≡ Menu

Ed Feng of The Power Rank has contributed today’s post. As always, we thank our guest contributors for their hard work.


Before Super Bowl XXV, Bill Belichick had a plan. The Giants defensive coordinator wanted his defense to let Bills running back Thurman Thomas rush for 100 yards.

The Giants defense was appalled. They prided themselves on a tough, physical brand of defense.

Nevertheless, the short, stout coach looked up into the eyes of Lawrence Taylor and Pepper Johnson and said, “You guys have to believe me. If Thomas rushes for a hundred yards, we win this game.”

In case his defense didn’t listen, Belichick took it upon himself to make sure Thomas got his yards. He took out a lineman and linebacker for two extra defensive backs, playing a 2-3-6 defense.

Did Belichick go insane? I thought so when I first read this story in David Halberstam’s book Education of a Coach.

However, Belichick’s strategy has analytics on its side. Let me explain. [Editor’s note: See also this article by Chase about Super Bowl XXV.]

Passing versus rushing in the NFL

To break down Belichick’s strategy, we’ll look a team efficiency by yards per play. Teams that excel at running the ball will gain more yards per carry on offense than they allow on defense.

Hence, team rush efficiency is defined as yards per carry gained on offense minus yards per carry allowed on defense. Teams with good rush efficiency will have large, positive values. Teams that struggle in the run game will have negative values.

From 1997 through 2016, 57% of playoff teams (138 of 240) had a positive team rush efficiency. The bottom panel of the visual below shows playoff teams as a random scatter of positive and negative points.

Let’s compare this with passing. We’ll define team pass efficiency as the difference in yards per pass attempt on offense minus defense. These plays include sacks.

Over the same twenty year period, 84.6% of playoff teams (203 of 240) had a positive team pass efficiency, a much higher rate than for rushing.

The data suggests the insignificance of rushing in the NFL compared to passing. [Editor’s note: Or, at least, the insignificance of yards per carry to measure the value of running.]

Yards per play isn’t the perfect statistic to measure rush efficiency. Teams that excel in pass efficiency will tend to have leads late in games. In these late game situations, they will run the ball to take time off the clock. The defense expects the run, which should lower their rush efficiency.

From 2012 through 2016, teams up 7 or more points in the 4th quarter rushed for 3.29 yards per carry, below the NFL average of 4.18 (data from Pro Football Reference). While this hurts the rush efficiency of good teams, these carries only make up 12% of all carries during this period.

Excluding these late game carries would not change the main claim about the insignificance of rushing compared to passing in the NFL.

Let’s look at passing and rushing efficiency for playoff teams in 2017.

The 2017 season

Of the 12 playoff teams, 7 had a positive team rush efficiency during the regular season. This 58% rate matches the historical trend over the previous 20 years.

Of these 12 playoff teams, New England (26th) and Pittsburgh (29th) had the two worst team rush efficiencies. Many considered these two teams the primary Super Bowl contenders when the playoffs started.

In addition, Cleveland had the best team rush efficiency by a wide margin over Philadelphia in 2017. The Browns didn’t win a game.

How about passing? 11 of 12 playoff teams had a positive team pass efficiency for a rate of 91.7%. Only Buffalo, a team that gutted its roster over the last year and hadn’t made the playoffs since the 1999 season, allowed more yards per pass attempt than they gained.

Let’s dig further into this season’s rushing numbers. These rankings show the 12 playoff teams in terms of yards gained per carry on offense (data includes playoff games).

1. Kansas City, 4.68
3. New Orleans, 4.47
6. Los Angeles Rams, 4.40
7. Philadelphia, 4.34
8. Tennessee, 4.29
9. Carolina, 4.28
10. Jacksonville, 4.27
12. Atlanta, 4.21
14. Buffalo, 4.14
15. New England, 4.12
23. Minnesota, 3.87
25. Pittsburgh, 3.85

Despite the difficulty in running the ball while up late in the game, 10 of 12 playoff teams were better than NFL average.

However, playoff teams really struggled to defend the run this season. In terms of yards per carry, here are the ranks of the 12 playoff teams.

4. Tennessee, 3.60
5. Minnesota, 3.65
7. Philadelphia, 3.79
9. Carolina, 3.87
21. Atlanta, 4.17
23. Jacksonville, 4.26
25. Buffalo, 4.30
26. New Orleans, 4.31
27. Kansas City, 4.41
28. Pittsburgh, 4.42
30. Los Angeles Rams, 4.56
31. New England, 4.57

8 of 12 playoff teams were 21st or worse in rush defense by yards per carry. New England, who will face Philadelphia in the Super Bowl, is the worst of these playoff teams.

