Players with the best NFL Combine results for fantasy football success

The NFL Combine has come and gone, leaving in its wake a wave of winners, losers and shifting big boards. Not to be left out of the fun, I dove in with an analytical look at the 2020 prospects based on which Combine metrics matter for offensive and defensive players.

The technique I used in those analyses is called ridge regression, which is great at quantifying the importance of each metric for comparison. We can now say with some confidence which drills matter most for each position group based on historical data. In this analysis, my process changes in a few ways.

First, the target variable, i.e. what we’re predicting, is going to be the fantasy football success of prospects, not draft position and player value (PFF WAR). Second, fantasy success is measured as a binary variable (yes/no), based on whether each player achieved a particular season-end scoring threshold in his first three seasons. These thresholds are top-12 scorer for quarterbacks, top-20 for running backs, top-30 for wide receivers and top-eight for tight ends. Lastly, the algorithm employed here doesn’t assume there is the linear relationship between performance in each feature variable and the target, and there’s no assumption that features are independent.

[Editor's Note: PFF's new college-to-pro projection system is powered by AWS machine learning capabilities.]

This analysis projects the likelihood of fantasy success in a prospect's first three seasons using the random forest algorithm. Random forest predicts classifications (success or not) by building many decision trees, randomizing samples and features, creating a “forest” of decision trees whose collective predictions are more accurate than an individual tree.

This technique differs from the more linear relationships in classification algorithms like logistic regression. A linear relationship would tell us that a prospect who runs 0.2 seconds faster than average would have a feature effect twice that as one who runs 0.1 seconds faster, but we know that relationships aren’t that simple and the benefits of being faster, bigger, quicker can see diminishing returns as some point. The downside of using random forest versus logistic regression comes in interpretability. While forest-based algorithms often produce better predictions, the exact manner in which they have split data and assigned probabilities is opaque.

I’ll illustrate the process below for every position by including an example decision tree first, then detailing results of the random forest predictions on historical prospects and the 2020 class. You’ll see that a single decision tree contains some useful insights but with rigid cutoffs that are likely overfit to past data.

Quarterbacks

Our sample decision for the quarterback tree includes the 110 prospects in our 2007-2017 data who participated in the majority of Combine drills. The first split is by broad jump at 118 inches, with those at or above this number falling into the highest success node. This is the node in the lower right corner, that includes three numbers: the ratio of data points that fall into the node classified as successful (0.41), the total number of prospects who fall into the node (n=22) and the rough percentage that fall into the node of the entire population (20%).

Those who don’t make the 118-inch cut for broad jump are then sorted by short shuttle times, first at less than 4.3 seconds and then greater than or equal to 4.5 seconds. This is a perfect example of how using a single decision tree can overfit the data. There’s no logical reason why a quarterback would be more successful with a short shuttle at or above 4.5 seconds than at 4.4 seconds, which is the case for the last split in the bottom left. The random forest we use for the results below is informed by so many trees that overfitting is less of an issue.

Top-10 Historical Success Probabilities: Quarterbacks (2006-2017)

Year	Pick	Player	Success	Success Prob	Height	Weight	Forty	Bench	Broad	Vertical	3cone	Shuttle
2014	3	Blake Bortles	Yes	62.6%	77	232	4.93	NaN	115	32.5	7.08	4.21
2016	26	Paxton Lynch	No	60.0%	79	244	4.86	NaN	118	36	7.14	4.26
2010	25	Tim Tebow	No	59.9%	75	236	4.71	NaN	115	38.5	6.66	4.17
2009	17	Josh Freeman	Yes	58.1%	78	248	4.9	NaN	119	33.5	7.11	4.43
2017	300	Mitch Leidner	No	56.4%	75	226	4.93	NaN	123	34.5	6.96	4.25
2009	5	Mark Sanchez	Yes	56.3%	74	227	4.93	NaN	116	32.5	7.06	4.21
2015	103	Bryce Petty	No	55.2%	75	230	4.87	NaN	121	34	6.91	4.13
2016	135	Dak Prescott	Yes	54.3%	74	226	4.79	NaN	116	32.5	7.11	4.32
2017	2	Mitchell Trubisky	No	50.7%	74	222	4.67	NaN	116	27.5	6.87	4.25
2011	300	Jerrod Johnson	No	49.3%	77	251	4.75	NaN	119	29	7.28	4.31

As I mentioned earlier, it’s more difficult to interpret random forest results, but the consistent themes in the top-10 are strong broad jumps and middling-to-strong performance in the agility drills. In addition, all 40-yard dash times were under five seconds, and all weights are over 220 pounds.

