SITE NEWS: We are moving all of our site and company news into a single blog for Sports-Reference.com. We'll tag all PFR content, so you can quickly and easily find the content you want.
Also, our existing PFR blog rss feed will be redirected to the new site's feed.
Pro-Football-Reference.com ยป Sports Reference
For more from Chase and Jason, check out their work at Football Perspective and The Big Lead.
ShareThisMost Dominant RBs: Playoff Edition
On Monday, I explained the system I've used to rank every running back in every season in NFL history. On Tuesday, I showed the most dominant 50 RB seasons in NFL history along with the top RB seasons for each franchise. Yesterday, we looked at the career list -- the 100 most statistically dominant RBs in NFL history. Today I want to discuss some of the most dominant -- and memorable -- postseason performances by any running back. Tomorrow, check in for an updated version of Tuesday's and Wednesday's lists, with playoff performances included.
For single game performances, the conversation starts and ends with Keith Lincoln's performance in the 1963 AFL Championship Game. Lincoln totaled 329 yards and two scores, easily the most yards from scrimmage in any playoff game.
When you think of great post-season performances, there are three that stand out from the crowd. John Riggins, 1982. Terrell Davis, 1997. Marcus Allen, 1983. Not only did all three capture Super Bowl MVP honors, but the three RBs rushed for over 100 yards in all of their 11 post-season victories.
rsh rshyd rshtd rec recyd rectd year rd opp Riggins 25 119 0 0 0 0 1982 w det Riggins 37 185 1 0 0 0 1982 d min Riggins 36 140 2 0 0 0 1982 c dal Riggins 38 166 1 1 15 0 1982 s mia Allen 13 121 2 5 38 0 1983 d pit Allen 25 154 0 7 62 1 1983 c sea Allen 20 191 2 2 18 0 1983 s was Davis 31 184 2 4 11 0 1997 w jax Davis 25 101 2 1 17 0 1997 d kan Davis 26 139 1 1 2 0 1997 c pit Davis 30 157 3 2 8 0 1997 s gnb
Riggins rushed for over 600 yards, Allen totaled over 190 yards per game and Davis rushed for 8 TDs. And while it's easy to remember these three historical postseasons, how do we rank every postseason performance ever?
It's not that hard; I'm going to use mostly the same formula that I used to rank each RB season. One note: I'm going to weigh all Super Bowl games twice -- they're so important and such a part of post-season lore that they deserve extra weight. I'm going to rate each post-season game each RB played relative to the league average that season (excluding that RB from the league average). There's no pro-rating here -- if you played four post-season games or one, you get what you get.
Let's use Riggins and my boy Keith Lincoln as examples. Lincoln had 206 adjusted rushing yards (206/0 fumbles), while the average starting RB that season averaged 47 adjusted rushing yards per game. So Lincoln's +159 in the rushing category. He scored 2 TDs, while the average RB scored 0.57 TDs/game; so Lincoln's up 1.43 touchdowns, or +29 adjusted yards. He had 7 catches and 123 receiving yards (133.5 ACY) in the championship game; the league average RB had 32.3 adjusted catch yards per game. So Lincoln added 101 adjusted catch yards over average, giving him a total of 289 adjusted yards over average. Wow.
How about Riggins? Remember we're counting his SB performance twice. So he's got 5 games played, 776 rushing yards, 0 fumbles, 2 receptions, 30 receiving yards and 5 TDs. That's 155.2 ARY/G (league average was 56.7) and 1.0 TD/G (league average was 0.68); his receiving numbers were obviously below average and therefore ignored. So he averaged 98.5 more rushing yards per game than average, over five games; that's +493; he scored 0.32 more TD/game over five games, so that's +1.6 TDs and +32 adjusted yards, for a grand total of 525 adjusted yards over average. That's the best mark in post-season history, over Davis (who had five fumbles and only two fumble recoveries in the '97 post-season) and Allen (who played one fewer game but was better on a per game basis). Here's the list of the top 50 post-season performances of all time.