How Belichick will defend Philadelphia

In Super Bowl XXV, Thurman Thomas got his yards. The Bills running back rushed for 135 yards on 15 carries for an amazing 9 yards per carry.

However, the Giants won the game 20-19 when a last minute field goal attempt sailed wide. Belichick opted to let Thomas get his yards instead of allowing Jim Kelly to throw the ball. Despite a 31 yard touchdown run from Thomas, he knew focusing on pass defense gave the Giants the best chance of winning.

Belichick has used a 2-3-6 type defense more recently. Against Peyton Manning in 2013, he put extra defensive backs on the field. Manning checked down to run plays, and Denver ran for 5.83 yards per carry compared to 3.47 yards per pass attempt. New England made a furious comeback to win 34-31 in overtime.

However, I doubt we’ll see this in the Super Bowl. While Nick Foles has played well in the playoffs, he doesn’t demand the same respect as Peyton Manning.

Ed Feng is the founder of The Power Rank. To get a free booklet with his full manifesto on passing and rushing in the NFL, click here.

{ 35 comments }
  • sacramento gold miners

    I don’t know if Bill Belichick considered analytics back in 1991, but his strategy was based on understanding the mentality of the opponent. The Giants had slowed the Buffalo attack in a 17-13 loss during the regular season, using their conventional defense. Belichick knew the Bills were red hot in the postseason, and playing that same defense in the rematch would have been problematic. Buffalo was accustomed to usually doing whatever they wanted offensively, Jim Kelly was a gunslinger. Belichick dared the Bills to run, gambling on the stubbornness of the Bills to work in his favor.

    Thurman Thomas wasn’t a factor rushing the ball until the fourth quarter of that game, when he gained more than half of those 135 yards. Buffalo just didn’t adjust in time, and the physicality of the Giant’s offense wore down their defense for most of the game. For three quarters, New York shortened the game using ball control, and by not adjusting earlier, the Bills became less patient as the game progressed.

    • You are right that the Giants’ conventional defense slowed the Bills in Week 15, but there is also more to that than meets the eye. Kelly got hurt midway through the second quarter and Frank Reich played the rest of the game. So in a little over a quarter and a half in horrible weather conditions, Kelly was 7/11 for 115 and a TD and led two touchdown drives. I think Belichick figured that, given the Bills playoff success and the success they had against the Giants with Kelly in December, they’d try to attack through the air again. As you mentioned, he gambled on the Bills stubbornness (and Kelly’s ego, since Kelly called the plays) that they would keep trying to throw and recreate their success of Week 15 and the prior playoff games, even if it didn’t work at first. As you said, Buffalo didn’t adjust until their margin for error was gone. And even then, they almost won. Belichick’s bold gamble worked…just barely.

  • Tom

    Great post, I especially like the insight and rush yards per carry in the context of game situations, etc.

    Regarding the Super Bowl 25:

    Belichick’s game plan was great, but I’m pretty certain it wouldn’t get the attention it has – which is deserved, because it WAS a great game plan – if Norton hits that field goal.

    Thomas ran for 33 yards on the Bills last drive…on 2nd-4 with 29 seconds left on the Giants 40, he ran for 11 yards to put them on the 29 with 9 seconds left. That single run put them in field goal position, poised to win the game.

    So yes, great game plan; we can guess that if the Bills’ passing game was more successful they would have put up more points and would not have been in a position where they needed a field goal to win at the end.

    But there was a single moment where a Thomas run almost won the game for the Bills. And we might be talking about how interesting it was that the Bills didn’t win the game with passing, they won the game by running the ball…and how smart it was for them to go to Thomas at the end to pick up the needed yards.

    • sacramento gold miners

      Super Bowls 25 is also a relic of the game’s past, we almost never see this formula anymore. I’m talking about bottling up a great offense with ball control, and physicality on defense without forcing multiple turnovers. SB 42 had the defensive effort without forcing turnovers, but the time of possession and running game weren’t significant factors in the game.

      If Super Bowl 25 was played under today’s rules, I’m certain the Giants would have been flagged multiple times for hits, and those 15 yard gifts would have led to more Buffalo scores. And the tighter restrictions on receiver contact would have been another hurdle for the Giants to overcome.

