If you enjoyed Part I of our NFL Turnover Analysis, then you will especially enjoy Part II as it has even more strongly correlated data which can be of great assistance when projecting the final standings for the upcoming season – and if your team has a chance at reaching the playoffs and ultimately winning the Super Bowl.
In this part we want to focus on turnovers themselves – actual TO Margins (TOM) year to year – without making any additional assumptions such as how much each may be worth, and how a points margin metric correlates to actual win/loss records. What we want to drill down on and really tune into is comparing year over year (YOY) changes, and how these figures tell such a massive part of the story as far as if a team will improve or regress season over season in wins. But please keep in mind when reading all these statistical analyses I post here – these can only be used as a piece of the puzzle, and not the entire puzzle to project anything. Think of the exercise of analyzing stats for any sport as a 1,000 piece jigsaw puzzle – each stat is a little piece that goes towards putting together the entire puzzle, but there are a lot of pieces that need to be put in place to get the entire picture.
For this analysis I will only focus on two scenarios, but that means 4 years of data (2008, 2009, 2010, 2011) as the main portion of this part will focus on YOY comparisons.
2008
|
2009
|
YOY
|
2009
|
2010
|
YOY
|
2010
|
2011
|
YOY
| ||||||
ARI
|
0
|
(7)
|
(7)
|
(7)
|
(5)
|
2
|
(5)
|
(13)
|
(8)
| |||||
ATL
|
(3)
|
3
|
6
|
3
|
14
|
11
|
14
|
8
|
(6)
| |||||
BAL
|
13
|
10
|
(3)
|
10
|
7
|
(3)
|
7
|
2
|
(5)
| |||||
BUF
|
(7)
|
3
|
10
|
3
|
(17)
|
(20)
|
(17)
|
2
|
19
| |||||
CAR
|
6
|
6
|
0
|
6
|
(8)
|
(14)
|
(8)
|
1
|
9
| |||||
CHI
|
5
|
(6)
|
(11)
|
(6)
|
4
|
10
|
4
|
2
|
(2)
| |||||
CIN
|
0
|
0
|
0
|
0
|
(8)
|
(8)
|
(8)
|
0
|
8
| |||||
CLE
|
5
|
(12)
|
(17)
|
(12)
|
(1)
|
11
|
(1)
|
1
|
2
| |||||
DAL
|
(11)
|
0
|
11
|
0
|
0
|
0
|
0
|
6
|
6
| |||||
DEN
|
(17)
|
7
|
24
|
7
|
(9)
|
(16)
|
(9)
|
(10)
|
(1)
| |||||
DET
|
(9)
|
(18)
|
(9)
|
(18)
|
4
|
22
|
4
|
11
|
7
| |||||
GB
|
7
|
24
|
17
|
24
|
6
|
(18)
|
6
|
24
|
18
| |||||
HOU
|
(10)
|
(1)
|
9
|
(1)
|
0
|
1
|
0
|
7
|
7
| |||||
IND
|
9
|
2
|
(7)
|
2
|
(4)
|
(6)
|
(4)
|
(12)
|
(8)
| |||||
JAC
|
(8)
|
2
|
10
|
2
|
(15)
|
(17)
|
(15)
|
5
|
20
| |||||
KC
|
5
|
1
|
(4)
|
1
|
9
|
8
|
9
|
(2)
|
(11)
| |||||
MIA
|
17
|
(8)
|
(25)
|
(8)
|
(11)
|
(3)
|
(11)
|
(6)
|
5
| |||||
MIN
|
(6)
|
6
|
12
|
6
|
(11)
|
(17)
|
(11)
|
(3)
|
8
| |||||
NE
|
1
|
6
|
5
|
6
|
28
|
22
|
28
|
17
|
(11)
| |||||
NO
|
(4)
|
11
|
15
|
11
|
(6)
|
(17)
|
(6)
|
(3)
|
3
| |||||
NYG
|
10
|
(7)
|
(17)
|
(7)
|
(3)
|
4
|
(3)
|
7
|
10
| |||||
NYJ
|
(1)
|
1
|
2
|
1
|
9
|
8
|
9
|
(5)
|
(14)
| |||||
OAK
|
1
|
(13)
|
(14)
|
(13)
|
(2)
|
11
|
(2)
|
(4)
|
(2)
| |||||
PHI
|
3
|
15
|
12
|
15
|
9
|
(6)
|
9
|
(14)
|
(23)
| |||||
PIT
|
4
|
(4)
|
(8)
|
(4)
|
16
|
20
|
16
|
(13)
|
(29)
| |||||
SD
|
(2)
|
9
|
11
|
9
|
(6)
|
(15)
|
(6)
|
(7)
|
(1)
| |||||
SF
|
(15)
|
9
|
24
|
9
|
(1)
|
(10)
|
(1)
|
28
|
29
| |||||
SEA
|
(7)
|
(8)
|
(1)
|
(8)
|
(9)
|
(1)
|
(9)
|
8
|
17
| |||||
STL
|
(5)
|
(13)
|
(8)
|
(13)
|
5
|
18
|
5
|
(5)
|
(10)
| |||||
TB
|
4
|
(5)
|
(9)
|
(5)
|
9
|
14
|
9
|
(16)
|
(25)
| |||||
TEN
|
14
|
(4)
|
(18)
|
(4)
|
(4)
|
0
|
(4)
|
1
|
5
| |||||
WAS
|
1
|
(11)
|
(12)
|
(11)
|
(6)
|
5
