On Saturday, the Hall of Fame selection committee will meet, lock themselves in a room, and debate the relative merits of the 15 modern-era finalists for induction. After an intense discussion, the results will be announced nationally as the final event in the festivities leading up to Sunday’s Super Bowl.
While the list of 15 finalists includes several names who have been waiting longer than they should for their call, the one that stands out the most to me is Terrell Davis, who has been a semi-finalist more than anyone else in this year’s class, reaching the top 25 ten times in his ten years of eligibility.
Hopefully the Hall of Fame committee can manage to make room for him in what could easily be a stacked class. Whatever they do this Saturday, however, will not change one simple fact: Terrell Davis should have long ago been elected to the Hall of Fame.
The Case for Terrell Davis
The positive case for Terrell Davis is already widely known, but I suspect many underestimate just how strong it really is.
Terrell Davis, in his first four seasons in the NFL, was named to three pro bowls, was a three-time first-team AP All Pro, won two AP Offensive Player of the Year awards, and was named MVP by the Associated Press.1 Davis was also a seven-time AP Offensive Player of the Week, including four times in 1998 alone; with the caveat that it appears the AP has only been issuing that award since 1990, Davis’ total is bettered by only three Hall of Famers (Barry Sanders with 11, Emmitt Smith with 10, and Marshall Faulk with 8).
His “trophy case” compares favorably to any other running back. Since the merger, (and discounting fullbacks or players who received honors for their work as a returner), Davis is one of 11 running backs to earn three first-team AP All Pro honors; the other ten are Barry Sanders (6); Eric Dickerson, Walter Payton, and O.J. Simpson (5); Adrian Peterson and Emmitt Smith (4); Earl Campbell, Marshall Faulk, Priest Holmes, and LaDainian Tomlinson (3). With the exception of Holmes and, (perhaps to some), Earl Campbell, that’s essentially a “who’s who” of modern “inner circle” Hall of Fame running backs. Again with the exception of Holmes, each of those backs was either elected to the Hall of Fame on their first ballot of eligibility, or else are not yet eligible but are widely expected to earn first-ballot status when they are.
Davis joins Peyton Manning, Drew Brees, Marshall Faulk, Barry Sanders, Jerry Rice, and Earl Campbell as the only repeat winners of the AP’s Offensive Player of the Year honors, (and if not for a 2,000 yard season by Sanders in 1997, Davis might have joined Campbell and Faulk as the only players with three consecutive).
These are some impressive accolades, and Davis’ production on the field was worthy of them. One method I like to use to compare production is an adaptation of baseball’s “black ink test” and “grey ink test”. The idea is to count how many times a player led the league or ranked in the top ten in certain key statistics. This test in a single stroke can level the playing field between players who were very good for a long time, (thus having many different chances to add a little bit of black ink / grey ink), and those who were exceptional for a short time, (thus accumulating a lot of black ink / grey ink at once).
There have been fifteen post-merger2 Hall of Fame running backs. Here is how Davis compares to those backs in top-3, top-5, and top-10 finishes in what I believe are the six key statistical categories for running backs: carries, rushing yards, yards per attempt, rushing touchdowns, yards from scrimmage, and total touchdowns.
Unsurprisingly, Davis ranks last in career top-10 finishes. He finishes right at the median for career top-5 finishes, though, (tying Faulk for 8th with 18 each), and ranks 6th in the sample in career top-3 finishes, despite accumulating all sixteen of them in a three-season span. Davis might be light on grey ink, but he sails through the black ink test with flying colors.3
The grey ink and black ink test are simplistic, but they do a phenomenal job of matching our intuitions; they clearly highlight Sanders, Payton, and Smith as a class unto themselves at the top, they clearly highlight Csonka, Riggins, and Bettis as a class unto themselves at the bottom, and they clearly illustrate the divide between the players in the middle who were very good for a long time (Dorsett, Harris, Martin), and players who were exceptional for a short time (Davis, Campbell, Simpson).4
Perhaps most interestingly, we can look at the players who fared the worst on the “grey ink”, (Davis’ weakness), to see what factors allowed them to overcome their liability and earn enshrinement.