  • Johhny Ohrl

    I preached passing year in year out… Throw on every down. Just look at the TMW plays. If you pass, the D can´t defend even though they know what´s coming. Nice side-effect: The draw works perfectly. 6 yards per play a given… One day, when stubborn the Norv-The-Smurf-Turners & Co are no longer between us, coaches will call agressive plays right from the start (as they should. Like AFL coaches do). Beli was one of the first to understand that simple concept when he let Brady throw 50+ times against SEA in the SB… To run the ball is useless, unless you lead by 20 minimum deep in the 3rd to run down the clock…

    • What’s funny is that strategy for Belichick in Super Bowl XLIX is “great” because of some late game luck, just like Super Bowl XXV. Had Russell Wilson not thrown that extremely rare 1 yard line interception and the Seahawks score, then just like Super Bowl XXV, the game is exactly the same but the narrative is “New England wasn’t big or strong enough to stop the Seahawks in the trenches. Seattle outrushed the Patriots by a ratio of over 3 to 1 and forced New England to pass every play. You need balance to win the Super Bowl”

      • Johhny Ohrl

        “You need balance to win the Super Bowl”…
        No you don´t. As another example (if the above wasn´t to your liking) just have a look at last years SB. Or look at the “miracle” in MIN this year, or any other “miracle” (they come in dozens per season nowadays).

        Once teams stop shooting for 2 yard gains by running straight into a wall of 300 pound bodies, the scoreboard explodes (and that is the aim in a game: score as much as possible to extend your chances to win): It mostly happens in TMW warnings, and 4th quarters. Even bad teams (those without proper respectively continious OL play, and thus without solid pass protection) come up with a comeback now and then. Did any “miracle” ever happen by running the ball? Rhetorical.

        Sometimes run on 3rd/4th and 1, at the GL, that´s about it. Those plays I guess make up for max 5% of all plays*.

        (* not including the above mentioned milking the clock runs)

        • I wasn’t saying you need balance. I was saying that’s what the talking heads and pundits would have said if Seattle scored at the end and they won that game. Most strategy discussions start at winning and work backwards. There are many ways to win, but conventional wisdom (incorrectly) always praises the strategy of who won. Having Brady throw 50 times is treated as smart because they won. Seattle, adopting almost the complete opposite, ground and pound strategy, came within one yard of winning. Ironically, one more power run probably would’ve done it for them. I was just making the point that, much like Belichick’s famed Super Bowl XXV gameplan, a strategy only looks smart if it wins. In these cases, it’s often forgotten that these strategies came within a razor’s edge of not winning.

          • Johhny Ohrl

            Ah ok… Fully agree with what you say… Thx for your explanation.

          • Four Touchdowns

            I wonder if the power run would have actually worked. The Patriots had a lot of guys on the line of scrimmage — if it didn’t work, the Seahawks would have burned a lot of clock with no timeouts. https://uploads.disquscdn.com/images/b09c1627cd1945df6471918ee8ad881e233ae0cbb5be014e2ba78ec284628cef.jpg

            • Richie

              According to that image, the Seahawks still had a timeout remaining.

            • Tom

              It looks like they have one timeout? In any event, it’s been discussed a bunch of times, but Caroll’s thought process was (if I recall), if they miss on the pass, then they can take a shot on the run, or something like that. He’s obviously not considering that the worst of the “three things happen when you throw the ball” was going to happen.

              • Johhny Ohrl

                “the worst of the “three things happen when you throw the ball” was going to happen.”

                Thats how losers think. Good coach Caroll. Especially seeing the D formation in the pic above. Right call. If coaches would always think of the 2% Int-chance, they´d better become a Turner-clone. We all know how that worked out for the smurf: Peter principle at its finest.

            • eag97a
          • Tom

            Right on…could not agree more. Some of us were having same discussion on FP on another topic…you can’t start with the win of a single game and then decide that a strategy worked, or that a QB is great. Belichick’s strategy was a good one, regardless of whether or not the Giants won the game.

            • sacramento gold miners

              Fascinating 30 for 30 ESPN show tonight, about the Parcells/Belichick relationship. Long before Belichick began implementing successful strategy, he got his first pro coaching job with the 1975 Colts, making $25 per week. It all starts with dedication, how many 23 year old college graduates would have accepted that kind of paltry salary?

