Maurile Temblay - Fantasy Points Per Target

Each week, Footballguys.com publishes sortable target data for players at each position. (A target is a pass that is intended for a particular player; for example, if Drew Bledsoe attempts four passes intended for Josh Reed, that means Josh Reed has four targets.)

How useful is that data? For purposes of predicting future fantasy points for a wide receiver, is past target data as useful as past fantasy points? More useful?

I looked at target data from the first four weeks of last season to see how effective it was at predicting fantasy points per game over the remainder of the season. It turned out that targets per game was a more useful statistic than fantasy points per game.

The following chart partially illustrates that point. The first column shows each WR's ranking in terms of fantasy points per game from weeks 5-17 - that's what we're trying to predict. To the right of each player, you can see where he ranked after four weeks in terms of targets per game and fantasy points per game.

'02 PPG Rk Wks 5-17	Player	'02 Targets Rk Wks 1-4	'02 PPG Rk Wks 1-4
1	Terrell Owens	28	24
2	Marvin Harrison	6	8
3	Chad Johnson	50	71
4	Hines Ward	5	1
5	Laveranues Coles	41	69
6	Randy Moss	1	25
7	Amani Toomer	21	27
8	Plaxico Burress	18	56
9	Koren Robinson	22	62
10	Joe Horn	6	17
11	Eric Moulds	3	2
12	Peerless Price	12	6
13	Jerry Rice	12	10
14	Isaac Bruce	26	39
15	Travis Taylor	46	79
16	Keyshawn Johnson	28	53
17	Derrick Mason	15	29
18	Bill Schroeder	73	61
19	Torry Holt	4	14
20	Quincy Morgan	41	31

For 16 out of the 20 receivers listed, after the first four weeks of the season, their targets per game was a better indicator of future fantasy performance than their fantasy points per game was.

To be a bit more rigorous, I looked at all WRs who had gotten at least 12 targets over the first four weeks of the 2002 season, and examined how well their targets per game up to that point correlated with their future fantasy points per game. By every measure, targets per game in weeks 1-4 was better correlated with fantasy points per game in weeks 5-17 than were fantasy points per game in weeks 1-4. In other words, if you were trying to predict how many fantasy points per game a particular WR would get in weeks 5 though 17 of the season, and you could choose whether to view his targets per game before that point or his fantasy points per game, you should choose to view his targets per game.

(For the statistically inclined among you, the coefficient of determination between targets/game in weeks 1-4 and fantasy points/game in weeks 5-17 was 0.271, while the coefficient of determination between fantasy points/game in weeks 1-4 and fantasy points/game in weeks 5-17 was 0.147. The correlation coefficients were 0.520 and 0.384, respectively. Using the best-fitting linear equations to predict fantasy points per game in weeks 5-17, the sum of the squared errors was 17% higher using fantasy points per game than using targets per game.)

The next two charts are last year's Top 20 WR lists after the first four weeks - the first in terms of targets per game, and the second in terms of fantasy points per game.

'02 Rk	Player	'02 T/G: 1-4	'02 PPG: 5-7
1	Randy Moss	15.5	12.4
2 tie	Troy Brown	13.3	6.6
2 tie	Eric Moulds	13.3	10.8
4	Torry Holt	12.3	9.5
5	Hines Ward	12.0	13.3
6 tie	Marvin Harrison	11.0	15.2
6 tie	Joe Horn	11.0	10.9
8	Marty Booker	10.8	8.9
9 tie	Tim Brown	10.3	5.8
9 tie	Deion Branch	10.3	2.0
9 tie	Rod Smith	10.3	8.1
12 tie	Peerless Price	10.0	10.3
12 tie	Jerry Rice	10.0	10.2
12 tie	Kevin Johnson	10.0	5.2
15 tie	David Patten	9.8	6.1
15 tie	Derrick Mason	9.8	9.5
15 tie	Az-Zahir Hakim	9.8	7.8
18	Plaxico Burress	9.7	12.2
19 tie	Jimmy Smith	9.0	9.2
19 tie	David Boston	9.0	6.8
.	Average	.	9.0