The model was fairly poor at picking out actual success, with only fleeting successes like Blake Bortles, Josh Freeman and Mark Sanchez qualifying. The results tell us what we already knew: a lot more goes into being a successful quarterback than athleticism.

2020 Quarterbacks

Year	Player	Success Prob	Height	Weight	Forty	Bench	Broad	Vertical	3cone	Shuttle	Rank
2020	Justin Herbert	50.0%	78	236	4.68	NaN	123	35.5	7.06	4.46	10
2020	Cole McDonald	42.5%	75	215	4.58	NaN	121	36.0	7.13	4.52	21
2020	James Morgan	34.5%	76	229	4.89	NaN	112	29.0	7.51	4.64	40
2020	Nate Stanley	33.1%	75	235	4.81	NaN	108	28.5	7.26	4.48	42
2020	Jalen Hurts	31.5%	73	222	4.59	NaN	125	35.0	NaN	NaN	44
2020	Jake Fromm	30.2%	73	219	5.01	NaN	111	30.0	7.27	4.51	48
2020	Kelly Bryant	27.3%	75	229	4.69	NaN	125	35.0	7.33	4.51	56
2020	Kevin Davidson	26.5%	76	224	5.00	NaN	NaN	NaN	7.13	4.37	61
2020	Jordan Love	20.0%	75	224	4.74	NaN	118	35.5	7.21	4.52	83
2020	Jacob Eason	19.6%	77	231	4.89	NaN	110	27.5	7.50	4.75	84
2020	Shea Patterson	18.7%	72	212	4.71	NaN	116	31.0	7.14	4.50	91
2020	Steven Montez	7.3%	76	231	4.68	NaN	117	33.0	7.25	4.43	144
2020	Brian Lewerke	7.2%	74	213	4.95	NaN	NaN	NaN	7.14	4.40	145

Justin Herbert checks all the Combine boxes as a tall, big, relatively fast and quick quarterback. Even then, the model only gives him a coin-flip chance at success based only on measurables. Jalen Hurts is shockingly fifth in the 2020 class, though he’s likely penalized significantly for skipping the agility drills despite his top-level speed and jumps.

Running Backs

The sample decision tree for running backs is more complex, mostly due to the larger sample of players at the position who participate at the combine in comparison to quarterbacks. It’s also the case that athleticism matters more for running backs, with the top three leafs (final nodes) of the tree with average success rates between 0.44 and 0.6. The first (most important) split is at 4.4 seconds in the 40-yard dash, with players below that mark finding fantasy success 60% of the time. Next, the broad jump, weight, three-cone drill and vertical jump all factor into various splits. Again, some of the splits directionally go against intuition, but that’s why we build a forest of decision trees for our final numbers below.

Top-10 Historical Success Probabilities: Running Backs (2006-2017)