        • Tom

          “Once teams stop shooting for 2 yard gains by running straight into a wall of 300 pound bodies”

          Best argument against running the ball I’ve ever heard…hahahaha

          • Johhny Ohrl

            Its always a pleasure that you enjoy my (sometimes controversial) posts 😉

      • Tom

        That is 100% correct. Too often, we let the outcome – which we know, at times, can be largely determined by random events – determine how we feel about the strategies. I think it was Brian Billick or some other coach said, about a play or 4th down decision, can’t remember: “If it works, it was a great idea, if it doesn’t work, it was a horrible idea”, something like that. If Lockette makes that catch, we’re all sitting around talking about what a mastermind Caroll is, because he passed when everyone expected run.

        • Richie

          One of the things I’m having trouble reconciling is that the Falcons lost the Super Bowl because they stayed too aggressive. But the Jaguars lost in the AFC Championship Game because they went too conservative.

          I guess it turns out, that either way you need to have some success, whether you are being conservative or aggressive, and neither team had much in the fourth quarter.

          • Tom

            Yep. I was one of those guys saying “Why did you guys keep passing? Why didn’t you run the ball more and bleed the clock?”, etc. And guess what I would have been saying had they just run the ball to kill the clock and still lost “Why did you guys stop passing? Why did you get conservative?”

            We know certain strategies are good and others aren’t, but when we look at a single game, it can be really hard to tell because we’re basing our judgement on the outcome. You’re totally right – “either way, you need to have some success”.

            Heck, I was screaming at the TV when the Jags went in at halftime with 55 seconds left and two timeouts. But then when I read Marrone’s explanation – basically saying “we’re not really good at that” – I thought, well maybe he knows his team and felt that they should just get in to half time with a lead and not risk a disaster happening (I still disagree though!).

            • Johhny Ohrl

              To go on with passing was right. Especially when the opponent is stacking the box to stop the run*… The reason the comeback worked was either for entertainment purpose only, Vegas had a risk to lose big, or ATL thought they had this in the bag and didnt gave 100% anymore (unlikely)…

              * I love to watch SB 20 now and then as prime example. The “brilliant” coaches of NE thought the Bears on offense were still Payton only (we all know how that worked out: zero championships the time the Bears had no QB. He could have run for 5.000 yds a season and still come up empty), thus stacking the box. Hence McMahon had a field day. Going long on most of his 20 passes (as usual he left the game injured, otherwise the final score was 61-10 or so) for an amazing 12.8 Y/P. If Gault wasnt shirt tackled on the bomb from the 4 yard line, that number would have sky rocketed…

            • eag97a

              And to add on to the discussion. For people who argue that the Seahawks should have run the ball during the end of SB 49, BB agrees with the pass play, here is the link;

              https://mobile.nytimes.com/2018/01/30/sports/football/patriots-belichick-zolak.html

              • Passing may might not have been the right call, but it wasn’t the “wrong” call like it’s remembered. Check out these stats:

                Since 2010, there have been 819 passes thrown from the 1 yard line
                -417 resulted in Touchdowns (50.9%)
                -27 resulted in a Sack (3.3%)
                -13 resulted in an Interception (1.3%)
                -4 resulted in a Fumble (0.5%)

                Passing wasn’t the problem. Russell Wilson made a poor decision/throw, Lockette ran a poor route, Bevell put in a poor formation/play or combos of all. But throwing in of itself wasn’t a problem. Throwing from the 1 yard line leads to a touchdown more than half the time and leads to an incompletion almost every other time.

                • eag97a

                  Well BB explained to Zo that the Pats stuffed the Hawks using the same formation earlier in the game and that factored in Carroll electing to pass in that situation. Browner jammed the guy running a rub route and Butler made the play of his life driving and intercepting the ball. I actually trust PC and the Hawks in deciding that play call it just didn’t work out for them. Same situation with the 4th and 2 call by BB that failed as well against the Colts which led to a loss. Those play-calls were actually well reasoned and just didn’t succeed. Rolling the dice will not always be successful.

                  • sacramento gold miners

                    BB is friendly with Pete Carroll, so I wouldn’t expect him to pile on about the bad play call. The Seahawks, like many other teams in today’s pass happy NFL, overthought the situation. As Eric Drews pointed out in his numbers, only 51% of passes thrown from the one yard line resulted in TDs since 2010. Your chances of running the ball with Lynch or Wilson would have been greater in getting the game winning TD. Even if we assume NE denies the TD on second down, Seattle would have called TO with about 18 seconds left, and a pass/run option with Russell makes more sense on third down. The Seahawk’s defense also let them down in the fourth quarter of that game.