g ARY/G TD ACY/G RSHV TDV CATV VALUE
John Riggins 1982 was 5 155 5 7 492 1.6 0 525
Terrell Davis 1997 den 5 133 11 12 332 8.2 0 497
Marcus Allen 1983 rai 4 152 7 40 350 4.1 15 447
Thurman Thomas 1990 buf 4 119 5 59 262 2.6 116 431
Terrell Davis 1998 den 4 149 3 32 319 0.7 20 354
Timmy Smith 1987 was 4 137 4 5 315 1.6 0 347
Larry Csonka 1973 mia 4 113 8 0 212 5.7 0 327
Franco Harris 1974 pit 4 113 7 2 236 4.5 0 325
Emmitt Smith 1993 dal 4 97 6 47 157 4.2 56 297
Keith Lincoln 1963 sdg 1 206 2 134 159 1.4 101 289
Marshall Faulk 2001 ram 4 98 3 49 138 1.0 79 236
Wilbert Montgomery 1980 phi 4 83 3 61 105 0.8 108 229
Emmitt Smith 1995 dal 4 87 8 18 108 5.7 0 221
Eddie George 1999 oti 5 99 5 25 176 2.2 0 220
Roger Craig 1988 sfo 4 68 2 73 46 0.0 165 212
Emmitt Smith 1992 dal 4 92 5 35 137 2.8 11 204
John Riggins 1983 was 4 93 7 1 118 4.2 0 203
Franco Harris 1979 pit 4 65 5 60 34 2.3 109 188
Matt Snell 1968 nyj 3 104 2 36 155 0.3 25 186
Joe Morris 1986 nyg 4 89 5 14 138 2.4 0 186
Earnest Byner 1987 cle 2 70 4 86 24 2.8 100 181
Dorsey Levens 1997 gnb 4 89 2 50 87 0.0 90 177
Natrone Means 1996 jax 3 119 2 17 168 0.2 0 172
Ricky Watters 1993 sfo 2 78 6 49 37 5.1 30 169
Ickey Woods 1988 cin 4 97 3 0 158 0.5 0 169
Chuck Foreman 1976 min 4 72 3 57 47 0.7 105 166
Merril Hoge 1989 pit 2 110 2 51 109 0.8 36 161
Eric Dickerson 1985 ram 2 135 2 6 142 0.6 0 155
Franco Harris 1975 pit 4 93 2 30 150 0.0 3 153
Freeman McNeil 1986 nyj 2 116 3 31 119 1.7 0 153
Kenneth Davis 1992 buf 5 91 2 25 152 0.0 0 152
Steve Van Buren 1949 phi 1 196 0 0 147 0.0 0 147
Roger Craig 1989 sfo 4 83 4 32 114 1.6 0 146
Elmer Angsman 1947 crd 1 159 2 - 3 115 1.3 0 142
Thurman Thomas 1989 buf 1 27 2 170 - 27 1.4 137 138
Jamal Lewis 2000 rav 5 88 5 2 104 1.6 0 137
Paul Lowe 1960 sdg 1 165 1 10 129 0.4 0 137
Curtis Martin 1996 nwe 4 77 6 25 61 3.7 0 135
Fred Taylor 1999 jax 2 123 2 30 118 0.8 0 135
Ottis Anderson 1990 nyg 4 88 2 7 135 0.0 0 135
Freeman McNeil 1982 nyj 3 100 1 11 129 0.0 0 129
Duane Thomas 1971 dal 4 75 4 14 94 1.7 0 128
Norm Standlee 1941 chi 2 72 4 19 69 3.0 0 128
Tony Nathan 1984 mia 4 38 2 87 - 99 0.0 225 126
Ricky Bell 1979 tam 2 113 2 8 114 0.6 0 126
Thurman Thomas 1993 buf 4 51 6 52 - 30 4.1 73 125
Thomas Jones 2006 chi 4 97 4 14 97 1.3 0 123
George McAfee 1941 chi 2 88 1 38 101 0.0 22 123
Earnest Byner 1985 cle 1 161 2 31 97 1.3 0 123
Robert Smith 1999 min 2 102 1 53 77 0.0 45 122
- Notice that Earnest Byner's 1987 postseason -- you know, the one Jeremiah Castille ended -- ranks in the top twenty-five. No one remembers it anymore, but Byner totaled 345 yards from scrimmage and four touchdowns in Cleveland's two playoff games that year.
