Data Mining: WRs On New Teams

 

Advanced Data Mining with Brandie Searle 

 

Big Name WRs such as...

Brandon Marshall, Terrell Owens, Anquan Boldin and Torry Holt all moved on to a new team this season.

What does history tell us?

During the offseason, nothing gets a football fan more excited than having their team sign an impact player. Playoff teams have "just added that final piece of the puzzle" and losing teams have "turned a corner." 

Randy Moss going to NE in 2007 had a HUGE impact. But do things usually work out well for a WR on a new team?   

Hint: Don't count on it. 

Interestingly, Wide Receiver movement appears to be on the rise. With defenses becoming more complicated, every offense is trying to find those quick-strike weapons that can make an impact.

In this article, I will do some historical research for you.

The Data (1980 to 2006) 
For the purposes of this study, I started with a database of player statistics going back into the 1940's. Since we are looking only at Wide Receivers for this study, I discarded all the other positions, including Tight Ends. Since offenses have undergone some major changes over time, I also limited the study to seasons that occurred after 1980.

Next, I limited the data to the year immediately before the WR changed teams, and the year immediately after the WR changed teams. (Remember we are looking only at the Year-1 impact of player movement, not 2-3 years down the road.) 

Because this article is focused on fantasy football, I didn't really want to concern myself with players who only had a minimal impact in football. However, I also did not want to filter out players who may have benefited from a scenery change due to a different QB change or position on the depth chart. For this reason, I decided to limit the study to WRs who had at least 32 receptions the year prior to the trade. I felt this number eliminated the fringe players who had minimal impact on either team, while still allowing us to view players who may have shown significant improvement with their new team.

Finally, I also wanted to filter out scenarios that were out of the players control and are largely unpredictable. To this end, I filtered out scenarios where a player missed the entire following season due to suspension. Since I didn't have the data to know the difference between injury or sever underperformance, I also eliminated the scenarios where a player appeared in less than half their team's games the year following the trade. In cases where a player was traded the following year, I combined the performance from both their new teams. I also eliminated scenarios where a player changed teams for the last year of their career.

Once all these filters were implemented, we were left with 158 different cases where a WR changes teams. Note that if a player changed teams multiple times in their career, all of these changes were identified. So now our study can begin.

The Analysis 
I started the analysis by looking at the total population of wide receivers in our study. In the year prior to changing teams, our 158 receivers averaged 53 receptions, 725 yards and 4 touchdowns. These are reasonable (though not great) numbers for a third Fantasy wide receiver. 

When I looked at the post-year statistics, I was a little surprised. The post-year statistics for our population dropped to 42 receptions, 569 yards, and 3 touchdowns. 

If you�re looking at percentages, that is a 20% drop in receptions, a 21% drop in yards, and a 25% drop in touchdowns! Those were pretty significant drops in performance. 

I knew that there may have been a difference in games played between the two years, even with the previous filtering I had implemented, so I looked at the games between the two years. 

Here, there was a slight difference. The year prior to moving teams, our receivers averaged 15 games. The year after changing teams, our receivers averaged 14 games. This was only a 6% difference. 

So if we look at our total population, with only a 6% average reduction in opportunities, our receivers declined over 20% in the three main categories we care about as fantasy football players.

Raising the Bar 
While those results were interesting, I still wasn't comfortable that the entire population was really focusing on those prime receivers we really care about for fantasy football. As I mentioned previously, the average results for our receivers were 53 receptions, 725 yards, and 4 touchdowns. While these numbers are reasonable statistics for a receiver, since these numbers are averages, this means that half our population had less than these numbers. So I decided to raise the bar and use those numbers as the starting point for the next part of my study.

So I now filtered my study down to include any wide receiver that had statistics that exceeded at least two of those average numbers. This means that a receiver that had 800 yards receiving and 6 touchdowns, but only 40 receptions would still be included. Since our 53/725/4 were average numbers, I thought this would be the best way to filter down the receiver sample without eliminating too many data points. 

Using my new criteria narrowed our sample size down to 79 wide receivers. Since this cut our sample almost perfectly in half, I was confident that we now filtered out some under performing wide receivers that would have declined even without switching to a new team.

As I expected, the numbers for our new sample definitely improved and now better relect the type of wide receivers we might consider drafting.

BEFORE-MOVE AVERAGES 
Pre-Move Year Statistics (n=79 Wide Receivers)
1) Receptions = 63
2) Yards = 893
3) TDs = 5.5. 
4) 15.0 Games

Above somewhat mirrors a #2 fantasy wide receiver. 

