A Beginner’s Attempt at an NFL Draft Value Chart

It was only fitting that a team could be so bad that it couldn’t even lose properly. Last Sunday, the 2024 New England Patriots seemingly picked the worst possible time to win a game, defeating the backups of the Buffalo Bills to finish the season 4-13 – and, as a result, losing the number one pick. Shortly following the “victory,” New England fired its rookie head coach Jerod Mayo. The dream of chasing successful ghosts of the past in the 2024 season came to a definitive, ugly ending.

Although the Patriots lost the number one pick, the team currently has the fourth pick in the draft. In fact, when writing this column, I initially wanted to examine how much New England lost by winning the final game of the season. As I continued my research, however, I became more interested in the topic of draft capital itself. What does history suggest about the typical return on a number one overall pick versus other picks in the draft? 

Today’s column is a little all over the place. But to summarize it, I analyzed the modern history of number one overall picks, their typical success rate, and what their returns on investment are. Afterward, I took a deeper look at what the typical return on investment was from each pick in the draft. 

Success Rate of Number 1 Picks

The first thing I did was look up every number one pick from 2000 to 2020 (I chose to exclude everything afterward due to the long-term trajectory of players from 2021 to now remaining up in the air). I used Pro Football Reference to look up the history of draft picks and then organized them as such, in order of draft picks with the most amount of weighted Approximate Value earned strictly for the team that drafted the player. 

NOTE: For a more specific explanation of how to evaluate AV, read Doug Drinen’s post here. The way I’d roughly describe it in single-season terms is that anything above a 18 is roughly an MVP season, in between 14 and 18 are Pro Bowl seasons, in between 9 and 14 are starting-caliber years, and beneath that gets into rotation player territory or busts.

Keep in mind that Approximate Value is not a perfect stat; it’s meant to give a general overview of ranges of players. It may not correctly tell you which individual has more value between a player with 15 AV versus another with 13 AV, but on the whole, it will capture the reality that most players with 15 AV are worth more than most players with 13 AV. By and large, it’s also a convenient tool for our purposes of evaluating thousands of draft picks later. Nonetheless, I also added a separate column measuring whether or not a player received a follow-up extension to his rookie contract, excluding franchise tags. 

Year & Player	Approximate Value by Drafting Team (drAV)	Did he receive a follow-up to his rookie contract by the drafting team? (Excluding Franchise Tag)
2011 Cam Newton	107	Yes
2009 Matthew Stafford	98	Yes
2017 Myles Garrett	79	Yes
2012 Andrew Luck	72	Yes
2003 Carson Palmer	60	Yes
2001 Michael Vick	60	Yes
2019 Kyler Murray	57	Yes
2013 Eric Fisher	56	Yes
2015 Jameis Winston	54	No
2016 Jared Goff	47	Yes
2008 Jake Long	46	No
2020 Joe Burrow	44	Yes
2002 David Carr	42	No
2014 Jadeveon Clowney	38	No
2018 Baker Mayfield	39	No
2006 Mario Williams	36	No
2005 Alex Smith	28	Yes
2010 Sam Bradford	25	No
2000 Courtney Brown	21	No
2007 JaMarcus Russell	6	No
2004 Eli Manning*	0	No

Even our two rough measures of success, however, lead to a few interesting cases that don’t cleanly fit as objective measurements of success. Most notably, Eli Manning stands out as a near-lock Hall of Fame player who within the strict confines of our criteria counts as a “bust” for the Chargers, even though they ended up functionally swapping picks with the Giants and choosing Philip Rivers. There are many such cases in NFL history where a player drafted by a franchise ends up playing somewhere else – it doesn’t always, however, necessarily mean that the player was a bust. In similar fashion, a player not signing an extension to his rookie contract doesn’t always indicate whether someone is a success or not; however, it is a strong indicator. 

Nonetheless, we get a pretty good picture at the typical success rate of a number one pick. In summary, it’s not a surefire guarantee, but it’s well over fifty percent if you count anyone above the “Bradford” line in drAV or anyone who received a second contract as a “not a bust.”

What is the value of each draft pick? 

The next thing I did was way more time-exhaustive: research every single draft pick from 2000 to 2020, and organize them by average drAV. In order to ensure that there were no shenanigans – such as the case of a lower draft pick coincidentally having a higher expected value than a ‘better’ draft pick – I then ran a polynomial regression trendline, accounting for outliers above the 75 percentile threshold and beneath the 25 percentile threshold. Essentially, this became a new column: “expected drAV.”

NOTE: In addition to accounting for outliers, I also entirely removed any instance of a player being selected with a pick, only to be traded elsewhere. There ended up being 871 such instances (the most prominent ones being Manning, Rivers, Lawrence Guy, and Marc Bulger). However,  in the context of assessing around 4000 picks, this still left us plenty of data to compare with simplistic methods.

How does my draft chart compare to other draft value charts? 

After I had my final table of expected DrAV by the trendline, I figured that I would compare my results to how NFL experts in the past have attempted to capture draft value. It would have taken me too much time to look at every possible example of a ‘draft value chart,’ but I decided to go with two prominent choices: the famous Jimmy Johnson one from the 1990s (notably before the 2011 collective bargaining agreement), and another one shared by one of my several personal sports-writing inspirations in Chase Stuart, from a little over a decade ago.