                    • eag97a

                      BB might be friendly with Pete Carroll but the way the article framed it was BB heard about Zolack ranting on-air about the playcall at the end and asked ZO to review some film with him to point out that the Pats went goal-lime defense and stuffed a Lynch run earlier in the game. Now BB might be covering for PC and there was a BB interview where he mentioned that PC was being judged by people very much less qualified than him but in this case I believe BB was telling an honest opinion and no offense to anybody I’m inclined to believe a HOF hc over monday morning qbs/armchair HCs/spreadsheet fantasy stats warriors.

                      Numbers should be judged abut the context surrounding them. In that particular situation PC judged the risk of interception was low since it was a fast developing slant and a rub route was gonna help execute it while the Pats still have to guard runs while guarding against passes. The Pats ran 3 corners against 3 receivers man to man and they just executed very well stuffing the rub/pick play and of course Butler driving very hard and making the pick. If the receiver faked the slant and Wilson threw a fade and TD if would have been a different story. PC and the Hawks made a reasonable decision and lost. Roughly same thing happened to BB when he was skewered for the 4th and 2 call against Indy and lost. Tbf the stat warriors defended BB then because it was an aggressive 4th down call but it was a regular season game and lesser stakes where in play.

      • sacramento gold miners

        I’m unsure what you mean by late game luck in SB 25, Scott Norwood’s 47 yard FG attempt was a difficult kick by someone who was just 6-10 from that range in the regular season. Also keep in mind Norwood was accustomed to kicking on artificial turf, he had just three FG attempts on grass the entire season. Norwood converted on a chip shot 23 yard FG earlier in the game, but a 47 yarder was far more a gimme.

        Even if Norwood hits that FG, Belichick still had a great strategy. Holding that offense to just 12 points over three quarters was tremendous.

        • Just because it wasn’t a gimmie, doesn’t mean the Giants weren’t fortunate that it didn’t go in. 6-10 still means that the Bills had a better than 50% chance of making the kick. Norwood struggled from that range, but he didn’t shank it or miss it short. That kick had more than enough leg but it was two feet to the wrong side. And I agree about strategy and that was my original point. Belichick’s strategy held the highest scoring team in the league well below their season average. That would’ve been true even if Norwood would’ve made the kick. However, in that alternative scenario, I’m sure few would have given him credit for it.

          • Tom

            Exactly. We wouldn’t even be discussing this right now if the Giants had lost. In fact, people might say it was the *wrong* strategy because it was a run play – Thomas’ 11-yarder – that put the Bills in field goal range.

            We’re all in agreement that what made it a great strategy was “holding that offense to just 12 points over three quarters” as sacramento noted, NOT the fact that the Giants actually won the game.

            Since we’re talking about this, the numbers below clearly show the rushing/passing imbalance. I’d almost say Belichick’s strategy “worked” too well!

            WPA numbers for the Giants:
            OFF: +0.35
            DEF: -0.37 (-0.08 passing, -0.34 rushing)
            SpT: +0.70

            EPA numbers for the Giants:
            OFF: +5.4
            DEF: -8.8 (-1.8 passing, -8.4 rushing)
            SpT: +4.4

    • Richie

      What is a TMW play?

      • Johhny Ohrl

        two minute warning play…

    • Tom

      Was watching an old game YouTube and John Madden pretty much said the same thing, while watching a team go through a 2-minute drill at the end of a half, “I don’t know why they don’t play the whole game that way”.

      • Johhny Ohrl

        Lucky you watching old NFL games (I guess full length and not just those annoying 2 minute clips)… There was a guy how uploaded every single Bears-85 game. Then it got deleted by the greedy §uckers called NFL. If they dont like someone using their footage, they should upload the games themselves. I like to watch real football of the 80s…

        Btw, I didn´t know Madden said this. And I had to listen/endure 😉 a lot of him when I started watching the NFL. Anyway, always liked Madden and his “painting” of the screen. Some kind of an art 😉

  • ed

    Okie dokie, so Bellichek schemed the Bills into missing a field goal late in the game? This entire article is a load of shit. It just boils down to ‘defend the pass against passing teams, don’t get distracted by an unproductive running game’.