- George McAfee rushed for 200 yards, had 69 receiving yards and scored a touchdown for the Bears in two 1941 playoff games ... and he wasn't even the best RB on Chicago! Norm Standlee had him beat: he scored two touchdowns in each playoff games for Chicago. In the final week of the regular season, McAfee caught the go-ahead touchdown pass from Sid Luckman the day the Japanese struck Pearl Harbor. After the Bears won the title, both McAfee and Standlee left the NFL and served in World War II.
- In addition to Lincoln's terrific one game performance, Steve Van Buren, Elmer Angsman, Thurman Thomas, Paul Lowe and Earnest Byner make the list based on one terrific playoff game. Van Buren's legendary 31-carry, 196 rushing yard performance helped the Eagles run 70 plays to Los Angeles' 52. The host Rams hadn't scored fewer than 27 points at home that season, but were shutout in the title game.
- We all remember Timmy Smith's 200 rushing yard, two touchdown performance in the Super Bowl, but he was the Redskins top rusher in all three playoff wins. He rushed for 342 yards on 51 post-season carries, for an awesome 6.7 yards per carry average. Among RBs with 40 or more carries in a single post-season, only Marcus Allen's 8.0 YPC average in '83 (466 rushing yards on 58 carries) was higher.
- There's one other guy who sort of gets lost in the mix when you think of great RB performances. That's because he was great all the time -- Emmitt Smith. In just over one season's worth of games (seventeen), Emmitt rushed 349 times for 1586 yards and 19 TDs. He averaged over 100 yards from scrimmage per game and scored 1.2 touchdowns per game with just two net fumbles.
This entry was posted on Thursday, April 9th, 2009 at 8:06 am and is filed under History. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.
Very interesting, thanks for this!
Terrell Davis's numbers would have been even better if he hadn't sat out the entire 2nd quarter of SB 32.
I also find it interesting that Natrone Means makes it for '96 JAX but not for '94 SD and that some all-time greats like Barry Sanders and Walter Payton miss the cut.
I think it would be better if you weighed superbowls by 1.5 or championship games by 1.25 and superbowls by 1.5 (or 2). Will you post a career playoff list?
Interesting list. I like the idea of having the Super Bowl count as 2 games rather then just 1. I agree with pm that a career playoff list would be nice to see as well.
Riggins benefits from the strike in 1982, which is the only season so far where the #1 seed has had to play 4 playoff games to win the Super Bowl. If we throw out the game against Detroit, his yards per game rises to 164.3, which is now 107.6 above average, but with only 4 games, that leads to 430.4. His 5 touchdowns over 4 games turn into 45.6 yards, for a total of 476, dropping him all the way to #2 on the list.
Has anyone noticed the TD numbers are missing from each player's playoff gamelogs? I think it happened when all the new data was added.
Great series, but why does the Super Bowl count double but the NFL (or AFL) championship prior to the Super Bowl era count 1? I get that the championship counts more than the prelims, but there is plenty of lore from the pre-Super Bowl days.
I'm behind the times, since this is really a response to the first RB post, but here's something that seems odd about your system. Suppose there are three players who play 48 games in their careers and all end up with exactly the same stats. But suppose their games played are distributed as: A) three 16 game seasons (no interruptions or games missed), B) three 16 game seasons but missing one year in the middle, or C) four seasons playing 12 out of 16 games. I think you treat A and B the same, but not C, and don't understand the logic.
The big reason why I counted the Super Bowl twice and other championship games as a regular playoff game was because of NFL expansion. Winning the NFL title in 1958 involved being better than 11 other times; that's less than the number of teams you need to beat to win the AFC now. Once the leagues crowned a common champion, that became the gold standard. Being the best out of 32 (or 31 or 28) was, IMO, deserving of something special.