The 10th, 11th, and 12th ranked Hall of Famers in “grey ink” are Marshall Faulk, O.J. Simpson, and Earl Campbell. All three made it into the Hall of Fame based on nearly unmatched dominance at their peak; the three combined for seven AP Offensive Player of the Year awards and each was named league MVP once by the Associated Press, (with Faulk collecting MVP awards from non-AP organizations in two years, and Campbell doing it three times).5
The 13th, 14th, and 15th ranked Hall of Famers in “grey ink” are Jerome Bettis, John Riggins, and Larry Csonka. All three made it into the Hall of Fame in large part based on their postseason narratives; Riggins and Csonka both had one of the best postseasons of all time, becoming two of just four players with 300 rushing yards and 6 touchdowns in a single postseason.6 In 2009, Chase calculated that they had the 1st and 7th most valuable individual postseasons of any running back, respectively. Jerome Bettis was born in Michigan.
As the black ink and awards case demonstrate, Davis clearly belongs alongside that first group of backs in terms of peak performance. And while Riggins and Csonka were excellent postseason backs, (and Jerome Bettis was born in Michigan), Terrell Davis was arguably the best postseason back of all time; he owns two of the top five postseasons, and by Chase’s calculation, trailed only Emmitt Smith in career postseason value. Smith accumulated 891 points of value in 17 career games. Davis accumulated 865 points in just 8 games.
In fact, how good was Terrell Davis in the postseason? Better than you think, even accounting for the knowledge that Terrell Davis was better than you think in the postseason. In his eight career playoff games, Davis rushed 204 times for 1140 yards and 12 touchdowns, (at 5.6 yards per carry), adding 19 receptions for 131 yards through the air.7 Pro-rated over a 16-game season, that 2280 rushing yards and 2542 yards from scrimmage would both break the existing NFL records, the former by a full 175 yards. And remember, these numbers came against postseason competition, the best the NFL had to offer.
If that sounds like a huge 8-game total, it is. It’s more than O.J. Simpson ran for in any 8-game stretch during his 2,000 yard season in 1973. But lets put it into better context. Pro Football Reference has game logs going back to 1960. If we remove overlapping stretches,8 there have been 18 times where a back has topped 1,100 rushing yards in 8 games.9 And, if you were to include touchdowns and rushing yards in such a stretch, Davis’ accomplishments would look even better.10
Here’s the complete data for all 18 stretches, plus Terrell Davis’ postseason career. I’ve also added a column showing how many games above .500 each back’s opponents were during their stretch.11
|Name||Year||Carries||Yards||YPC||TDs||Begins (Game #)||Ends (Game #)||Games above .500|
Terrell Davis’ 8 playoff games register as one of the greatest 8-game stretches from any running back in NFL history, (or at least since 1960), even without the additional context about the quality of the opposition or the fact that every single game was a “win or go home” contest.
If you prefer the data in per-game form, Davis’ 142.5 career rushing yards per game in the postseason laps the field; with a five-game minimum, John Riggins (110.7 yards per game) is the only player within 33% of Davis’ total. Of the 75 running backs with at least 100 career rushes in the postseason, Davis’ 5.59 yards per carry is likewise head-and-shoulders above the rest of the field; Marcus Allen, (5.04 yards per carry), is the only other player over 5.0 for his career.
In Terrell Davis’ “worst” playoff game, he still rushed for 91 yards. In NFL history, there have been 144 backs who appeared in at least 4 games and rushed for at least 200 yards; of those 144 backs, only four, (Arian Foster, John Riggins, Eric Dickerson, Emmitt Smith), averaged more yards than Davis had in his least productive game. There have been 188 backs with at least 4 playoff games and at least 200 yards from scrimmage; similarly, only four, (Arian Foster, Marcus Allen, Eric Dickerson, John Riggins), have an average better than Davis’ lowest total, (115 yards).
If we prefer not to divide up those playoff games and treat them as a separate “season” unto themselves, but rather to add them on to his existing seasons, Terrell Davis’ 1997 and 1998 become easily two of the greatest years by any running back in history. The only back with more rushing yards in a single season, (regular and postseason combined), than Terrell Davis’ 2331 in 1997 was… Terrell Davis, with 2476 in 1998. According to Chase’s methodology from 2009, Terrell Davis’ 1998 and 1997 were the 1st and 5th most valuable running back seasons in history.