I won’t mince words here: there are several differences in methodology between Johnson and Stuart that make their models far more rigorous than mine. It’s no surprise that a first-time sports analytics grad student is not factoring in as many elements to his analysis as people who literally think about football for a living. As such, reviewing their methodologies and contexts behind their charts was a bit of a humbling experience. To broadly summarize some of our key differences:

• I captured raw drAV per pick from 2000 to 2020. 
• Stuart estimated Approximate Value provided by each player strictly within his first five years, with a data set that spanned from 1980 to 2007 (notably before the collective bargaining agreement of 2011).  
• As the 33rd team describes, Johnson’s value chart was also made with salary cap and cash considerations, but to a far bigger degree, and with an emphasis on trade value (not the typical return on investment from a draft pick).

Out of curiosity, I decided to compare my chart to Johnson’s and Stuarts. After capturing the values that each of us assigned our picks in the draft, I then normalized them into Z-Scores in order to ensure proper standardization. Following that, I then compared the values together, making sure to measure their respective Pearson correlation coefficients (essentially the linear correlation between our data values).

For the specific values, feel free to see the appendices. As of right now though, I don’t think I did too badly for a beginner, but clearly, I have a long way to go to catch up to my heroes of the past. No wonder I’m still a Patriots fan.

Appendices & Sources

• Pro Football Reference
• My Spreadsheet of NFL Drafts (2000-2020)
• My Football Projects GitHub

My Expected Value Chart

Pick	Expected_DrAV
1	39.24
2	38.73
3	38.23
4	37.74
5	37.25
6	36.76
7	36.28
8	35.81
9	35.34
10	34.87
11	34.42
12	33.96
13	33.51
14	33.07
15	32.63
16	32.2
17	31.78
18	31.35
19	30.94
20	30.52
21	30.12
22	29.72
23	29.32
24	28.93
25	28.54
26	28.16
27	27.78
28	27.41
29	27.04
30	26.67
31	26.32
32	25.96
33	25.61
34	25.27
35	24.93
36	24.59
37	24.26
38	23.93
39	23.61
40	23.29
41	22.98
42	22.67
43	22.36
44	22.06
45	21.77
46	21.47
47	21.18
48	20.9
49	20.62
50	20.34
51	20.07
52	19.8
53	19.54
54	19.28
55	19.02
56	18.77
57	18.52
58	18.28
59	18.04
60	17.8
61	17.57
62	17.34
63	17.11
64	16.89
65	16.67
66	16.45
67	16.24
68	16.03
69	15.83
70	15.63
71	15.43
72	15.23
73	15.04
74	14.86
75	14.67
76	14.49
77	14.31
78	14.14
79	13.96
80	13.79
81	13.63
82	13.47
83	13.31
84	13.15
85	13
86	12.85
87	12.7
88	12.55
89	12.41
90	12.27
91	12.13
92	12
93	11.87
94	11.74
95	11.61
96	11.49
97	11.37
98	11.25
99	11.14
100	11.02
101	10.91
102	10.8
103	10.7
104	10.59
105	10.49
106	10.39
107	10.3
108	10.2
109	10.11
110	10.02
111	9.93
112	9.85
113	9.76
114	9.68
115	9.6
116	9.52
117	9.45
118	9.37
119	9.3
120	9.23
121	9.16
122	9.09
123	9.03
124	8.96
125	8.9
126	8.84
127	8.78
128	8.73
129	8.67
130	8.62
131	8.57
132	8.52
133	8.47
134	8.42
135	8.37
136	8.33
137	8.28
138	8.24
139	8.2
140	8.16
141	8.12
142	8.08
143	8.04
144	8.01
145	7.97
146	7.94
147	7.91
148	7.88
149	7.84
150	7.81
151	7.79
152	7.76
153	7.73
154	7.7
155	7.68
156	7.65
157	7.63
158	7.6
159	7.58
160	7.56
161	7.53
162	7.51
163	7.49
164	7.47
165	7.45
166	7.43
167	7.41
168	7.39
169	7.37
170	7.35
171	7.33
172	7.32
173	7.3
174	7.28
175	7.26
176	7.24
177	7.22
178	7.21
179	7.19
180	7.17
181	7.15
182	7.13
183	7.12
184	7.1
185	7.08
186	7.06
187	7.04
188	7.02
189	7
190	6.98
191	6.96
192	6.94
193	6.91
194	6.89
195	6.87
196	6.85
197	6.82
198	6.8
199	6.77
200	6.74
201	6.72
202	6.69
203	6.66
204	6.63
205	6.6
206	6.57
207	6.54
208	6.51
209	6.47
210	6.44
211	6.4
212	6.36
213	6.33
214	6.29
215	6.25
216	6.2
217	6.16
218	6.12
219	6.07
220	6.02
221	5.98
222	5.93
223	5.88
224	5.82
225	5.77
226	5.71
227	5.66
228	5.6
229	5.54
230	5.47
231	5.41
232	5.35
233	5.28
234	5.21
235	5.14
236	5.07
237	4.99
238	4.91
239	4.84
240	4.76
241	4.67
242	4.59
243	4.5
244	4.41
245	4.32
246	4.23
247	4.13
248	4.04
249	3.94
250	3.84
251	3.73
252	3.62
253	3.52
254	3.4
255	3.29
256	3.17
257	3.06
259	2.81
260	2.68
261	2.55

​​