The Rams, the Bengals, and Guts & Stomps Revisited

Cincinnati Bengals WR Tyler Boyd
Cincinnati Bengals WR Tyler Boyd
Photo: USA Today Sports Images

NFL Preseason Week 1 - Guest column by Chris Johnson

In the regular season last year, the Cincinnati Bengals had three wins by 14 or more points against teams that finished with winning records; the L.A. Rams had none. However, the Rams didn't lose a single game to an opponent who finished below .500, while the Bengals lost to losing teams four times. Did any of these past results forecast L.A.'s win over Cincinnati in the Super Bowl?

Every time a good team beats a bad team by a large margin, there's a lot of debate about the degree to which we should care about it. According to one school of thought, what a contender does against a bad team means little or nothing for their long-term prospects, because in order to win in the playoffs you have to beat good teams. Winning a close game against a good team is a better indicator of postseason potential because hard-fought games against good teams are what the playoffs are all about. According to another school of thought, however, a big win against a bad team is an equal or better indicator of overall team quality compared to a close win against a good team, meaning that we should take such a blowout seriously when considering a team's long-term prospects.

Early in the history of Football Outsiders, Aaron Schatz wrote an article about this topic called "Guts and Stomps" in which he argued in favor of the second position. Yes, he argued, championship teams are often defined by their ability to destroy bad teams, not their ability to win close games. That article broke down the numbers from 1995 to 2004 to support Schatz's point. It hasn't been updated in almost 20 years. So today, we're going to compile those same numbers from 2005 to 2021 and see whether anything has changed.

Definitions

Wins in the original "Guts and Stomps" article were put in the following categories:

  • A "Gut" is a one-score win (eight points or fewer) against a team that eventually finishes over .500.
  • A "Stomp" is a win by 14 points or more against a team that will eventually finish under .500.

Those were the two main categories, but Schatz went further (as will I) to look at two other categories of wins:

  • A "Skate" is a one-score win against a team that will finish under .500.
  • A "Domination" is a win by 14-plus points against a team that finishes over .500.

For this updated article, I have decided to at least define the four analogous categories of losses:

  • A "Gutless" loss is a one-score loss against an eventual over-.500 team.
  • An "Outclassed" loss is a loss by 14-plus points against an eventual over-.500 team.
  • A "Stinger" loss is a one-score loss against an eventual under-.500 team.
  • An "Embarrassment" is a loss by 14-plus points against an eventual under-.500 team.

If losses follow a similar logic to valuing Guts over Stomps (quality of competition is more important than margin of victory), you might expect Stinger losses to be worse for a team than Outclassed losses as well. We'll see. For the record, here's how the Rams and Bengals fared in the regular season last year:

2021 Games by Category
Category LAR CIN
Gut 2 1
Stomp 4 3
Skate 2 3
Domination 0 3
Gutless 2 2
Outclassed 2 1
Stinger 0 3
Embarrassment 0 1

The Rams then added a Domination and three Guts in the playoffs, while the Bengals added three Guts themselves before suffering a Gutless loss in the Super Bowl.

Super Bowl Winner Deep Dives

For Super Bowl winners since the 1995 season, I went through all eight of the categories we just discussed. Here are the numbers for all of those Super Bowl winners in the regular season:

  • Overall record: 324-109
  • 42 Guts, as opposed to 42 Gutless losses.
  • 58 Skates, as opposed to 17 Stinger losses.
  • 45 Dominations, as opposed to 18 Outclassed losses.
  • 102 Stomps, as opposed to 4 Embarrassing losses.
    • I didn't keep track of nine- to 13-point wins, but I can tell you that Super Bowl winners in this timeframe have only 25 total losses against sub-.500 teams, so the 14-or-more-point Stomps alone quadruple the number of losses.
    • Trivia question: How many of those four Embarrassments can you name?

And here are notes more similar to the ones in Schatz's original article:

  • 19 champions have more Stomps than Guts; only four have had fewer (2003 Patriots, 2006 Colts, 2011 Giants, 2015 Broncos).
  • Nine champs have had more Guts than Gutless losses; 12 have had fewer.
  • Only the 2004 Patriots, 2012 Ravens, and 2014 Patriots have more Dominations than Stomps. Eighteen have had fewer.
    • Schatz's original article included the 1995 Cowboys by mistake in this category. They had three of each.
  • Five teams have more Skates than Stomps, as opposed to 19 who have had fewer.