Year	Pick	Player	Success	Success Prob	Height	Weight	Forty	Bench	Broad	Vertical	3cone	Shuttle
2007	7	Adrian Peterson	Yes	77.5%	73	217	4.40	NaN	127	38.5	7.09	4.40
2008	73	Jamaal Charles	Yes	76.1%	71	200	4.38	NaN	122	30.5	6.80	4.22
2016	73	Kenyan Drake	Yes	74.9%	73	210	4.45	10	123	34.5	7.04	4.21
2008	55	Ray Rice	Yes	73.6%	68	199	4.42	23	119	31.5	6.65	4.20
2017	8	Christian McCaffrey	Yes	73.3%	71	202	4.48	10	121	37.5	6.57	4.22
2008	44	Matt Forte	Yes	72.5%	73	217	4.44	23	118	33.0	6.84	4.23
2006	30	Joseph Addai	Yes	71.8%	71	214	4.40	18	125	38.5	7.09	4.47
2013	48	Le'Veon Bell	Yes	71.8%	73	230	4.60	24	118	31.5	6.75	4.24
2011	73	Stevan Ridley	Yes	70.0%	71	225	4.65	18	118	36.0	6.78	4.21
2016	45	Derrick Henry	Yes	69.8%	75	247	4.54	22	130	37.0	7.20	4.38

The random forest model is much more accurate at running back, with all of the top-10 in success probability being actual successes in the NFL. Some of this is the fact that the model was fit on the same data that it is now predicting, but it’s also due to the strength of athleticism in predicting positional success. The tenth-highest predicted probability for running backs (Derrick Henry, 69.8%) is higher than the top probability at quarterback.

2020 Running Backs

Year	Player	Success Prob	Height	Weight	Forty	Bench	Broad	Vertical	3cone	Shuttle
2020	Jonathan Taylor	45.8%	70	226	4.39	17	123	36.0	7.01	4.24
2020	AJ Dillon	37.6%	72	247	4.53	23	131	41.0	7.19	NaN
2020	Brian Herrian	33.8%	71	209	4.62	18	126	38.5	7.12	4.40
2020	Darius Anderson	32.2%	70	208	4.61	19	128	36.0	NaN	4.19
2020	Cam Akers	29.6%	70	217	4.47	20	122	35.5	NaN	4.42
2020	JaMycal Hasty	29.4%	68	205	4.55	15	123	39.0	NaN	4.03
2020	Anthony McFarland	26.9%	68	208	4.44	NaN	116	29.5	NaN	NaN
2020	Darrynton Evans	25.4%	70	203	4.41	20	125	37.0	NaN	NaN
2020	Rico Dowdle	24.0%	71	213	4.54	NaN	127	38.0	NaN	NaN
2020	James Robinson	22.4%	69	219	4.64	24	125	40.0	7.03	4.19
2020	Scottie Phillips	21.7%	68	209	4.56	29	114	30.0	7.40	4.53
2020	Joshua Kelley	20.0%	70	212	4.49	23	121	31.0	6.95	4.28
2020	Raymond Calais	19.6%	67	188	4.42	20	120	37.5	NaN	NaN
2020	Eno Benjamin	19.0%	68	207	4.57	12	122	39.0	6.97	4.25
2020	Javon Leake	17.8%	72	215	4.65	NaN	125	28.5	NaN	NaN
2020	LeVante Bellamy	16.9%	68	192	4.50	16	125	39.5	NaN	NaN
2020	Clyde Edwards-Helaire	16.9%	67	207	4.60	15	123	39.5	NaN	NaN
2020	Patrick Taylor	16.5%	73	217	4.57	15	123	34.0	NaN	4.34
2020	Tony Jones Jr.	14.6%	70	220	4.68	13	119	32.5	7.18	4.21
2020	Sewo Olonilua	13.8%	74	232	4.66	25	123	36.0	NaN	4.28
2020	D'Andre Swift	12.7%	68	212	4.48	NaN	121	35.5	NaN	NaN
2020	J.J. Taylor	12.6%	65	185	4.61	19	118	34.5	7.00	4.15
2020	Ke'Shawn Vaughn	9.7%	69	214	4.51	NaN	117	32.0	NaN	NaN
2020	DeeJay Dallas	6.6%	70	217	4.58	NaN	119	33.5	7.18	4.32
2020	Salvon Ahmed	4.1%	70	197	4.62	NaN	120	34.5	NaN	NaN
2020	Benny LeMay	3.8%	68	221	4.75	24	112	34.0	NaN	NaN
2020	Lamical Perine	3.7%	70	216	4.62	22	118	35.0	7.13	4.31
2020	Zack Moss	1.6%	69	223	4.65	19	NaN	33.0	NaN	4.37

Jonathan Taylor performs best in the fantasy success model, though at 45.8% probability perhaps not as well as you would have suspected. A.J. Dillon isn’t far behind in the class, while PFF’s pre-Combine RB1, Zack Moss, was given the lowest probability of success by the model at a paltry 1.6%.