And Chase’s system was hardly the only advanced statistical system to love Terrell Davis. Before the 2013 season, Football Outsiders looked back at their data, (which dated back to 1991). They found that despite his abbreviated career, Davis ranked 3rd in career rushing value, just a tenth of a percent behind LaDainian Tomlinson for second. Davis’ 1998 season the best they had measured and his 1997 season tied for second, (despite Football Outsiders not including any postseason production for either).
Terrell Davis at his best was clearly one of the greatest running backs to ever play football.12 Despite his shortened career, he packed so much production into such a short time that his regular-season resume is still Hall of Fame caliber. Advanced stats went nuts for him. He supplemented all of this with two Super Bowls and a track record as the greatest postseason running back, (and potentially the greatest postseason performer at any position— Lynn Swann has nothing on him), in NFL history.
So why isn’t Terrell Davis in the Hall of Fame yet? While the case in his favor is exceptional, the case against him is strong, as well.
The Case Against Terrell Davis
Part 1: Longevity
Due to an injury-abbreviated career, Terrell Davis ranks 55th in career rushing yards, behind names such as Chris Warren, James Brooks, and Garrison Hearst. Even the much-maligned Larry Csonka finished his career with nearly 500 more rushing yards than Terrell Davis. How can a player who doesn’t even rank in the top 50 at his own position be a Hall of Famer?
The problem with this argument? Longevity, by itself, means nothing. Vinny Testaverde and Drew Bledsoe rank 9th and 10th in career passing yards, yet neither will ever even be a finalist for the Hall of Fame. Playing for a long time isn’t enough; a player must be adding substantial value during that time.
Allow me to demonstrate with a comparison. While I hate when people argue for one player to make the Hall of Fame by comparing him to one of the worst players at his position who has already been elected, I think a comparison between Terrell Davis and Jerome Bettis, member of the class of 2015, can be illustrative.
Bettis was a Hall of Fame finalist four times from 2011 to 2014, while Davis was cut at the semifinalist stage each time. Both backs were finalists in 2015, but Bettis was inducted over Davis. The Hall of Fame voters pretty clearly expressed a preference for Bettis in each year he was eligible, so a comparison seems fair.13
Bettis made the Hall of Fame in large part because of his remarkable longevity.14 When Bettis retired in 2005, he ranked 5th in NFL history in career rushing yards. By the time he was eligible for the Hall of Fame, he still ranked 6th, (having been passed by LaDainian Tomlinson, who retired with 22 more rushing yards).
How much more need Davis have done to match Bettis’ remarkable longevity? Let’s investigate. During his peak, (1995 to 1998), Terrell Davis played 61 games, amassed 1343 carries, gained 6413 yards and 56 touchdowns, and added 152 receptions for 1181 yards and 5 more scores. In addition, as mentioned, he added 204/1140/12 rushing and 19/131/0 receiving in the postseason. While we could discuss whether the postseason performance was more valuable than the regular season performance, that’s not important— at a bare minimum, hopefully we can agree that it was equally valuable and deserves to be counted.
Combining those totals, Davis’ “peak production” spanned 69 games and consisted of 1547/7553/68 rushing and 171/1312/5 receiving. We can subtract those totals from the career production of Jerome Bettis, (including his own postseason games), then divide by the additional number of games Bettis played, and that will tell us exactly what Davis would have had to average, and for how long, to precisely match Bettis’ career totals.
Doing this, we find that Davis would have needed to play roughly 8.5 more seasons, and during those 8.5 seasons he would need to average roughly 249 carries for 792 yards, (3.2 yards per carry), catch an additional 4 passes for 23 additional yards, and score roughly 3.5 touchdowns.
Imagine Davis had given us an extra 8-9 years like this. Would this make him more of a Hall of Famer? Certainly not. Those are the definition of “replacement level” seasons; we’re far more concerned with how much value a player provides above and beyond that.