'02 Rk	Player	'02 PPG: 1-4	'02 PPG: 5-7
1	Hines Ward	15.5	13.3
2	Eric Moulds	14.8	10.8
3	Curtis Conway	14.6	8.3
4	Donald Driver	14.4	9.1
5	James Thrash	13.9	5.7
6	Peerless Price	13.7	10.3
7 tie	Marty Booker	13.6	8.9
7 tie	Marvin Harrison	13.6	15.2
9	Troy Brown	11.7	6.6
10	Jerry Rice	11.0	10.2
11 tie	Deion Branch	10.8	2.0
11 tie	Qadry Ismail	10.8	2.9
13	Andre Davis	10.7	3.0
14 tie	Torry Holt	10.6	9.5
14 tie	Tim Brown	10.6	5.8
16	Donte Stallworth	10.5	7.3
17	Joe Horn	10.4	10.9
18	Corey Bradford	10.2	5.4
19	David Patten	10.0	6.1
20	Brian Finneran	9.8	7.0
.	Average	.	7.9

Note that the first group outscored the second group over the remainder of the season by just over one point per game, again tending to show that targets are a better predictor of future performance than fantasy points are.

How can you use this information to your advantage? It may be able to help you identify players who are currently under- or overrated. If targets mean more than fantasy points, then players who rank very high in targets but low in fantasy points may be primed to start scoring more fantasy points in the near future. On the other hand, players who rank much higher in fantasy points than they do in targets may see their production decline in the coming weeks. Calculating each player's fantasy points per target should allow us to quickly identify which players are getting more opportunities than their current points per game reflect, and which are not getting as many as it might appear.

Somewhat counterintuitively, players with a high fantasy points per target - those who are currently doing more with less - are the ones we should look at as possibly being overvalued. If we'd calculate every WR's fantasy points per target after four weeks last year (ignoring WRs without at least five targets per game), the Top 20 list would have been as follows:

'02 Rk	Player	PPT	'02 PPG: 1-4	'02 PPG: 5-7
1	Curtis Conway	2.2	14.6	8.3
2 tie	Andre Davis	1.9	10.7	3.0
2 tie	Donald Driver	1.9	14.4	9.1
4 tie	Eddie Kennison	1.8	9.6	5.8
4 tie	Donte Stallworth	1.8	10.5	7.3
4 tie	James Thrash	1.8	13.9	5.7
7	Corey Bradford	1.6	10.2	5.4
8 tie	Keenan McCardell	1.4	8.9	6.7
8 tie	Josh Reed	1.4	7.4	2.8
8 tie	Qadry Ismail	1.4	10.8	2.9
8 tie	Brian Finneran	1.4	9.8	7.0
8 tie	Peerless Price	1.4	13.7	10.3
13 tie	Quincy Morgan	1.3	8.8	9.4
13 tie	Hines Ward	1.3	15.5	13.3
13 tie	Marty Booker	1.3	13.6	8.9
16 tie	Marvin Harrison	1.2	13.6	15.2
16 tie	Antonio Freeman	1.2	7.0	4.7
16 tie	Steve Smith	1.2	9.2	6.2
16 tie	Rod Gardner	1.2	9.0	9.4
16 tie	Terrell Owens	1.2	9.3	17.6
.	Average	.	11.0	8.0

Note that 16 out of those 20 players averaged fewer points over the remainder of the season than they did in the first four games. The average decline was three points per game.

Here are the players who had the lowest ratios of fantasy points per target through four weeks last year.