Wide Receivers

The sample tree for wide receivers splits first at weight (200 pounds), then broad jump (120-121 inches), and for lower leaves short shuttle and bench press. Looking back to the 2006-2017 sample could be overvaluing receiver size, as it’s become less of a driver in today’s space-seeking NFL offenses. Still, the big prototype WR1s are probably the best bets to attract targets and see fantasy success early in their careers.

Top-10 Historical Success Probabilities: Wide Receivers (2006-2017)

Year	Pick	Player	Success	Success Prob	Height	Weight	Forty	Bench	Broad	Vertical	3cone	Shuttle
2018	91	Trequan Smith	No	62.2%	74	203	4.49	12	130	37.5	6.97	4.50
2018	174	Marquez Valdes-Scantling	No	56.6%	76	206	4.37	15	124	30.5	NaN	NaN
2012	43	Stephen Hill	No	52.8%	76	215	4.36	14	133	39.5	6.88	4.48
2014	90	Donte Moncrief	No	51.2%	74	221	4.40	13	132	39.5	7.02	4.30
2007	2	Calvin Johnson	Yes	48.8%	77	239	4.35	NaN	NaN	NaN	NaN	NaN
2013	34	Justin Hunter	No	45.8%	76	196	4.44	NaN	136	39.5	NaN	4.33
2015	7	Kevin White	No	44.2%	75	215	4.35	23	123	36.5	6.92	4.14
2008	84	Harry Douglas	No	42.7%	71	176	4.51	NaN	120	31.0	6.57	4.12
2007	157	David Clowney	No	42.5%	72	188	4.36	NaN	123	32.5	7.00	4.15
2009	124	Louis Murphy	No	41.8%	74	203	4.32	12	NaN	NaN	NaN	NaN

The wide receiver success model was fairly awful at finding successful wide receivers, only hitting on Calvin Johnson. The other athletic marvels on the top-10 list weren’t able to translate those gifts into fantasy production, another sign that we shouldn't read too much into wide receiver results from the Combine.

