Forgotten facts about fumbles

I originally planned this article as a defense of Antonio Gibson, one of my favorite Commanders. There were more over-the-top comments about Gibson’s fumbles in recent Hogs Haven discussion threads than I can count. KyleSmith4GM led the charge with an article entitled “It’s Time to End This Antonio Gibson Experiment”. Exchanges were heated. When I pointed out that the leading rushers in NFL and Washington history fumbled at a far higher rate than Gibson, someone argued that Emmitt Smith and Walter Payton could not succeed in the NFL today because of “analytics,” which left me scratching my head.

There were two primary criticisms of Antonio Gibson on Hogs Haven. One was that he wasn’t a natural running back, which I find peculiar when talking about the sixth leading rusher in the league last year, despite having a broken leg and an injury-riddled offensive line. But that topic is not the focus of this article. The other criticism was that he fumbled “all the time,” based on leading the league with 6 fumbles last year. I’ll cover that issue here.

In some ways, my article here seems pointless now. After Gibson fumbled early in the first preseason game – the worst possible timing – he is riding the bench. Shiny new toy Brian Robinson Jr. is the starter based on his usage in the past two preseason games, and he is clearly the new fan favorite. The Gibson-as-a-starter ship has sailed. But was there a rush to judgment based largely on his fumbles?

Before I get into the data, some methodological notes.

• NFL statistics on fumbles are for all activities – rushing, receptions, and returns. In fact, Gibson fumbled 4 times on runs and 2 times on passes last year. So, you have to look at touches for all plays to assess fumble rates.
• I ignore fumbles lost. Fumbles are random enough. Fumbles lost are pure luck, depending literally on how the ball bounces – and whether one’s teammates are alert.
• Raw data for some tables were pulled from multiple sources (NFL.com, Pro-Football-Reference.com, and statmuse.com). Different sources can use slightly different methods of compiling data. I apologize in advance for any accidental errors.
• I limited my analysis to data from recent years. This addresses complaints about using data from earlier eras (e.g., “I don’t care what Ernest Byner did in 1988”).

I came to four primary conclusions after looking at the data.

1. Fumbles by leading RBs are mostly random events.

Anyone with statistical training would look at tables showing the small number of running back fumbles versus touches and would wonder if their fumbles are anything other than random events.

Gibson had only one fumble in 2020 on 206 touches, which everyone forgets. Were his 6 fumbles on 300 touches in 2021 a sign that he has developed the incurable disease, Fumbilitis Pigskintitis?

Table 1 shows the fumble history of the ten running backs who have the greatest number of yards from scrimmage during the past five years (per statmuse.com) by career year (per Pro-Football-Reference.com). Gurley began in 2015, and three others began in 2016, and we have at least 5 years for all of them. What do the data show?

Data Sources: statmuse.com, Pro-Football-Reference.com

• All RBs fumble. The number fluctuates wildly from season to season. Only Joe Mixon had more than one career year in which he did not fumble at all, and he had one season with 3, one with 2, and one with 1. Our old friend the Ewok had the craziest pattern, with 5, 1, 6, 3, 6, and 1 in consecutive years. Dalvin Cook and Melvin Gordon had multiple years with 4 or more.
• You might think that younger players fumble more, then learn to hang onto the ball. You would be wrong. The career high year for fumbles fell all over the place.
• Based on Gibson’s first two years, what can we safely predict about his fumbles in future years? Absolutely nothing. Right now, he is fumbling slightly more than the players in this list, but one year like his rookie year and he could be way up on the board. One year’s fumbles does not predict the next year’s.

2. Antonio Gibson led the NFL with 6 fumbles last year. That is a not a high number, even today.

There are two tables that support this conclusion. Table 2 looks at the fumble leaders for each year of the past five. What does it show?