Bettis is, admittedly, an extreme example, (though as I said, I believe he is a fair one, given how he and Davis directly competed for a spot and the committee explicitly preferred him). Few backs have had as many “filler” seasons as Bettis, who was below 1,000 yards from scrimmage in five different years, accumulating 3758 total yards during that span. Had I chosen Curtis Martin as the point of comparison, Davis would have needed 1400 yards for about 7 more years. Had I chosen Thurman Thomas, Davis would have needed 1200 yards for about 8 more years.
Without the rushing yardage from those five filler seasons, Bettis falls from 6th to 28th in career rushing yards. He falls from 22nd to 90th in yards from scrimmage. Does he make the Hall of Fame without them? I think not.
Theoretically, we never want to punish a player for not having filler seasons. In practice, using career totals always does. Davis gained 983 rushing yards at 4.0 yards per carry in the 13 games he played after his knee injury. Had he added another three seasons at that pace, he would have been over 10,000 career rushing yards.
But would another three seasons of his post-injury play really have made him more deserving of enshrinement? We must be very careful of what exactly we hope to reward when we discuss “longevity”.
Part 2: Replaceability.
Much has been made of the success of running backs in Denver after Terrell Davis went down. In 1999, Olandis Gary had 1159 rushing yards in 12 games. In 2000, Mike Anderson topped 1600 yards from scrimmage in just 13 games. Clinton Portis rushed for 1500 yards in each of his first two seasons, and then in 2004 Rueben Droughns rushed for 1200 yards in the final 12 games after winning the starting job. The 2005 Broncos very nearly got 1000 yards rushing from both Mike Anderson and Tatum Bell.
The argument is that the success of all of these different backs prove that it was the system, and not any individual talent, that resulted in Terrell Davis’ incredible production.
But how successful were these backs, really? Not nearly as successful as Terrell Davis. From 1999 to 2010, Denver’s running backs combined for just one pro bowl appearance, (by Clinton Portis). From Davis’ injury in 1999 to Peyton Manning’s arrival in 2012— a span of 13 years— all Denver running backs combined, (including Portis), won four Offensive Player of the Week awards; that’s as many as Davis had in 1998 alone.
The ten seasons from 1995 to 2004 represent four years with Davis as the primary back, two years with Portis as the primary back, and four years with a hodgepodge of players. Of those ten years, the top four seasons in total rushing yards are 2003, (with Portis)… then 1998, 1997, and 1996, in that order.
In six of those seasons, Denver as a team averaged more than 4.5 yards per carry. They were 2002 and 2003, (with Portis), plus 1995, 1996, 1997, and 1998, (with Davis). The traditional statistics suggest that the Davis and Portis offenses were clearly better than the others at running the ball.
Advanced statistics tell the same story. Per Football Outsiders, here’s Denver’s year-by-year rushing DVOA rating, (a statistic to measure how much more effective a certain player or unit is relative to league average), as well as where that value ranked among the rest of the league:
1995 – 13.5% (2nd)
1996 – 13.1% (7th)
1997 – 24.0% (1st)
1998 – 31.4% (1st)
1999 – 7.4% (10th)
2000 – 12.8% (6th)
2001 – 2.0% (13th)
2002 – 22.3% (3rd)
2003 – 9.0% (8th)
2004 – 4.0% (13th)
Denver topped 13% DVOA rushing in each of Davis’ four healthy seasons. They only topped it once in the six subsequent years, with Clinton Portis, (who was also pretty good). Denver ranked in the top 5 in rushing efficiency four times, and three of them were with Davis, (the other with Portis). Denver’s average rushing DVOA before Davis’ injury was 20.5%, and its average rushing DVOA after was 9.6%.
More evidence that Terrell Davis was a lot better than Olandis Gary and Mike Anderson comes from the 2001 season, when a shadow-of-his-former-self Terrell Davis in his final eight games of his career kept both on the bench. In the eight games Davis played that season, he rushed 167 times for 701 yards, (4.2 yards per carry). In the eight games he missed, Anderson rushed 141 times for 537 yards, a 3.8 yards per carry clip (in the games Davis played, Anderson received just 34 carries).15
Brian Burke, (formerly of Advanced Football Analytics, now of ESPN), created an “Estimated Points Added”, (or EPA), model of production that measures how much each individual play contributes to a team’s bottom line. Unfortunately, Burke’s statistics don’t go back to before 1999, so we can never know how valuable Davis was by that method. It’s fairly safe to say that the Broncos would have ranked rather high, though.
