The Stats Room: Projecting Rookies Within Team Context

Written by

Updated on September 3, 2018 12:18PM EST

Projecting football players is tough, and I'm going dive into one of the impossible areas, rookies. There is just not a lot of information to go off with them. It's not like baseball where they've played least a couple seasons in the minor leagues to get a profile. In the NFL, there is their varied college performance, some combine measurements and maybe a few quarters of televised practice. I'm here to provide a clear answer using historical rookie results and team context.

Previously, I just projected rookie running by going off an average production level. There didn't seem to be any measurable production differences between early and late picks.

One item that really bothered me about this analysis was team context. The rookie running backs who got the most work were likely on teams who needed their help. These teams were bad and the running back's production likely suffered from a bad supporting cast. I was looking for a way to add some overall team context and I may have found an answer in Las Vegas.

For years, fans have been able to bet the over/under on a team's wins. SportsOddsHistory.com collected all the preseason expectations from 2001 to 2017. While not a perfect proxy for team talent, it's a starting point. And I have 17 years of data to match up.

I decided to group the preseason odds into three groups:

1. More than nine wins
2. Seven to nine wins
3. Fewer than seven wins

The seven-to-nine

Previously, I just projected rookie running by going off an average production level. There didn't seem to be any measurable production differences between early and late picks.

I decided to group the preseason odds into three groups:

1. More than nine wins
2. Seven to nine wins
3. Fewer than seven wins

The seven-to-nine group is the average team projection while the other two are seen as better or worse than average. By using the three groupings, a distinct separation can be seen. I'll start with the rookie running backs.

Running Backs

For running backs, I split them by backs drafted in the first three rounds and then those drafted later. These two groups then got divided into the three win categories. Here are the results.

Note: I used rookies drafted in the first 96 picks (32 x 3) to represent the first three rounds.

PROJ. TEAM WINS	YDS/RUSH	TD/RUSH	FUMBLE/TOUCH	REC/TRG	YDS/REC	TD/REC	AVG PICK
First Three Rounds
>9 Wins	4.4	2.87%	0.92%	77.22%	8.2	1.87%	57
7-9 Wins	4.3	3.01%	0.70%	74.54%	8.3	3.06%	49
<7 Wins	4.1	2.51%	0.73%	72.08%	8.4	2.52%	39
Fourth Round or Later
>9 Wins	4.1	1.99%	1.06%	69.59%	7.1	3.15%	161
7-9 Wins	4.0	2.56%	0.59%	72.05%	8.0	3.07%	163
<7 Wins	4.0	2.47%	0.91%	69.67%	7.5	1.90%	155

There is a lot going on, but here are some highlights.

• The bad-team running backs were taken earlier in the drafts and performed worse. Of the backs taken in the first three rounds, the ones taken by projected winning teams had a distinctive advantage in yards per carry.

• The backs drafted early by bad teams had similar yards per carry compared to backs drafted by winning teams more than 120 rounds later.

• As previously researched, the backs drafted earlier are better receivers with higher completion rates and yards per completion.

Wide Receivers

With wide receivers, I divided the results into the same six groups and here are the results.

PROJ. TEAM WINS	REC/TRG	YDS/REC	TD/REC	FUMBLE/TOUCH	AVG PICK
First Three Rounds
>9 Wins	58.41%	14.0	8.63%	1.51%	44
7-9 Wins	56.12%	13.1	8.30%	1.55%	48
<7 Wins	50.60%	14.1	8.41%	1.13%	44
Fourth Round or Later
>9 Wins	60.70%	11.0	7.37%	1.62%	153
7-9 Wins	54.25%	12.8	6.95%	2.59%	168
<7 Wins	50.32%	12.7	6.48%	2.34%	154

The completion percentage is rather stark with a 10 percent spread for receivers drafted after the third round. A lot of this difference likely comes down to the quarterback play as I recently researched. Bad overall teams usually have bad quarterbacks. And it's time to move onto those quarterbacks.

Quarterbacks

Hopefully, owners don't have to use rookie quarterbacks, as their historical production is horrible compared to the rest of the league's starters.

PROJ. TEAM WINS	COMP%	YDS/COMP	TD/COMP	INT/ATT	AVG PICK
First Three Rounds
>9 Wins	60.6%	10.9	5.1%	2.2%	62
7-9 Wins	57.1%	11.7	6.7%	3.5%	31
<7 Wins	57.1%	11.4	5.5%	3.2%	21
Fourth Round or Later
>9 Wins	53.9%	12.1	9.1%	3.7%	183
7-9 Wins	55.8%	10.8	5.3%	3.3%	153
<7 Wins	53.5%	10.4	4.3%	4.8%	152

A couple items really stick out. Of the quarterbacks drafted in the first three rounds, those on projected good teams had similar production to those on bad teams, but the ones on bad teams went 40 picks earlier. Second, the only noticeable trend is the lower completion percentage of the quarterbacks taken later.

Conclusion

Team context matters when creating a rookie's projection. That statement isn't ground shattering. The key is how much to adjust projections based on the team context.

For running backs, it's +/- 0.2 yards per carry likely based on the offensive line.

For wide receivers, their catch rate can vary +/- 5 percent.

Additionally, these projections help to cap expectations on certain players. In the confusing world of rookie projections, hopefully, this provides some guidance.

For reference, here are the projections for the top-25 rookies being drafted at FanTrax. The assumptions are:

• Running backs: Total yards and touchdowns projected from 250 carries and 40 targets

• Wide receiver: Receiving yards and touchdowns projected from 120 targets

• Quarterbacks: Passing yards and touchdowns projected from 500 targets

PLAYER	TEAM	POS	DRAFT POS	PROJ. WINS	YARDS	TD
Saquon Barkley	NYG	RB	2	7	1,319	8.7
Rashaad Penny	SEA	RB	27	8	1,319	8.7
Royce Freeman	DEN	RB	71	7	1,319	8.7
Sony Michel	NE	RB	31	11	1,357	7.9
Ronald Jones II	TB	RB	38	6.5	1,270	7.3
Kerryon Johnson	DET	RB	43	7.5	1,319	8.7
Nick Chubb	CLE	RB	35	5.5	1,270	7.3
Calvin Ridley	ATL	WR	26	9	881	5.6
D.J. Moore	CAR	WR	24	9	881	5.6
Michael Gallup	DAL	WR	81	8.5	881	5.6
Nyheim Hines	IND	RB	104	6.5	1,203	6.9
Christian Kirk	ARI	WR	47	5.5	854	5.1
Anthony Miller	CHI	WR	51	6.5	854	5.1
Courtland Sutton	DEN	WR	40	7	881	5.6
James Washington	PIT	WR	60	10.5	984	6
Jordan Wilkins	IND	RB	169	6.5	1,203	6.9
Baker Mayfield	CLE	QB	1	5.5	3,248	15.7
Lamar Jackson	BAL	QB	32	8	3,348	19.1
Kalen Ballage	MIA	RB	131	6.5	1,203	6.9
Keke Coutee	HOU	WR	103	8.5	832	4.5
Dante Pettis	SF	WR	44	8.5	881	5.6
Tre'Quan Smith	NO	WR	126	9.5	802	5.4
Josh Allen	BUF	QB	7	6	3,248	15.7
Chase Edmonds	ARI	RB	134	5.5	1,203	6.9
Bo Scarbrough	DAL	RB	236	8.5	1,240	7.6