|
(6)
|
(15)
|
(9)
| |||||
AVERAGE
|
13
|
(14)
|
13
|
(12)
|
13
|
(17)
| ||||||||
AVERAGE
|
(12)
|
17
|
(13)
|
13
|
(12)
|
14
| ||||||||
This data shows each of the 32 NFL teams, their TOM each season from 2008, 2009, 2010 & 2011, along with the grey columns which represent YOY change in TOM – these grey boxes will eventually become the crux of this analysis below. Next, let’s discuss the color coding: the numbers highlighted in green are teams that were +9 or better in TOM during a season, while the red highlights are for teams that were (9) or worse. Typically, teams that play with an extreme impact from TOM one season typically see that stat revert back closer to the mean in the very next season – and the grey box shows the change in the following season that each team had – for example, BAL shows a +13 TOM in 2008, which means we would expect to see a decline in 2009, so I maintain the same color coding in the grey column – each grey column’s color coding will be the same as the first year in each of the 3 boxes – to keep track of each team’s expected move. Now that we are all on the same page and you know what the data represents, let’s analyze it and show how it can add value – and also, how it sets the foundation for the true key portion of the analysis that will be discussed below. Here are the results from analyzing this data:
- 18 instances of a team having +9 or more TOM in one season (for the last three years ’08-’10), and all 18 teams had a decline in their TOM the next season. The average decline of the 18 teams was (14.3)
- 16 instances of a team having (9) or worse TOM in one season, and 14 (88%) of the 16 increased their TOM the following season. The average increase of the 14 teams that fit this scenario was +14.6
- Key take away from the straight annual analysis is you can estimate the TOM move YOY for teams that were +/- 9 or more the prior year to be +/- 14
As mentioned above in the second bullet point there were two teams over the last three seasons that did not have a change in TOM that we would have projected: DET in 2009 worsened by (9) compared to 2008, which was mostly driven by the fact the Lions played a rookie QB that season who was responsible for 24 turnovers on his own; DEN last season was the other outlier as they worsened by only (1) compared to 2010 due to similar reasons as DET – playing a first year QB who was responsible for 15 turnovers on his own. Based on those two outliers, one area to keep an eye on when attempting to use this data is the QB position, which is obvious in the sense it is the only position on the field that will handle the ball on every play – if a team is playing a QB that lacks experience and is getting his first true taste of the NFL, it’s very likely even though he could be successful (Stafford and Tebow), and turn into an All-Pro type signal caller eventually (Stafford), its often a rocky start in their first season.
As we can see by now in this analysis along with Part I, if a team is amongst the league leaders in one season they almost always revert backwards the next year – but by how much is the next question? We can use those average marks below each of the three seasons discussed above as a barometer for usage in any models, but we can also go another step forward and use YOY comparisons which will yield another estimate.