With these results in place, I expected my comparison results (post-move year) to be much more favorable. I was completely mistaken. Instead of improving, our numbers became significantly worse.

AFTER-MOVE AVERAGES 
Post-Move Year Statistics (n=79 Wide Receivers)
1) Receptions = 47
2) Yards = 663
3) TDs = 3.7 
4) 14.3 Games

Wow, what a big drop! 

Receptions decreased by 25%, yards decreased by 25%, and touchdowns decreased by a whopping 30%.

Remove Injury Factor 
Now I began wondering what other variations I could eliminate to make sure I was looking at an apples-to-apples comparison between the two years to ensure that this data was meaningful. The next step I took was eliminating the injury factor between the two seasons. 

So I now eliminated all players who participated in less than 14 games in either the prior-move or post-move years. This reduced our sample size to 52 instances. In both pre and post move seasons, our players now averaged a little over 15.5 games. 

Even after eliminating the Injury Factor, things did not get much better for wide receivers playing on their new team.

BEFORE-MOVE AVERAGES 
Pre-Move Year Statistics (n=52 Wide Receivers)
1) Receptions = 67
2) Yards = 918
3) TDs = 5.3 
4) 15.5 Games

AFTER-MOVE AVERAGES
Post-Move Year Statistics (n=52 Wide Receivers)
1) Receptions = 54
2) Yards = 759
3) TDs = 4.2 
4) 15.5 Games

So now, even with eliminating the difference due to games played, we still see a reduction in statistics of 19% for receptions, 17% for yards and 20% for touchdowns.

Did Anybody Improve? 
Decreased wide receiver production was surprisingly consistent across all three steps of filtering I had completed so far. At this point, I began to wondering - which (if any) wide receivers had actually improved after switching teams?

Of the 51 players still remaining in our study, only 9 players actually improved their performance after switching teams. These 9 players were:

Prior To Move

 

After Move

 

Difference

Last_Name

First_Name

Year

Team

G

REC

YDS

TD

 

Year

Team

G

REC

YDS

TD

 

REC

YDS

TD

Moss

Santana

2004

nyj

15

45

838

5

 

2005

was

16

84

1483

9

 

39

645

4

Ellard

Henry

1993

ram

16

61

945

2

 

1994

was

16

74

1397

6

 

13

452

4

Rison

Andre

1989

clt

16

52

820

4

 

1990

atl

16

82

1208

10

 

30

388

6

Rice

Jerry

2000

sfo

16

75

805

7

 

2001

rai

16

83

1139

9

 

8

334

2

Fryar

Irving

1995

mia

16

62

910

8

 

1996

phi

16

88

1195

11

 

26

285

3

Martin

Tony

1997

sdg

16

63

904

6

 

1998

atl

16

66

1181

6

 

3

277

0

Fryar

Irving

1992

nwe

15

55

791

4

 

1993

mia

16

64

1010

5

 

9

219

1

Haynes

Michael

1993

atl

16

72

778

4

 

1994

nor

16

77

985

5

 

5

207

1

Owens

Terrell

2003

sfo

15

80

1102

9

 

2004

phi

14

77

1200

14

 

-3

98

5

Some interesting points about these players:
1. Only three of these moves occurred since 2000 (S. Moss, J. Rice, T. Owens) 
2. Irving Fryar actually appears on this list twice. 
3. Of the four times, Andre Rison switched teams in his career, only one time did he appear on our list for actually improving his statistics when he played a whole season. 
4. In only 1 season (1993) did more than 1 WR improve!

So of our 51 players, only 18% improved performance the year after moving to a new team. 

Just to validate this fact against our original data, I went back and looked at all the players that improved their performance a year after moving to a new team, regardless of games played, and regardless of whether they met our minimum criteria of 53 receptions, 725 yards, and 6 touchdowns. My thought was that we would pick up a number of players at the lower end of the spectrum who improved their mediocrity.