As it turns out, it's possible that Schatz's original article actually undersells his case, due to a failure to look at the losing end of the spectrum. Not only are Stomps much more than twice as common as Guts for champions, but Super Bowl champions since 1995 are exactly .500 in one-score games against teams who finish over .500.

I was somewhat on the Stomps > Guts bandwagon before I wrote this essay, but the Guts vs. Gutless stat surprised even me. To be fair, even an 84-game sample has some margin for error (so there's still room to speculate that Super Bowl champs should have some advantage in these games, up to winning 60% of the time or so), but the ratio (only one Gut for every Gutless loss) is remarkably unremarkable, to the point that it made me doubt for a second whether my data was even correct.

On the other hand, the Skates-to-Stingers ratio (3.4:1) and the Dominations-to-Outclassed ratio (2.5:1) are much more similar. The confidence intervals for those two samples overlap, so there isn't much we can conclude from that discrepancy.

Of course, the ratio of Stomps to Embarrassments (25.5:1) totally laps the field, although that's not necessarily news. (Super Bowl winners don't lose badly to bad teams very often!)

Here are particular cases of teams who bust the Guts > Stomps narrative:

  • The 1999 Rams had a ludicrously easy schedule. They racked up 10 Stomps (plus two Stomp-level wins against 8-8 teams), and played a grand total of ONE regular-season game against an over .500 team (a Gutless loss to the Titans).
     
  • All three of the 2013 Seahawks' losses were Gutless. Of course, those are the best losses to have from an "indication of team quality" standpoint. But if you went with the "who can find a way to pull out close wins against good teams" narrative, you could play games surrounding the fact that Seattle went 9-0 against below-average teams and "only" 4-3 against above-average ones (and 1-3 when those contests were decided by one score).
     
  • The 2005 Steelers (four Stomps, one Gut, three Gutlesses) and 2010 Packers (four Stomps, two Guts, three Gutlesses) are similar examples.

Playoff Breakdowns, 2001-2021

Another element of the original Guts and Stomps article was a table looking at whether the team with an advantage in a particular category won each year in the Super Bowl and conference championship games. I did this exercise with every playoff game in the last 21 seasons, for all four categories of wins, as well as DVOA, overall wins, home-field advantage, and the number of wins that don't belong in any of the four "buckets" (either nine- to 13-point wins, or wins against exactly .500 teams).

Wins by Category and Playoff Advantages
  DVOA Wins HFA Stomps Guts Skates Dominations Other Ws
2011-2021 73-52 65-42 75-41 57-42 46-50 40-58 54-33 51-55
2001-2010 67-43 52-42 60-40 50-39 42-41 33-54 46-30 46-41
2001-2021 140-95 117-84 135-81 107-81 88-91 73-112 100-63 97-96
Only SBs 11-10 7-11 2-0 5-11 8-11 9-11 10-5 9-8
Only SBs and CCGs 35-28 28-24 30-14 23-27 24-25 24-31 26-17 24-30
2001-2021, Pct 59.6% 58.2% 62.5% 56.9% 49.2% 39.5% 61.3% 50.3%
90% conf. interval 54.0%,
64.9%
52.2%,
64.1%
56.8%,
68.0%
50.7%,
63.0%
42.8%,
55.6%
33.4%,
45.7%
54.6%,
67.7%
44.1%,
56.4%

To help understand this table: under the HFA column, the 135-81 in the 2001-2021 row means that from the 2001 season through the 2021 season, playoff teams were 135-81 when they had home-field advantage. Teams with more Stomps than their opponent in the regular season went 107-81, teams with more Skates went 73-112, and so on.

In the later rounds, Stomps haven't separated themselves from Guts in the last 21 years. However, in a larger sample of the whole playoffs, they definitely have separated themselves. Meanwhile, teams with the most Dominations are clearly winning at high rates regardless of how you slice it.

What Schatz seemed to be missing in his original article is that although Dominations are more rare than Stomps for Super Bowl winners, they're also more rare in general. When you can pull a Domination off, it actually is impressive and suggests you're in good shape going forward. Dominations are, in fact, the only win type which competes with the three "baseline" indicators (DVOA, overall wins, HFA) in terms of predicting playoff success (at least in a positive sense; you could say Skates are a reasonably good negative indicator).

Looking at overall percentages and confidence intervals, a clear hierarchy of Dominations > Stomps > Guts > Skates emerges. It's mostly in line with intuition/conventional wisdom, with the exception that Stomps > Guts isn't quite as intuitive.