Let’s examine the YOY grey boxes from above: for this first example, we will compare the grey box from the 2008/2009 analysis to the grey box from the 2009/2010 comparison:
'09/'08
|
'10/'09
|
Variance
| |
ARI
|
(7)
|
2
|
9
|
ATL
|
6
|
11
|
5
|
BAL
|
(3)
|
(3)
|
0
|
BUF
|
10
|
(20)
|
(30)
|
CAR
|
0
|
(14)
|
(14)
|
CHI
|
(11)
|
10
|
21
|
CIN
|
0
|
(8)
|
(8)
|
CLE
|
(17)
|
11
|
28
|
DAL
|
11
|
0
|
(11)
|
DEN
|
24
|
(16)
|
(40)
|
DET
|
(9)
|
22
|
31
|
GB
|
17
|
(18)
|
(35)
|
HOU
|
9
|
1
|
(8)
|
IND
|
(7)
|
(6)
|
1
|
JAC
|
10
|
(17)
|
(27)
|
KC
|
(4)
|
8
|
12
|
MIA
|
(25)
|
(3)
|
22
|
MIN
|
12
|
(17)
|
(29)
|
NE
|
5
|
22
|
17
|
NO
|
15
|
(17)
|
(32)
|
NYG
|
(17)
|
4
|
21
|
NYJ
|
2
|
8
|
6
|
OAK
|
(14)
|
11
|
25
|
PHI
|
12
|
(6)
|
(18)
|
PIT
|
(8)
|
20
|
28
|
SD
|
11
|
(15)
|
(26)
|
SF
|
24
|
(10)
|
(34)
|
SEA
|
(1)
|
(1)
|
0
|
STL
|
(8)
|
18
|
26
|
TB
|
(9)
|
14
|
23
|
TEN
|
(18)
|
0
|
18
|
WAS
|
(12)
|
5
|
17
|
AVERAGE
|
(15)
|
23
| |
AVERAGE
|
14
|
(26)
|
This is where it get’s slightly complex, but read through, soak it in step by step using the information in the first matrix to help you understand this last one, and you will be able to use this information in your own handicapping and predictive efforts.
In the first column I have the YOY change in TOM from the 2008 to 2009 season – so I would have used this information heading into the 2010 season. The teams highlighted in red were teams that had a YOY +9 or more change from ’08 to ’09 – meaning, in football discussion terms, they increased their TOM performance significantly in 2009 vs. 2008. Color coding these teams red can be thought of as a potential “red flag” as they increased their performance so much in 2009 that in 2010 they were almost certainly going to experience the exact opposite – a red flag warning. The green colored teams represent the same terminology, except these teams were likely to experience a favorable impact from turnovers in 2010 based on their subpar performance in this area during 2009. Now, let’s discuss the far right column – the crux of this analysis – as that column represents the change in YOY change. Using ARZ as an example, they registered a 0 in 2008 followed by a (7) in 2009 which equals the (7) you see in the first column; the next column is calculated by taking the (7) they were in 2009 and the (5) they were in 2010, which equals the +2. Remember, we would not be able to calculate that last # prior to the 2010 season, but after it, we can and that is what I just explained. The black column on the far right is then simply the difference between the two columns to the left, for ARZ it is (7) to a +2 move equals an improvement of 9. Now that you understand the figures just look at how high the correlation is between the teams that are colored and what happened to their numbers – let’s look closer at the green teams. Heading into 2010 I expected all those teams to enjoy a favorable impact from turnovers, and look at the results: +21, +28, +31, +22, +21, +25, +23, +18, +17. Nine teams I projected to see a significant turnovers improvement in 2010, and look not only how high those corrections were, but how closely in range they are to one another – which shows how accurate, and consistent the analysis is, so attempting to use this method as an estimate for any team if they were a candidate in the coming season to shift one way or another is reliable. The red teams tell a similar story, with opposite results.