In the 6 years after Davis’ injury, Denver ranked 12th, 9th, 18th, 4th, 2nd, and 9th in rushing EPA. Outside of the two exceptional Clinton Portis years, Denver typically settled in the “above average” range for rushing production. Which means Davis’ Hall of Fame candidacy is being dinged because (A) three years after his injury Denver drafted a genuinely great running back, and (B) Denver got good but unspectacular production from Davis’ immediate replacements.
This is such a loaded argument, though. The year after Jim Brown retired, Cleveland sent not one but two running backs to the Pro Bowl; one of Brown’s replacements, Leroy Kelly, would make three straight first-team AP All Pro teams, make the All-Decade squad, and eventually make the Hall of Fame.
Joe Montana’s direct replacement, Steve Young, would likewise make three straight first-team AP All Pro squads, win two league MVPs, and make the Hall of Fame. Steve Young’s direct replacement, Jeff Garcia, would earn a trip to three straight pro bowls after Young retired. In fact, in his first full year as a starter, Garcia would pass for 4,000 yards and rush for 400 yards, becoming just the second player in history to reach both marks behind… Steve Young.
It could be argued that Montana was himself a successor of Ken Anderson, who many consider the best player not currently in the Hall of Fame. Marshall Faulk had two direct successors in his career, (Edgerrin James and Steven Jackson), and both of them accumulated 15,000 career yards from scrimmage. St. Louis finished in the top 5 in passing yards in 2002, 2003, 2004, 2005, and 2006 as they moved on from Kurt Warner to Marc Bulger. Two years after Barry Sanders retired, James Stewart had 1400 yards from scrimmage with Detroit.
How many times was Jeff Garcia mentioned when Steve Young was brought to the floor for debate? I’m not on the Hall of Fame committee, so I don’t know. But I’d venture it was probably as many times as he deserved to be mentioned: zero. Steve Young’s Hall of Fame case depended on the merits of Steve Young’s play and nothing more.
Similarly, Terrell Davis’ candidacy should rest on how good Terrell Davis himself was, and the answer is “exceptionally good”. His peak was incandescent, including two of the greatest seasons in NFL history by any measure. His trophy case is stuffed full and his profile is positively littered with black ink, finishing in the top half of modern Hall of Famers at the position. He was perhaps the greatest postseason performer in NFL history and he retired with two rings and a Super Bowl MVP for one of the most iconic big-game performances the league has seen, a 160-yard, three-touchdown affair, (including the game-winner), where Davis missed literally the entire second quarter to a migraine. Davis did more in his four seasons before injury robbed him of his ability than most backs do in an entire, full career.16
The Hall of Fame has made exceptions for brilliant players with short careers in the past. Gale Sayers, Paul Hornung, Kellen Winslow Sr., Dwight Stephenson, Earl Campbell, Lynn Swann, Floyd Little, and Dick Butkus all had much shorter careers than their peers. At the time of his enshrinement, (on the first ballot, no less), Roger Staubach ranked 31st in career passing yards, behind names like Steve Bartkowski, Jim Plunkett, and Charley Johnson.
I am not saying that the Hall of Fame should make another such exception for Terrell Davis now. They should have already done so a decade ago.17 But at least it’s not too late to correct their oversight.