A career playoff list is hard because everything is rated on a pre-game basis. A player who gets in a couple of playoff games at the end of his career will look pretty bad. The best thing I could think of is the last paragraph in Friday's post.
mtr,
Here's my best explanation. Assume that a team wins 60% of its games with your RB playing and only 45% of its games when he's injured. If a RB plays in four 12-game seasons, we'd project 36 wins. If a RB plays in three 16 game seasons, we'd expect his team to win 28.8 games. What about in that fourth season? I'd project an average season, of 8 wins. So 36.8 wins is better than 36.
Now maybe you're thinking why am I making the team average in the fourth full season but below average in the four four game seasons? Because when a RB gets injured, a team is hampered, and it has to play its second best RB. When a RB retires, they can plan around this and sign or draft another RB. Perhaps more importantly, the goal should be to make the playoffs and more. Being great for 16 games three times is more likely to result in playoff wins and SB appearances than being great for 12 games four times.
Do you buy that, mtr?
Thurman Thomas being up there bothers me. He had a few great playoff games, but he also was terrible in three superbowls. In a career perspective, ignoring the bad seasons makes sense, but 10 rushes for 13 yards, 11 rushes for 19 yards, and 16 rushes for 37 yards? These are in his prime, too.
Mattie,
In the prime of his career -- '89 to '96 -- he appeared in 17 playoff games. He rushed 316 times for 1318 yards, totaled 669 receiving yards and scored 19 TDs. Over a TD a game and 117 yards from scrimmage per game is pretty awesome in the playoffs. Sure, he had three bad Super Bowls -- he also had two 200+ yards from scrimmage games, two more 190+ YFS games and three more 150+ YFS games.
I suppose so. When i glanced at his numbers, I didn't notice one of his low rush yardage playoff games had 150 yards receiving. I guess he's just the paragon of choking in the superbowl in my head, so it seems odd. I do remember him having monster games in the playoffs before the superbowl.
I think he (with help from Kenneth Davis) ran for like 340 against Miami in a playoff game once. I imagine you have the stats for that. I think it was the year before Brian Cox flipped off everyone in Rich Stadium.
http://www.pro-football-reference.com/boxscores/199512300buf.htm
It was Holmes and Tindale, not Davis, but that's the game you're looking for.
I don't remember that game at all! Its a good example of why I have issues with QB playoff stats though... Miami was down big so they threw over 80% passes. It doesn't matter if you're Dan the man or Tim Couch, when you drop back to throw 83% of the time, you're going to get picked apart.
Hmm, that'd be an interesting way to weight QB stats... Find average pass/rush ratio, then weight games based on how close they are to average ratio. Then ideally you'd get how they perform in a close game, which might "fix" the stats for QB's hopeless teams.
Sorry for the tangent
Chase, thanks for the reply, but I'm afraid I don't buy it. Why does the RB get any credit (or blame) for what happens when he's not on the field? Couldn't you take the same logic and say I deserve some credit for the Steelers winning the Super Bowl because they didn't try to use me as a running back?
I can see how your logic works in making the GM's job easier. But I don't see how that should matter to, say, a Hall of Fame voter.
MattieShoes (#16.) You better check that Boxscore again. Marino and his Offense put up 0 (that's ZERO) points in that game in the 1st three quarters and then put up 22 points in the 4th. I'm guessing Dan's 3 INT's came BEFORE the 4th quarter.
I wasn't suggesting that the INTs were in the fourth quarter
I never saw the game so I really don't know, but they were down two scores in the first quarter and appear to have gone nuts on the pass for the entire game, not just at the end. I suppose it's kind of like the RB total rushing yards vs yards-per-rush debate. Dan threw the ball 64 times, so obviously the coach felt that was the way to win. But when you're losing and you're throwing over 80% of the time... Well, that's what happens.