Now, let’s examine this same data, but using ‘09/’10 vs. ‘10/’11 YOY changes:
'10/'09
|
'11/'10
|
Variance
| |
ARI
|
2
|
(8)
|
(10)
|
ATL
|
11
|
(6)
|
(17)
|
BAL
|
(3)
|
(5)
|
(2)
|
BUF
|
(20)
|
19
|
39
|
CAR
|
(14)
|
9
|
23
|
CHI
|
10
|
(2)
|
(12)
|
CIN
|
(8)
|
8
|
16
|
CLE
|
11
|
2
|
(9)
|
DAL
|
0
|
6
|
6
|
DEN
|
(16)
|
(1)
|
15
|
DET
|
22
|
7
|
(15)
|
GB
|
(18)
|
18
|
36
|
HOU
|
1
|
7
|
6
|
IND
|
(6)
|
(8)
|
(2)
|
JAC
|
(17)
|
20
|
37
|
KC
|
8
|
(11)
|
(19)
|
MIA
|
(3)
|
5
|
8
|
MIN
|
(17)
|
8
|
25
|
NE
|
22
|
(11)
|
(33)
|
NO
|
(17)
|
3
|
20
|
NYG
|
4
|
10
|
6
|
NYJ
|
8
|
(14)
|
(22)
|
OAK
|
11
|
(2)
|
(13)
|
PHI
|
(6)
|
(23)
|
(17)
|
PIT
|
20
|
(29)
|
(49)
|
SD
|
(15)
|
(1)
|
14
|
SF
|
(10)
|
29
|
39
|
SEA
|
(1)
|
17
|
18
|
STL
|
18
|
(10)
|
(28)
|
TB
|
14
|
(25)
|
(39)
|
TEN
|
0
|
5
|
5
|
WAS
|
5
|
(9)
|
(14)
|
AVERAGE
|
(16)
|
28
| |
AVERAGE
|
15
|
(24)
|
This time we are familiar with what the figures & colors above represent, so let’s jump right into the analysis and see what the projections and results showed. Again, what we want to do is isolate teams with a mark of 9 or more either way – and of the 9 teams that fit the plus 9 or more premise in YOY move from 2009 to 2010, all 9 had a massive drop from 2010 to 2011 – an average decline of (24). There were also 9 green teams that had a negative move from 2009 to 2010 – and all 9 of those teams enjoyed an increase in the 2010 to 2011 metric, with the average being +28.
For the upcoming 2012 season, we can use the YOY comparison charts above to isolate teams that I expect to move significantly – all I am doing here for this list of teams is using the middle column from the last matrix, which shows the ‘11/’10 TOM move, and searching for +/- 9:
- YOY expected increase (favorable TOM impact): KC, NE, NYJ, PHI, PIT, STL, TB, WAS
- YOY expected decrease (unfavorable TOM impact): BUF, CAR, GB, JAC, NYG, SF, SEA
Now that we have another list of candidates from this type of turnover analysis, we can compare and combine it with Part I’s candidate list to see what we can expect in the 2012 season. Here is a list of the team’s that showed up in both Part I and Part II, along with some comments:
- Green Bay Packers (15-1): last season they enjoyed a whopping +24 TOM, which helped boost their record from the 11-5 their true performance indicated, all the way to losing only one game during the regular season. However, heading into 2012, both turnover analyses support a drop in their record – not that this one isn’t obvious, but there is an extremely high chance GB wins 14 or less. VERDICT: 14 OR LESS WINS
- Kansas City Chiefs (7-9): talk about an interesting team, the pair of turnover indicators point in different directions in 2012 – Part I of this analysis projected the Chiefs to drop from their 7 wins mostly because of their true performance, not so much the impact of (2) TOM. Part II above states KC should enjoy a significant increase to their TOM, which in turn would favorably impact their record. If my assumption above holds true and they increase their YOY TOM mark by the estimated 28 located on bottom of the last matrix, that would move their TOM in 2012 to +12 – all four teams last year that were + double digits in TOM reached the playoffs. But unless their true performance improves, it is unlikely TOM will be enough to push them into the playoffs. VERDICT: Neutral
- New England Patriots (13-3): the Patriots are like KC on steroids – it is true that NE enjoyed a nice +17 in TOM during 2011, but they also played extremely well as even normalizing their points margin for the impact of TO’s they still performed 6th best in NFL; including the TO impact their points margin was 3rd, only behind GB and NO. The data in this article suggests their TOM will increase based on the fact it dropped from +28 in 2010 to +17 LY – and any double digit drop in one season is very often followed by a double digit increase the next. NE will likely remain near the top of the NFL in TOM, but either way they will also likely play well enough that the impact of the TOM will not be critical to their success. VERDICT: Neutral
- New York Giants (9-7): the defending champs have enjoyed back to back seasons of an improving TOM, which as we talk to above is often a recipe for a reversion, especially when the jump is double digits. Last year the Giants, without the impact of a +7 TOM were around a .500 team, which would have excluded them from the playoffs. Heading into 2012, looking at the data above, and the data in the prior article, my theory that they will drop from 9 wins to .500 or worse is supported by both analyses. Combine this data with an improving division including a pair of talented, hungry and often underachieving teams in DAL & PHI, and it looks to be a down year and a missed playoffs season for the G-Men. VERDICT: 8 OR LESS WINS