Fun Factoids

  • Embarrassment losses by Super Bowl champions in history (not just the era I was focusing on):
    • 1981 Week 3: ATL 34, SF 17
    • 1988 Week 3: ATL 34, SF 17
    • 1993 Week 1: WAS 35, DAL 16
    • 1994 Week 5: PHI 40, SF 8
    • The next four games are the answers to the earlier trivia question!
    • 2003 Week 1: BUF 31, NE 0
    • 2011 Week 1: WAS 28, NYG 14
    • 2016 Week 4: BUF 16, NE 0
    • 2018 Week 3: DET 26, NE 10
       
  • The Brady/Belichick Patriots (2001-2019, excluding 2008) had 64 Stomps, 33 Guts, 38 Skates, 34 Dominations, 27 Gutlesses, 12 Outclasseds, 10 Stingers, and five Embarrassments.
    • The five Embarrassments were the three games listed above, plus a 21-0 loss to Miami in 2006 and a 31-14 loss to Cleveland in 2010.
       
  • The 1972 Dolphins are infamous for their easy schedule, so I was curious what they look like with these definitions. They had eight Stomps and one Skate, then added three Guts in the playoffs.
     
  • The 2007 Patriots, on the other hand, had six Stomps, two Guts, one Skate, five Dominations, and of course added one supremely Gutless loss in the playoffs.
     
  • The 1999 Rams lap the field among teams I looked at for this exercise with their 10 Stomps. I'm tempted to say "unbreakable record," but I will admit that I haven't looked at every team in history to be sure that it is a record. And with weak divisions, it might be possible to see another team racking up double-digit Stomps one day.
     
  • Among teams I analyzed before 2021, the leaders in Guts are the 2010 Falcons and the 2017 Panthers, tied at five. Both were one-and-done in the playoffs; the Falcons got Outclassed by the Packers and the Panthers ironically lost in Gutless fashion to New Orleans. In 2021, the Buccaneers joined the Five Guts club, and of course lost in the divisional round of the playoffs.
     
  • I thought the 1972 Dolphins would compete for the Skates title, but no, the title among teams I looked at belongs to their successors in Miami from 2016, who racked up eight Skates throughout the season. Another one-and-done.
     
  • The Domination title is tied between the 2007 and 2014 Patriots. They were tied at five (although it's worth noting that the 2014 team added a sixth in the AFC title game).
    • The 2014 team also has the distinction of having Dominated all three of the other AFC division winners that year, leaving little doubt as to their supremacy in the conference.
    • However, the 2007 team did Dominate three division winners, just not all in the AFC (they Dominated Pittsburgh, San Diego, and Dallas from the NFC).
       
  • Seven teams I looked at had four Dominations or more: the 2021 Bills, 2019 Ravens, 2002 Buccaneers, and the Patriots in 2004, 2007, 2014, and 2017. Five of those teams made it to the Super Bowl and three of them won it.

Conclusion

Beating up on bad teams, far from useless, is actually one of the main characteristics of winning postseason teams. Close losses to good teams are important because they're losses in the standings, and because of their effect on tiebreakers for playoff position, but the identity of the best teams is generally not determined by which teams play best in close games against other winning teams.

Christopher Johnson is a recent college graduate from Massachusetts, so he cheered the Tom Brady Patriots on as they Stomped and Dominated through the league for the last two decades. You can email him at bluecjj@gmail.com.

Comments

26 comments, Last at 12 Aug 2022, 3:36pm

1 Thanks to Mr. Johnson for…

Thanks to Mr. Johnson for the article, and thanks to Mr. Schatz for running it! I would love to see additional articles revisiting some of the research from early-era FO to see if their conclusions continue to hold true. I suspect they generally do, but the game has changed in the years since -- and we've got many more seasons of data -- so I think it's a worthwhile endeavor.

2 Thanks for updating the…

Thanks for updating the information.

 

Though it's a bit disappointing to me, that with all of these categories, the biggest predictor of winning postseason games is still HFA. 

I guess the next question would be how guts and stomps correlate with teams *earning* HFA.

4 Great article!  I expect to…

Great article!  I expect to link to it regularly this year after the frequent "beating bad teams doesn't matter, who have they beaten?" comments.

5 Championship teams are…

Championship teams are defined by winning a title. They may be characterized by stomping bad teams. 

14 LOL.  I'm so glad we're…

LOL.  I'm so glad we're clear that championship teams are defined by winning a championship.  So all I have to do to best predict who the most likely champion is of all the contenders...is wait until after the game.  So helpful!

23 You missed the point

The sample of SB winners will only include...SB winners.

