that conference is.

Toronto, Charlotte & Chicago: here is a trio of teams we made a lot of cash backing this season, and they check in at #8, #9 & #10 overall – the top 3 teams in the Eastern Conference.  Although they are the top 3 teams in the conference using the equally weighted method they slip in the linear weighted model because of average to poor EFG%, especially offensively.  These teams, especially the latter two, are best known for their work on the defensive end – which is where they will no question need to excel in order to advance in the playoffs.

Brooklyn: the Nets check in at #25, easily the lowest rated team in the playoffs this season – even the lowly Hawks are ranked #20!  What does it all mean?  Brooklyn is very much a middle of the road team AT BEST when using full regular season statistics as even their EFG% numbers are middle of the pack – and they frankly do not excel in any one area besides possibly defensive pressure/forcing turnovers (which was assisted by the Livingston addition to the starting lineup in early calendar year 2014) and getting to the FT line offensively – both areas they leveraged to pick up a win in Toronto in Game One.

Next let’s dive into the Power Ratings which use the Performance Rankings as a base:


HOME
ROAD

Power Rating
Power Rating
L.A. Clippers
117.0
113.0
San Antonio
116.2
112.7
Miami
116.0
112.5
Golden State
115.5
112.5
Oklahoma City
116.2
112.2
Dallas
114.7
111.7
Toronto
114.3
111.3
Phoenix
114.2
111.2
Washington
113.3
110.8
Houston
114.5
110.5
Memphis
113.1
110.1
Indiana
113.7
109.7
Minnesota
112.2
109.7
New York
111.6
109.6
Portland
112.7
109.2
Charlotte
111.9
108.9
Chicago
111.8
108.8
Brooklyn
110.9
107.9
New Orleans
110.3
107.8
Detroit
109.5
107.5
Atlanta
109.6
107.1
Denver
109.4
106.9
Cleveland
108.8
106.8
Sacramento
108.6
106.6
Boston
108.5
106.5
Orlando
106.5
104.5
L.A. Lakers
105.6
103.6
Utah
105.4
103.4
Milwaukee
105.2
103.2
Philadelphia
102.0
100.0

These are actually Power Ratings that can be used when handicapping nightly matchups.  

Keep in mind however that when the playoffs commence ratings become more hyper-sensitive to variance in performance, and are adjusted slightly from what you see above.  But these can still be used as a solid estimate – let’s take a look at projected lines for Game One of each series versus the actual line posted in Vegas:

[#1] PACERS       -6.5       [#8] hawks                          Vegas line: -8                     Variance: 1.5 points
[#2] HEAT            -7            [#7] bobcats                       Vegas line: -9.5                 Variance: 2.5 points
[#3] RAPTORS    -6.5        [#6] nets                           Vegas line: -3.5                 Variance: (3.0) points
[#4] BULLS          -1            [#5] wizards                       Vegas line: -4.5                 Variance: 3.5 points

[#1] SPURS         -4.5        [#8] mavericks                  Vegas line: -9.5                 Variance: 5.0 points
[#2] THUNDER   -6            [#7] grizzlies                     Vegas line: -8                     Variance: 2.0 points
[#3] CLIPPERS    -4.5        [#6] warriors                      Vegas line: -7.5                 Variance: 3.0 points
[#4] ROCKETS    -5.5        [#5] blazers                      Vegas line: -5                     Variance: (0.5) points


What initially really jumps off the page is the fact 6 of the 8 lines (and another is just 0.5 points below & may still move on Sunday to positive territory) are HIGHER on the favorite in Vegas compared to my ratings - which makes perfect sense since Vegas routinely adjusts odds on favorites up based on the public typically backing the favorites.  For Saturday’s games my ratings suggested plays on Hawks, Raptors, Grizzlies & Warriors – 3 of 4 underdogs; the pair of 3pt variances split while the two smaller variance plays also split.  However, remember, line value is only one small piece of the handicapping puzzle – and unlike the regular season there are small adjustments made to these ratings in the playoffs designed to further accentuate the performance fluctuation by team.