2020 Wide Receivers

Year	Player	Success Prob	Height	Weight	Forty	Bench	Broad	Vertical	3cone	Shuttle
2020	Denzel Mims	46.3%	74	207	4.38	16	131	38.5	6.66	4.43
2020	Chase Claypool	44.1%	76	238	4.42	19	126	40.5	NaN	NaN
2020	Isaiah Hodgins	39.2%	75	210	4.61	9	124	36.5	7.01	4.12
2020	Michael Pittman Jr.	36.8%	76	223	4.52	13	121	36.5	6.96	4.14
2020	Dezmon Patmon	29.9%	75	225	4.48	15	132	36.0	7.28	4.38
2020	Juwan Johnson	29.3%	76	230	4.58	14	124	33.0	6.94	4.37
2020	Brandon Aiyuk	28.6%	71	205	4.50	11	128	40.0	NaN	NaN
2020	Jalen Reagor	26.5%	70	206	4.47	17	138	42.0	7.31	4.46
2020	Tyrie Cleveland	23.1%	74	209	4.46	13	126	39.5	NaN	NaN
2020	Jerry Jeudy	22.4%	73	193	4.45	NaN	120	35.0	NaN	4.53
2020	Henry Ruggs III	22.3%	71	188	4.27	NaN	131	42.0	NaN	NaN
2020	Antonio Gandy-Golden	21.9%	76	223	4.60	22	127	36.0	7.33	4.55
2020	Quez Watkins	21.3%	72	185	4.35	NaN	125	36.5	7.28	4.36
2020	Darnell Mooney	20.6%	70	176	4.38	9	124	37.0	NaN	NaN
2020	Isaiah Coulter	19.6%	73	198	4.45	NaN	121	36.0	7.28	4.62
2020	Donovan Peoples-Jones	19.2%	73	212	4.48	NaN	139	44.5	NaN	NaN
2020	Devin Duvernay	19.1%	70	200	4.39	NaN	123	35.5	7.13	4.20
2020	Gabriel Davis	16.8%	74	216	4.54	14	124	35.0	7.08	4.59
2020	Justin Jefferson	15.9%	73	202	4.43	NaN	126	37.5	NaN	NaN
2020	Antonio Gibson	15.7%	72	228	4.39	16	118	35.0	NaN	NaN
2020	Kalija Lipscomb	14.7%	71	207	4.57	16	127	32.0	NaN	NaN
2020	K.J. Osborn	14.7%	71	203	4.48	18	123	37.5	7.00	4.35
2020	Kendrick Rogers	13.8%	76	208	4.51	17	124	35.5	7.13	4.48
2020	Chris Finke	13.4%	69	186	4.57	7	NaN	40.0	NaN	NaN
2020	Binjimen Victor	11.5%	75	198	4.60	9	128	35.0	7.10	NaN
2020	Jeff Thomas	10.7%	68	170	4.45	NaN	125	36.5	NaN	NaN
2020	Stephen Guidry	10.6%	75	201	4.47	NaN	125	34.0	7.31	4.46
2020	Omar Bayless	10.6%	72	212	4.62	11	123	36.0	7.35	4.50
2020	Joe Reed	10.5%	72	224	4.47	21	123	38.0	NaN	NaN
2020	Trishton Jackson	9.9%	72	197	4.50	NaN	117	36.0	NaN	NaN
2020	Marquez Callaway	9.8%	73	205	4.55	NaN	126	38.0	NaN	NaN
2020	Jauan Jennings	9.6%	75	215	4.72	NaN	119	29.0	NaN	NaN
2020	Cody White	8.8%	75	217	4.66	NaN	120	35.5	7.19	4.52
2020	Malcolm Perry	8.5%	69	186	4.63	10	122	36.0	7.12	4.31
2020	CeeDee Lamb	8.5%	73	198	4.50	11	124	34.5	NaN	NaN
2020	Aaron Parker	8.0%	73	209	4.57	12	112	26.5	6.94	4.23
2020	Tony Brown	6.9%	72	198	4.65	14	119	33.5	7.21	4.27
2020	Quintez Cephus	5.7%	72	202	4.73	23	124	38.5	7.20	4.33
2020	Freddie Swain	5.5%	72	197	4.46	16	124	36.5	7.05	4.26
2020	John Hightower	5.5%	73	189	4.43	NaN	122	38.5	7.07	4.21
2020	Aaron Fuller	3.9%	70	188	4.59	NaN	118	34.0	7.14	NaN
2020	Quartney Davis	3.0%	73	201	4.54	NaN	NaN	35.5	NaN	NaN
2020	Laviska Shenault Jr.	2.7%	72	227	4.58	17	NaN	NaN	NaN	NaN
2020	Austin Mack	2.4%	73	208	4.59	NaN	117	31.5	NaN	4.42
2020	K.J. Hill	0.8%	71	196	4.60	17	114	32.5	NaN	NaN

Buzz is building post-Combine for names at the top of the list like Denzel Mims and Chase Claypool, but the wide receiver model still sees them at less than 50% for fantasy success. I’m not suggesting that you ignore the results of the Combine, as they’ll at least affect draft positions. But caution is probably in order before thinking about adding them to dynasty rosters ahead of more accomplished receivers.

Tight Ends

The sample decision tree for tight ends first splits at 40-yard dash (4.5 seconds), then for the tight ends who don’t make that steep threshold at three-cone drill (7.2 seconds), and broad jump. In fantasy football, we don’t care much about tight-end versatility or how well they block, but the ability to stretch the field and get open with quickness are paramount.