A property or attribute of those teams is how often they stomp cupcakes, which will have some sort of distribution. The extent to which that distribution can be used to 'predict' who will, or even just who is more likely, to win a championship is then just naive empiricism, made even worse by excluding all other teams who aren't SB winners. In fact the game is given up in this line: "In the later rounds [of the playoffs], Stomps haven't separated themselves from Guts in the last 21 years". 

6 Interesting observation

about the "embarassments":

 

All 8 occurred in the first 5 weeks of the season

6 of 8 occurred in the first 3 weeks of the season

3 occurred in the first week of the season

 

The lesson - as always - September football is mostly teams using those weeks to figure who their best players are and what their identity is as a team.  Pre-season used to be used for this but now that there's very few padded practices and roster regulars rarely play in PS games the first weeks of the season now serve as "pre-season".  

Which is why RESULTS are more important than FORM in September. After all, September win count the same as December wins at the end of the season.  But what we see on the field in first quarter of the season is rarely an indication of a team's eventual performance / outcome. 

12 Results > Form early

Interesting if this could apply to weighted DVOA, whether ignoring the first 'N' weeks would have more predictive power for the playoffs than full season DVOA

17 IIRC, Aaron or other staff…

IIRC, Aaron or other staff has revisited this idea on a fairly regular basis, and full-season DVOA has consistently come out as more predictive than Weighted DVOA... Which would seem to imply that early-season form is, in fact, relevant.

19 Yeah

Or at least no significant difference. Would love to hear the updated correlations though to confirm. 

Sure things are rougher earlier but every team is trying to find themselves so it's kinda moot point. Essentially punish teams for doing better early and having their ducks in order. 

20 I'm not sure it was that…

I'm not sure it was that simple: I think it was full-season DVOA is more predictive for next year but weighted DVOA is more predictive in-season. Basically how you do over the full year is more indicative of how strong your roster is (and how competent you are as a franchise) but obviously in-season drops (often due to injuries) are relevant for continued performance in that season.

Even as recent as last season, Aaron commented: "The big story, however, may be in our weighted DVOA ratings that have a slightly higher correlation with future performance."

Note that there's an older article from Bill Barnwell (vanished from the interwebs) that concluded that total DVOA was more predictive for playoff success, but that's a little different since he was doing a multivariate analysis with splits (so the data set was already sliced and diced).

9 Interesting Results

But regardless of how many cupcakes the Cowboys stomp in their division, I won't take them seriously unless they win some of their early games against the Bucs, Bengals, and Rams.

If Michael Gallup returns by the time they face the Packers (and win), then maybe I'd be willing to write off those losses.

15 "Interesting facts, but Ima…

"Interesting facts, but Ima go with my gut which runs counters to the facts."  Our guts and the narratives we build around certain regular season games...lol  Tanier wrote a hilarious article about how the Bills had no heart, shouldn't be taken seriously etc. etc. after their first loss to Pats last season (the no-pass game).  Then they beat down the Pats in 2nd game and playoffs.

18 It's probably better to say…

It's probably better to say it "I won't take them seriously unless they're competitive versus etc. etc. etc."

Tanier wrote a hilarious article about how the Bills had no heart, shouldn't be taken seriously etc. etc. after their first loss to Pats last season

So to be clear, I don't disagree with you that the article was silly (but it is Walkthrough, it's a columnist's opinion, he's allowed to have the "just don't buy them" gut feeling). But...

The point of that article wasn't that the Patriots were better than the Bills. It was, to quote the article, that the Bills lack "an "it factor," some combination of intangibles that converts into an extra dozen yards or so in a close game[.]" In fact, it was a hook for the prediction that the Bills would lose to the Buccaneers that next Sunday.

Which... they did. In overtime. In kinda gutless fashion. I mean, 1st and 10 in the red zone (with a first down available!), plenty of time, down 3... field goal. I mean... meh. But then they win the coin toss! There's your "it" factor, manipulating a piece of metal with the sheer force of Josh Allen's will! And... not even a first down. But! 63 yard punt (holy crap) pinning them inside the 10 (where the Bills are now more likely to score)! There's your it factor, right? Matt Haack's ROBO-PUNTER leg. And... short pass to Perriman turns into a 58-yard touchdown with the Bills in a high-comedy short-yardage defense that just screamed "pass to the right side."

Sooo.... maybe not the best example, because that specific article was pretty prescient, actually.

21 It's probably better to say…

It's probably better to say it "I won't take them seriously unless they're competitive versus etc. etc. etc."