When looking at these players, it did increase our count to an additional 21 players. These players are listed below: 

Prior To Trade

 

After Trade

 

Difference

Last_Name

First_Name

Year

Team

G

REC

YDS

TD

 

Year

Team

G

REC

YDS

TD

 

REC

YDS

TD

Horn

Joe

1999

kan

16

35

586

6

 

2000

nor

16

94

1340

8

 

59

754

2

Metcalf

Eric

1994

cle

16

47

436

3

 

1995

atl

16

104

1189

8

 

57

753

5

Rison

Andre

1996

jax

10

34

458

2

 

1997

kan

16

72

1092

7

 

38

634

5

Burress

Plaxico

2004

pit

11

35

698

5

 

2005

nyg

16

76

1214

7

 

41

516

2

Jackson

Michael

1995

cle

13

44

714

9

 

1996

rav

16

76

1201

14

 

32

487

5

McCardell

Keenan

1995

cle

16

56

709

4

 

1996

jax

16

85

1129

3

 

29

420

-1

Owens

Terrell

2005

phi

7

47

763

6

 

2006

dal

16

85

1180

13

 

38

417

7

Johnson

Keyshawn

2003

tam

10

45

600

3

 

2004

dal

16

70

981

6

 

25

381

3

Graham

Jeff

1998

phi

15

47

600

2

 

1999

sdg

16

57

968

2

 

10

368

0

Graham

Jeff

1993

pit

15

38

579

0

 

1994

chi

16

68

944

4

 

30

365

4

Langhorne

Reggie

1991

cle

14

39

505

2

 

1992

clt

16

65

811

1

 

26

306

-1

Davis

Willie

1995

kan

16

33

527

5

 

1998

nor

15

53

823

1

 

20

296

-4

Conway

Curtis

1999

chi

10

44

426

4

 

2000

sdg

14

53

712

5

 

9

286

1

Brooks

Bill

1992

clt

14

44

468

1

 

1993

buf

16

60

714

5

 

16

246

4

Shaw

Bobby

2002

jax

16

44

525

1

 

2003

buf

16

56

732

4

 

12

207

3

Patten

David

2000

cle

14

38

546

1

 

2001

nwe

16

51

749

4

 

13

203

3

Taylor

Travis

2004

rav

10

34

421

0

 

2005

min

16

50

604

4

 

16

183

4

Monk

Art

1993

was

16

41

398

2

 

1994

nyj

16

46

581

3

 

5

183

1

Thrash

James

2000

was

16

50

653

2

 

2001

phi

15

63

833

8

 

13

180

6

Hakim

Az-zahir

2001

ram

16

39

374

3

 

2002

det

10

37

541

3

 

-2

167

0

Jefferson

Shawn

1995

sdg

16

48

621

2

 

1996

nwe

15

50

771

4

 

2

150

2

 Some interesting notes about these players:
1. Only 8 of these players have changed teams since 2000. 
2. Jeff Graham makes a double appearance. 
3. In 12 of these 21 scenarios, the player appeared in more games the following season, thus attributing to their improvement.

So if we look at our total initial population of 156 receivers who changed teams, only 19% improved their performance from year to year.

What caused this improvement?
Now that we have identified the players that did improve, are there any conclusions we could draw from these experiences?

One item that clearly jumps out is that only 2 of the 9 players who switched teams after performing well for their current team significantly increased their number of receptions. However, if we look at the player who improved from our total sample, we can see that ever single player improved their receptions, and 12 of the 20 appeared to move into the #1 role for their new team. This clearly indicates that the best odds for identifying a player who will increase their performance is looking for a situation where the player is moving from a #2 or #3 receiver into the #1 role.

Honestly, there doesn't appear to be another factor that would indicate what players will improve significantly. There is no individual year that seems to show significance for improved performance. In many cases, these receivers do appear to be main end-zone target for their new teams, but only half improved the touchdown receptions by 4 or more, so this isn't consistent either.

So what does this mean?

General Conclusions
After looking at the numbers, it appears that is best to stay away from any receiver who changes teams, even if they are a big-name player. Historical data indicates that you can expect a 20% reductions in receptions, yards, and touchdowns for receivers who change teams. Even using the most conservative measurement standard, there appears to be at least a 17% reduction for even the best wide receivers.

Outside of the amazing 2007 16-0 Patriots and Randy Moss, while there have been some wide receivers who have improved after changing teams (19% overall), only 9 of these players would be what we would consider quality players at the time they changed teams. Only 13 receivers who have changed teams since 2000 have shown improvement after changing teams.

If you are going to gamble on a receiver who has changed teams, you are better off looking for somebody who is moving into a more prominent role with their new team, and who may be the primary end-zone receiving target for their new team.

Final Point
Based on all the research that was included in this article, very few receivers seem to fit all the indicators of being a breakout player with their new team. We may never see a WR on a new team approach anything like Randy Moss' 2007 performance for years to come.
 

We are your friend's secret weapon.

  • Get 4 FREE downloads
  • Receive breaking news alerts & analysis
  • BONUS: Learn how to play DFS.
  • Battle-Tested by 40,000+ fantasy football diehards since 1999.