Lastly let’s take a look at wins by team in three critical buckets: ATS Win %, Final Wins, Regression Wins.  Also we will examine the variance between actual & regression wins:


ATS WIN %

FINAL ACTUAL WINS

REGRESS WINS

ACTUAL - REGRESS WINS

Wins
RANK

Wins
RANK

Wins
RANK

Wins
RANK
L.A. Clippers
0.568
4

57.0
3

59.4
1

(2.4)
21
San Antonio
0.549
5

62.0
1

56.8
2

5.2
9
Miami
0.463
24

54.0
5

56.4
3

(2.4)
20
Golden State
0.519
10

51.0
8

55.0
4

(4.0)
15
Oklahoma City
0.538
6

59.0
2

54.3
5

4.7
11
Dallas
0.537
7

49.0
10

51.4
6

(2.4)
22
Toronto
0.582
3

48.0
11

49.6
7

(1.6)
25
Phoenix
0.642
1

48.0
11

49.0
8

(1.0)
27
Washington
0.531
9

44.0
14

48.7
9

(4.7)
12
Houston
0.500
13

54.0
5

48.4
10

5.6
6
Memphis
0.456
25

50.0
9

48.2
11

1.8
24
Indiana
0.469
20

56.0
4

47.1
12

8.9
2
Minnesota
0.476
18

40.0
17

45.4
13

(5.4)
7
New York
0.463
23

37.0
19

43.9
14

(6.9)
4
Portland
0.537
7

54.0
5

43.5
15

10.5
1
Charlotte
0.595
2

43.0
16

42.4
16

0.6
28
Chicago
0.506
12

48.0
11

42.3
17

5.7
5
Brooklyn
0.512
11

44.0
14

40.4
18

3.6
17
New Orleans
0.487
17

34.0
21

37.7
19

(3.7)
16
Detroit
0.432
29

29.0
23

37.6
20

(8.6)
3
Atlanta
0.469
20

38.0
18

36.2
21

1.8
23
Denver
0.476
18

36.0
20

35.4
22

0.6
29
Cleveland
0.488
15

33.0
22

34.3
23

(1.3)
26
Sacramento
0.468
22

28.0
24

30.7
24

(2.7)
19
Boston
0.488
16

25.0
26

28.1
25

(3.1)
18
Orlando
0.430
30

23.0
28

27.5
26

(4.5)
14
L.A. Lakers
0.500
13

27.0
25

21.8
27

5.2
10
Utah
0.434
28

25.0
26

20.4
28

4.6
13
Milwaukee
0.444
26

15.0
30

20.3
29

(5.3)
8
Philadelphia
0.439
27

19.0
29

19.5
30

(0.5)
30

The first column ATS Win % is not only impacted by actual team performance but also certainly by other factors that are tough to measure but we know they exist such as expectations & public teams.  Team expectations at the beginning of the season certainly plays a critical role in ATS performance because perceived “good” teams people will continually wait on if they start slow (New York), while perceived “bad” teams people will continually wait and/or bet on to flounder after a hot start which sometimes doesn’t occur (Phoenix).  We also see teams such as Miami who checks in at #24 in ATS performance largely driven by the fact they are more actively managing the regular season grind compared to most other teams, with a clear eye on winning another championship.  Understanding these largely unmeasurable variables is absolutely critical to having success betting on the NBA because the season is such a long grind, and emotional levels, lineups, injuries, motivation and other human behaviors are key in the NBA compared to say the NFL where every team plays once per week and is typically motivated equally.

Perhaps the most critical piece of this final NBA analysis entry of the season is the far right column.  As a reminder let’s define regression wins: this figure represents how many games a team is projected to win based on their performance in specific statistical categories I have defined (with help from many other books & intelligent NBA executives) as having the greatest impact on winning games, properly weighted using linear regression.  The statistics & weights will remain for my eyes only as that is proprietary, but we can discuss the results in some detail that will provide insight into my process to the astute reader.  One item of note is the rank column for the actual minus regression wins metric uses absolute values of the variances to better represent which teams were furthest or closest to actual wins.  Based on this metric let’s take a closer look at the playoff teams from each conference, and possibly identify teams that are over or under valued based on their true statistical performance this season.  In total of the 16 playoff teams 10 posted a record better than their true performance suggested (5 in each conference) – but who has the biggest variances?  

CHECK BACK TOMORROW FOR A DETAILED ANALYSIS OF EACH PLAYOFF SERIES.


Email me directly: boss@thesportsboss.com
Visit my website: www.thesportsboss.com
Follow me on Twitter: @SportsBoss
Follow me on Facebook: The SportsBoss
Leave comments here on the blog

COPYRIGHT: THE SPORTSBOSS, 2014