- San Francisco Forty-Niners (13-3): SF is very similar to the NYG as far as TOM analysis goes heading into 2012. Both articles support a drop in wins for SF as no team enjoyed more of a favorable impact from TO’s last season than SF. They checked in at an incredible +28 on the season, leading the league; normalizing their points margin for this +7.00 impact yields a modest +2.44, which is equivalent to about a 10-6 type team (another way to look at it is last year’s +28 TOM was really worth approximately 3 wins.) It’s pretty safe to assume SF will suffer a double digit drop in their TOM this year, which would then lead to a drop from last year’s 13 wins. VERDICT: 12 OR LESS WINS
- Philadelphia Eagles (8-8): heading into last season very big things were expected from the Eagles, and it never materialized after a slow start including numerous inexplicable home losses as huge favorites. One major reason they suffered only their second .500 or worse season since the turn of the century was TOM – PHI was 3rd worst in the NFL at (14), with only TB and WAS worse. And boy did those TO’s have a HUGE impact on their won/lost record: removing the estimated (3.5ppg) TO impact from their points margin yields a true performance margin of 7.56, which would slide them into an approximate 12-4 team! They were tied with MIA last year for team’s that had the biggest difference between their true performance margin and actual record, 4 games under. Both of my turnover analyses suggest PHI will indeed improve their record this season, and in actuality if they get double digit favorable TOM this season I expect PHI will contend for home field advantage in the NFC Playoffs. VERDICT: 9 WINS OR MORE
- Washington Redskins (5-11): the Redskins have both TOM indicators pointing towards an improvement for the '12 campaign, but this is a perfect spot where we must revert back to the comment on rookie QB's - typically they are prone to more TO's than a more seasoned veteran which would have me concerned about drawing a conclusion that their TOM will flip next year despite the fact both indicators suggest that will be the case. This would be a team to keep an eye on especially if RG3 proves to be strong taking care of the football - I will call it neutral with a lean towards WAS improving to 6+ wins. VERDICT: Neutral
There are the seven teams that fit both turnover analysis articles I posted. Five of the seven teams had the same conclusions from both articles, firming my expectation those teams move in said direction sans WAS - I will cautiously lean towards a TO improvement for the Skins as both indicators suggest, but keeping an eye on the rookie QB's TO's will give us a better feel for their end of year result; two teams had mixed reviews, so no conclusion can be reached.
The beauty of this YOY comparison analysis is there are ZERO outliers to the underlying thesis – unlike purely analyzing TOM by the year. Don’t get me wrong, looking at TOM from a pure YOY angle definitely still adds value – only 2 outliers out of 34 instances is very solid. But, thinking outside the box some, and coming up with your own types of analysis that are rarely discussed and used by others is how you can truly find value, and add accuracy to your own modeling and handicapping.
When reading all the different types of analysis offered on sites all over the Internet remember two things: one, is it relevant or is it random correlation? Two, although a lot of analysis can add value to your decision making process, no piece of information alone should be the reason you formulate an opinion. Please, use all the resources that are available to you! There are lots of reading options these days all over the Internet – but make sure it fits the two criteria I mentioned above for it to truly add value to your own style of handicapping.
I will most likely be completing a Part III to this Turnover Analysis in the next few weeks, discussing the turnover impact on playoff berths. That analysis will tie together the results from the first two Part’s to the teams that earn, or fall out of a playoff spot year to year.
Thanks again for reading, please feel free to:
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, 2012
No comments:
Post a Comment