Data Source: statmuse.com

• Gibson’s league-leading 6 wasn’t exceptionally high in the current era. It would not have led the league in 2017, 2018, or 2019.
• J.D. McKissic, who gets lots of love on this site, was on the leaderboard with 3 in 2020, a year in which Gibson had 1. So, is McKissic a bum too?
• The league’s greatest contemporary runners are among those who fumble the most. In fact, the league leader in total yards among running backs was among the league leaders in fumbles in three of these five years (Jonathan Taylor in 2021, Derrick Henry in 2020, Todd Gurley in 2017, all shown in bold). There is a ringer or two among the league leaders every year (such as Demetric Felton in 2021, Corey Clement in 2018), but top backs are always there on the leaderboard.
• If you want to be an All-Pro running back, fumble a few times. An All-Pro RB according to one or more major lists was on Table 1 in each year (Jonathan Taylor in 2021, Derrick Henry and Dalvin Cook in 2020, Dalvin Cook 2019, Ezekiel Elliott in 2018, Todd Gurley in 2017).

3. Touches per fumble are not high either.

What about fumble rate? Maybe we have a different conclusion if we look at touches per fumble rather than total fumbles. Table 3 addresses this. What does it show?

Data Sources: statmuse.com, Pro-Football-Reference.com

• Gibson’s career rate of one fumble per 72 touches is worst on the list, but of course that comes from averaging one very low and one high year. If his first five years were like his 2020, he would be second best on the list (just behind Mixon) at 1 fumble per 206 touches. If he fluctuates like other top backs in Table 1, his first 5 years will be in the middle of the pack.
• Gibson’s worst year is the biggest concern, but 2021 for him is not out of line for top backs. Elliott was slightly worse in his worst year – and was an All Pro. Melvin Gordon was similar in his worst year.
• Washington benches running backs who fumble, while others celebrate them. The last column shows that in their worst year for fumble rate, five of ten were an All-Pro, one was Offensive Player of the Year, and one was a Pro Bowler.
• There is very little practical difference between Gibson’s career average fumble rate and that of the best backs in the league over the past five years. Based on his first two years, we would expect him to fumble 3.5 times per year, or once every 4.6 games based on 253 touches/year or 15.8 touches/game. The best in the league in the last 5 years was Aaron Jones, with one per 213 touches, which is 1.6 per year at his current workload; if you give him the same workload as Gibson, he fumbles 2 times per year. Would you start Jones over Gibson based on a difference of 1.5 fumbles per year? Or how does he compare to the worst RB on the list, Dalvin Cook? Cook averages 3 fumbles per year on a similar workload to Gibson’s. Do those who have praised Cook care about a difference of less than a fumble per season between Cook and Gibson?
• These differences between backs are so small that it’s pointless to calculate statistical significance. All differences are statistically significant if the underlying dataset is big enough. Here, the practical differences are what is important, and they are trivial.

4. The best running backs in the history of the league fumble and nobody cares.

Table 4, showing the four players who are among the all-time league leaders in all-purpose yards and who retired within the past five years, with their fumble history.

Data Source: Pro-Football-Reference.com

• Gibson’s worst year was better than that of three of the four on the table, behind only McCoy.
• Touches per fumble were all over the place. Gibson was far superior to two of these (former Redskin Adrian Peterson and future Hall of Famer as well as Darren Sproles) and well behind LeSean McCoy and Frank Gore on touches per fumble.
• Somehow, these players earned honors despite fumbling, sometimes in their worst year for fumbling. Peterson had 9 fumbles in his worst year, in which he was MVP, Offensive Rookie of the Year, and All-Pro. Sproles was a second team All-Pro in his worst year; he had twice as many touches per fumble as Gibson at his worst.

Bottom line: Running back fumbles are one of the most overrated statistics that you can cite.

Other teams celebrate running backs who produce, even if they have a few fumbles per year. Washington fans want to bench them, turn them into wide receivers, or run them out of town. Antonio Gibson doesn’t deserve to be benched or trashed because of the number of times he’s fumbled.

It’s part of the coaches’ jobs to minimize turnovers, and they certainly have sent a loud message to Gibson to reduce the fumbles. We as fans don’t need to overreact to either the fumbles or the coaching.

Final point: If you really care about fumbles, you’re looking at the wrong position. Most NFL fumbles are by quarterbacks. In fact, the top 20 fumblers in the league last year were quarterbacks, led by Dak Prescott with 14. Look there if you want to cut down on fumbles – not to mention interceptions.