I admit it is probably better, but also hard for me to do. If Michael Gallup is out for all those games, and Ceedee is limited by double coverage, it'll be easy for a lot of people to say, "Well they had no deep threat, no wonder they lost some close games." There always seems to be a convenient excuse for this team underperforming, and that's likely to be the one trotted out.

For me, it really depends on the kind of competitiveness they show in the game. The Boys have a habit of getting behind early in games against playoff participants (Cardinals, 49ers, Browns in 2020, etc), then making a rally at the end to "make it competitive", even though winning those kinds of games require a lot going right, and unsurprisingly, they don't win them. If they lose all their games to the teams I mentioned the way they lost to the Bucs last year, where they keep pace the whole game, it'll be a lot easier to trust the Stomps > Guts numbers.

If they lose them by failing to rally in the 4th quarter, I'll probably be citing a curse by Jimmy Johnson or the extenuating circumstances of having an entitled country club mentality of softness that Jerry Jones instilled in the ranks of the team.

22 So...

"The point of that article wasn't that the Patriots were better than the Bills. It was, to quote the article, that the Bills lack "an "it factor," some combination of intangibles that converts into an extra dozen yards or so in a close game[.]" In fact, it was a hook for the prediction that the Bills would lose to the Buccaneers that next Sunday."

....this would defeat the purpose of statistical analysis then. If you an say 'Team X is better than Team Y', but 'Team X has no je-ne-sais-quoi to really prove that they're good', then what does that say about the number crunching?

10 It's been apparent to me,…

It's been apparent to me, for a very, very, long time, that the outcomes of close games very frequently turn on random events, but the best teams much more frequently remove random events as an important factor in determining outcomes.

11 Correlation question

Seems to me that there's some possible underlying correlation here. eg a good team in a crap division gets 4 or 6 shots at a Stomp, versus a good team in a good division getting 0 or 2 (but more shots at Guts). And that first team is also likely to win their division & get a better seed, leading to easier early playoff games (plus possibly a bye and/or a chance to rest starters in the final week if they've got the division sewn up). More Guts maybe comes from more tough opponents, correlating with a tougher schedule and a tougher division, and worse seeding + unrested starters.
ie maybe Stomps show team quality, or maybe to some extent they show the value of a weak schedule and a weak division.

I'd be curious to see if anything changes if you looked at Stomp% or Gut% (ie % of potential stomps become actual stomps), to correct for schedule strength.

Id also be curious about breaking opp quality down a bit further. Winning by 14 against a 7-9 opp is probably meaningful, winning by 14 against a 2-14 opp may not be very much. Or maybe it is, we can't tell from what we see so far.

16 I did this exercise with…

I did this exercise with every playoff game in the last 21 seasons, for all four categories of wins, as well as DVOA, overall wins, home-field advantage, and the number of wins that don't belong in any of the four "buckets" (either nine- to 13-point wins, or wins against exactly .500 teams).

W... wait. You extended the original guts/stomps by adding the losses... and didn't include them in the playoff % table too? I mean, at least add a teaser: "what about the losses? You'll find out in part 2!" or something. Not even at all? Nooo!

The question to me really isn't "does a team need to eke out a win versus a good team to be good?," it's "if a team is outclassed versus good teams and stomps the hell out of bad teams, how do they do in the playoffs?" The problem with that, of course, is that there aren't many teams like that at all anyway.

But at least the Super Bowl numbers kindof give that away anyway, since the "domination" vs "outclassed" ratio is just so big. I mean, if you reorder the table (domination/gut/gutless/outclassed and stomp/skate/stinger/embarassment), the distributions look like:

vs winning: 45-42-42-18
vs losing: 102-58-17-4

So in some sense you can at least see that if someone says "I don't care that they crushed the Jets, they need to at least show up versus the Chiefs" they're not crazy: Super Bowl teams stomp losing opponents 56% (*) of the time, and show up 87% (*) of the time. So in some sense, you can word it like "how much you beat a bad opponent by isn't that important - getting crushed by a good team is a way worse indicator."

*: these numbers aren't correct obviously because there are classes of games missing as noted in the article. But good enough to show the point. 

26 Skates also seem important

Based on the playoff chart, it appears the Skates are more predictive at picking the loser than Stomps are for picking the winner.  This makes some sense: if you're barely beating below .500 teams, playoff success may not be in the cards.  I wonder how a Stomp/Skate ratio could predict a team's playoff success.

Save 10%
& Support the Writers
Support Football Outsiders' independent media. Use promo code WRITERS to save 10% on any FO+ membership and give half the cost of your membership to tip the team of writers.