In a more general sense, if you calculate a QB's average attempts, I imagine you'd find the ones with above average attempts have worse per-attempt numbers than those with below-average attempts. 64 attempts is wayyy above average, even for Dan, so I'd expect worse than average numbers.
For kicks, I stuck some numbers in a spreadsheet.
Dan has 18 playoff games. If you split them by attempts into a high-attempt and low attempt group...
The low attempt group has 2.4 TDs per INT. The high attempt group has 0.8 TDs per INT.
The low attempt group has 7.5 Y/A
The high attempt group has 5.9 Y/A
The low attempt group has 62% completions
The high attempt group has 52% completions
The low attempt group has 65 QB rating
The high attempt group has 95 QB rating
The low attempt group has 7-2 record
The high attempt group has 1-8 record
Before you claim this is some sort of further evidence of Dan choking in the playoffs, take your boy Troy. Take Troy's 16 playoff games and split them by attempts the same way.
The low-attempt group has a better TD/INT ratio, 9% higher completions, 15 point better qb rating, and nearly 2 more yards/attempt. And the low attempt group has a 7-1 record while the high attempt group is 4-4.
I'm not suggesting more attempts CAUSES bad performances, but game situation is huge for *any* player's statistics and significantly above/below average attempts is a reasonable indicator that the game situation was good or bad.
In the playoffs, the number of games is too low to balance out, especially since the bracket system forces more lopsided matchups intentionally. Actually I don't really think it balances out in the regular season either -- QB's on winning teams have a huge advantage, stats-wise.
For kicks again, I took Dan's numbers and called his low attempt playoff games an average winning performance and his high-attempt games an average losing performance.
Now Dan on a virtual 12-4 team would have 31 TDs an 18 INTs. Dan on a virtual 4-12 team would have 26 TDs and 25 INTs. QB rating would go from 72 to 87. Completion percentage goes up 5%. Y/A goes from 6.3 to 7.1. This was a long string of assumptions and has nothing to do with RB's in the playoffs so I'm going to stop now.
Mattie---Thanks for the info., that is interesting. And yet in the end as one study shows, if a QB throws 1 INT. in a PO game his Team wins 56% of the time. But when he throws 2 INT's, his Team only wins 31% of the time. That's quite a difference, I think you would agree. Here's the problem-Marino threw 2 INT's in MORE PO games than anyone else ( 10 times in 18 games ). Here is the BIGGER problem-Dan threw 2 INT's in 75% ( 3 of 4 ) of his Championship games. Sorry, I know this isn't the theme of the thread but I felt compelled to answer your post. Oh by the way, how did Dan and his Team do in those games from above-not so good.
Correlation does not imply causation.:-)
This is the same problem as the "teams who rush more win more" statistic. It's absolutely true, but rushing more doesn't necessarily CAUSE the win. Incidentally, Dan was 7-0 when his opponents rushed less than 37 times and 0-9 when they rushed more than 37 times. And the team with more rushing attempts won 16 of the 18 playoff games he was in.
Obviously the team that is ahead runs the ball more, so the team who ends up winning generally has more rush attempts. I don't think many would claim that rushing 40 times vs Dan automatically gets you the W.
I contend that the same situation holds with interceptions. The ahead team throws few passes, and usually into a run defense, and without pressure to make something happen, so they throw less interceptions. The behind team throws more, perhaps deeper, passes, usually into pass defenses, and they may feel compelled to try and make an impossible pass rather than throw it away. So they throw more interceptions, but less important ones. I find comparing raw INTs to final score highly suspect.
I was going to try and be back on topic by suggesting that RBs on winning teams might have better stats because they're less often caught in situations where rushing isn't an option, but then I looked and it doesn't appear to be true. There's a good mix of players from good teams, bad teams, and everything in between. Well, the playoff list itself is dominated by players on good teams, but the career list is not. Barry and OJ both played for some pretty bad teams for at least part of their careers. That's pretty awesome actually - it makes the list "feel right" in a way that the QB and WR lists don't. Well, to me anyway.