- The majority of this article will be AP-centric, simply because that is widely accepted as the “most legitimate” or “most important” postseason honor; for those of you who feel that the other awards are just as legitimate, (and I count myself among your number), note that Davis fared well across the board; all five naming institutions listed him as a first-team All Pro in 1996, all four institutions named him a first-team All Pro in 1997 and 1998, and the Pro Football Writers of America likewise gave Davis their MVP award in 1998, (though the Bert Bell Player of the Year award went to Randall Cunningham, his third). The UPI also named Davis second-team All Conference in 1995. Incidentally, Davis also finished 3rd in MVP voting in 1996, behind eventual winner Brett Favre and teammate John Elway. [↩]
- Simpson was drafted in 1969, but obviously the majority of his career came after the merger. Two other Hall of Fame running backs played part of their career after 1970, but are excluded here. Leroy Kelly accumulated more than half of his career production prior to the merger, while Floyd Little entered the Hall of Fame via the seniors’ process, which makes him a poor point of comparison. [↩]
- An interesting, (but harder-to-quickly-calculate), alternative to the “black ink test” is a “yards above X” model. Doug Drinen in 2006, for instance, calculated how many yards above the 10th-leading rusher every running back since 1970 accumulated. Davis finished 7th. In yards from scrimmage, Davis finished 10th. In the summer of 2015, Chase did the same thing but used a lower threshold of “worst starter.” Obviously Davis fared worse here, but still respectable, ranking 31st in rushing yards (largely on the strength of two top-30 seasons) and 33rd in yards from scrimmage. Davis also ranks 28th in career VBD, which uses RB24 as the baseline. [↩]
- Looking at the individual categories also is a neat way to see a player’s relative strengths. Emmitt Smith was a touchdown machine, Barry Sanders lapped the field in yards per carry, Walter Payton was a yards from scrimmage monster. Perhaps most interesting to me is Larry Csonka; he played in a smaller league, but his six seasons in the top 10 in yards per carry trails only Barry Sanders’ nine, an outcome I certainly was not expecting. [↩]
- Of note: Davis, Simpson, and Campbell are three of the five running backs to average 110+ rushing yards per game over a three-year stretch. [↩]
- The other two are Davis himself and Hall of Famer Franco Harris. Emmitt Smith also once had 298 rushing yards and 6 touchdowns in the postseason. [↩]
- He also contributed a 2-point conversion. [↩]
- For example: in 2012, Adrian Peterson topped 1,100 rushing yards in every 8-game stretch from game 5 to game 17, six such stretches in total. I have removed all but his most prolific stretch, the 1322 rushing yards from week 9 to week 16. [↩]
- While we’re on the subject, if you don’t remove overlapping stretches, Terrell Davis’ 11 stretches with 1,000 yards in eight consecutive games is tied with Barry Sanders for second behind only Eric Dickerson’s 12. That’s not counting Davis’ 8 playoff games as a stretch. [↩]
- If you’ll permit a bit of endpoint cherrypicking, there have only been four stretches since 1960 where a back topped 1100 rushing yards and 12 touchdowns in eight games. Those stretches belong to Ladainian Tomlinson in 2006, Larry Johnson in 2005, Eric Dickerson in 1983, and… Terrell Davis himself, with 1150 rushing yards and 14 rushing touchdowns in the first eight weeks of 1998. [↩]
- “Games above .500” includes the regular season record of all teams each back faced in his respective 8-game run, outside of those eight games. For example: the eight teams Adrian Peterson faced to end the 2012 season had a combined record 25 games above .500 in the regular season, but that includes the fact that they went 3-5 against the Vikings during Peterson’s big stretch. Removing those 8 games from the sample improves the teams Peterson faced to 27 games above .500. On the other end of the spectrum, the teams O.J. Simpson faced to end 1976 were 16 games above .500 for the year, but Simpson’s Bills amazingly went 0-8 despite O.J.’s 1127 rushing yards; subtracting those games from the sample drops the combined record of Simpson’s opponents to 8 games above .500. Note that all numbers include regular-season records only, except for Terrell Davis’ playoff numbers, which include opponents’ postseason wins and losses as well. [↩]
- He also rushed for at least 50 yards in a game in 87% of his career games, the best rate since at least 1960. And Davis also rushed for 75+ or 100+ yards at record or near-record rates, too. [↩]
- Chase also mentioned both Bettis and Davis in the context of the HOF here. [↩]
- He made it in small part due to being born in a city where they very occasionally play significant football games. [↩]
- Gary couldn’t get on the field over either; he finished the season with 57 carries for 228 yards. [↩]
- The Broncos also went 49-7 when Davis had 20 carries, and that’s an incredible winning percentage even when you consider that every running back has a good record when they get 20+ carries. [↩]
- For those curious, Davis ranked as the 16th best running back in the Football Perspective Wisdom of Crowds experiment. [↩]