Top-10 Historical Success Probabilities: Wide Receivers (2006-2017)

Year	Pick	Player	Success	Success Prob	Height	Weight	Forty	Bench	Broad	Vertical	3cone	Shuttle
2017	19	O.J. Howard	No	58.9%	78	251	4.51	22	121	30.0	6.85	4.16
2008	141	Gary Barnidge	No	50.4%	78	243	4.61	22	117	31.0	6.92	4.23
2015	204	Darren Waller	No	46.9%	78	238	4.46	12	125	37.0	7.07	4.25
2013	222	Chris Gragg	No	43.4%	75	244	4.50	18	125	37.5	7.08	4.51
2012	34	Coby Fleener	No	42.7%	78	247	4.51	27	NaN	NaN	NaN	NaN
2007	31	Greg Olsen	No	40.8%	78	254	4.51	23	114	35.5	7.04	4.48
2006	72	Leonard Pope	No	39.0%	80	258	4.62	22	118	37.5	7.47	4.67
2008	300	Darrell Strong	No	38.9%	76	268	4.78	17	110	25.0	6.86	4.36
2008	158	Kellen Davis	No	38.2%	78	262	4.59	22	118	28.0	7.25	4.38
2015	117	Blake Bell	No	36.7%	78	252	4.8	14	116	33.0	6.85	4.32

The threshold of fantasy success for tight ends in this analysis is tougher at TE8, but the model going 0-for-10 isn’t a great endorsement. It has identified move tight ends, some of whom were eventually successes later in their careers, like Greg Olsen and Darren Waller. If anything, the model’s lack of accuracy points to some of the randomness of the position and how closely it’s linked to quarterback and touchdown scoring.

2020 Tight Ends

Year	Player	Success Prob	Height	Weight	Forty	Bench	Broad	Vertical	3cone	Shuttle
2020	Adam Trautman	37.6%	77	255	4.80	18	114	34.5	6.78	4.27
2020	Albert Okwuegbunam	35.7%	77	258	4.49	NaN	NaN	NaN	NaN	NaN
2020	Dalton Keene	23.9%	76	253	4.71	21	125	34.0	7.07	4.19
2020	Stephen Sullivan	22.9%	76	248	4.66	NaN	123	36.5	7.51	4.62
2020	Charlie Taumoepeau	22.5%	74	240	4.75	18	121	36.5	7.00	4.27
2020	Charlie Woerner	15.1%	76	244	4.78	21	120	34.5	7.18	4.46
2020	Josiah Deguara	14.0%	74	242	4.72	25	115	35.5	7.15	4.35
2020	C.J. O'Grady	13.8%	75	253	4.81	16	119	34.0	7.30	4.34
2020	Devin Asiasi	13.7%	75	257	4.73	16	115	30.5	NaN	NaN
2020	Cole Kmet	13.2%	77	262	4.70	NaN	123	37.0	7.44	4.41
2020	Colby Parkinson	12.6%	79	252	4.77	18	109	32.5	7.15	4.46
2020	Dominick Wood-Anderson	11.4%	75	261	4.92	NaN	119	35.0	NaN	NaN
2020	Hunter Bryant	9.1%	74	248	4.74	23	115	32.5	7.08	4.46
2020	Brycen Hopkins	6.2%	75	245	4.66	21	116	33.5	7.25	4.28
2020	Jared Pinkney	4.9%	76	257	4.96	23	NaN	NaN	NaN	NaN
2020	Harrison Bryant	2.9%	76	243	4.73	13	110	32.5	7.41	4.37
2020	Mitchell Wilcox	2.2%	75	247	4.88	NaN	112	31.0	7.37	4.43

You wouldn’t think a tight end with a 4.80 40-yard dash would top the list over one who posted a sub-4.5 time, but that’s what numbers say. Adam Trautman likely benefits significantly from his 6.78 three-cone time at 255 pounds, while Albert Okwuegbunam is probably hurt by skipping drills other than the 40-yard dash. Outside of those two, the model doesn’t see much chance of success for the rest of the class, at least only based on their Combine metrics.