'02 Rk	Player	PPT	'02 PPG: 1-4	'02 PPG: 5-7
1	Travis Taylor	0.4	2.6	9.6
2 tie	Derrick Alexander	0.5	3.3	1.5
2 tie	Laveranues Coles	0.5	3.6	13.3
2 tie	Koren Robinson	0.5	4.5	11.8
5 tie	Chad Johnson	0.6	3.5	13.3
5 tie	Peter Warrick	0.6	3.4	7.8
5 tie	Plaxico Burress	0.6	5.5	12.5
5 tie	T.J. Houshmandzadeh	0.6	3.5	3.4
5 tie	Randy Moss	0.6	9.3	12.4
5 tie	Az-Zahir Hakim	0.6	6.3	7.8
11 tie	Willie Jackson	0.7	4.4	2.1
11 tie	Antonio Bryant	0.7	6.0	7.4
11 tie	Kevin Dyson	0.7	4.3	7.6
11 tie	Keyshawn Johnson	0.7	6.0	9.6
15 tie	Dez White	0.8	5.0	5.9
15 tie	Rod Smith	0.8	8.6	8.1
15 tie	Brandon Stokley	0.8	6.5	6.3
18 tie	Darrell Jackson	0.9	7.1	9.3
18 tie	Torry Holt	0.9	10.6	9.5
18 tie	Troy Brown	0.9	11.7	6.6
.	Average	.	5.8	8.3

Thirteen of the 20 players in this group improved their performance over the remainder of the season, and the average improvement was 2.5 points per game.

So last year, examining fantasy points per target after four weeks might have helped identify some underrated players at the time - such as Laveranues Coles, Koren Robinson, and Plaxico Burress - as well as some overrated players - such as Curtis Conway, Andre Davis, Corey Bradford, or Qadry Ismail.

Over the next few weeks, you may want to keep your eye on fantasy points per target to help you identify which WRs may be primed for improvement and which may see their fantasy points drop off a bit. As always, overall statistical trends are no substitute for analysis of an individual player's own merits; but every little piece of information helps, and knowing the direction of certain overall trends can prove useful. A ratio greater than about 1.3 could mean that a WR is already making the most of all his opportunities and there's nowhere to go but down, while a ratio under 0.9 could be a sign that he is getting enough looks to excel, but for whatever reason has (temporarily?) been unable to capitalize on them.

Currently, here are the receivers with the highest ratios of fantasy points per target this year (excluding WRs without at least 5 targets per game).

'03 Rk	Player	PPT
1	Darrell Jackson	1.9
2	Chris Chambers	1.7
3 tie	Laveranues Coles	1.4
3 tie	Hines Ward	1.4
3 tie	Reggie Wayne	1.4
3 tie	Eric Moulds	1.4
3 tie	Torry Holt	1.4
8 tie	Santana Moss	1.3
8 tie	Joey Galloway	1.3
8 tie	Matthew Hatchette	1.3
8 tie	Anquan Boldin	1.3
8 tie	Koren Robinson	1.3
8 tie	Amani Toomer	1.3
14 tie	Eric Parker	1.2
14 tie	Ike Hilliard	1.2
14 tie	Randy Moss	1.2
14 tie	Plaxico Burress	1.2
18 tie	Chad Johnson	1.1
18 tie	Isaac Bruce	1.1
18 tie	Javon Walker	1.1

And here are the WRs with the lowest ratios of fantasy points per target.

'03 Rk	Player	PPT
1 tie	Todd Pinkston	0.3
1 tie	Quincy Morgan	0.3
1 tie	Troy Brown	0.3
4 tie	Marty Booker	0.4
4 tie	Peerless Price	0.4
6 tie	Josh Reed	0.5
7 tie	Dez White	0.6
7 tie	Terry Glenn	0.6
7 tie	Bryant Johnson	0.6
10 tie	Curtis Conway	0.7
10 tie	Rod Smith	0.7
10 tie	Travis Taylor	0.7
10 tie	Tim Brown	0.7
10 tie	Jerry Rice	0.7
10 tie	Keyshawn Johnson	0.7
10 tie	Kevin Johnson	0.7
10 tie	Marvin Harrison	0.7
18 tie	Jabar Gaffney	0.8
18 tie	Justin McCareins	0.8
18 tie	James Thrash	0.8

Scoring System:
4 pts = TD class="pagebody2" pass
6 pts = TD class="pagebody2" run/catch
1 pt = 20 YD pass
1 pt = 10 YD run/catch
-1 pt = INT