Risky Business
EdjSports examines critical decisions and their impact on GWC (Game-Winning Chance).
Chase Young
Photo: USA Today Sports Images

It doesn’t require a sophisticated analytics model to know that turnovers are bad. Surrendering possession near the line of scrimmage is almost always very costly to a team’s winning prospects. A second consideration is the advancement of the ball after the turnover. It is instinctual for players to try to score after an interception or a recovered fumble, but it often comes with risk. Defensive players, particularly those who are not accustomed to carrying the ball, are taught to fall on a fumble recovery rather than scoop it up and keep running. There are exceptions, of course, but we often see valuable turnovers squandered because a player tried to do too much. The trade-off between the value of securing the loose ball and the value of advancing it is worthy of some deeper analysis.

In the fourth quarter of last night’s game between Philadelphia and Washington, each team fumbled at a crucial moment when a considerable amount of GWC was at stake. We will look at one of those impactful plays and examine the comparative game states from the defensive teams’ perspectives. The GWC values were generated by the EdjSports simulation model based upon what would have been the resulting game state.

At 9:42 of the fourth quarter, the Eagles faced second-and-5 on their own 46-yard line trailing 17-14. After a botched snap, Nate Sudfeld was unable to pick up the ball, and it was deflected to the Eagles’ 36-yard line where Washington’s Chase Young scooped it up and attempted to score. He was pushed out of bounds at the Eagles’ 25-yard line.

Comparative Scenarios:

To assess the risk and reward of Chase Young’s attempt to score we will treat a fumble recovery by Washington at the 36-yard line as the baseline (assuming Young could have simply fallen on the ball).

With the generous assumption that he would score a touchdown:

Risk: (80.2 – 66.9) = 13.3%

Reward: (92.3 – 80.2) = 12.1%

Ratio: 13.3/12.1 = 1.1

The more realistic (actual) scenario:

Risk: (80.2 – 66.9) = 13.3%

Reward: (83.2 – 80.2) = 3.0%

Ratio: 13.3/3.0 = 4.4

To put this in perspective, Chase Young needed to be 1.1X more likely to score a touchdown than to botch the fumble recovery to justify his effort, and he needed to be 4.4X more likely to advance the ball 11 yards than to botch the recovery. Watching the video, it is difficult to fault him for picking up the ball and running, but this analysis illustrates how the risk/reward of such heroic efforts can often be misguided in the context of winning the game.


9 comments, Last at 07 Jan 2021, 5:53pm

1 How common are botched…

How common are botched fumble recoveries? Are they rare enough to make this decision trivial? If there's only say, a 1% chance of getting McCree'd on average, it seems like a safe bet that the returner has a better than 4.4x chance of getting down safely. 

2 I don't think the issue is…

I don't think the issue is getting McCree'd, but rather, whiffing and failing to secure the ball at all.

My favorite instance of this was from the Vikings against the Falcons in 2008. Atlanta was up by 10, just shy of the goal line, when the ball carrier fumbled. A Viking had the opportunity to just fall on the ball, which would have given the Vikes possession at their own 1 yard line, down 10, but the guy attempted to scoop and score. Alas, he whiffed, and Atlanta recovered in the end zone for a TD to go up by 17. They eventually won by 7. I was immensely frustrated at the time, but now that I think about it, the defender would have had a free path to the end zone, to narrow the deficit to 3. Even though it didn't work out, I think he was right to go for it.

3 Botched fumble recovery

I think the implication is the risk between "recover the fumble" versus "scoop up the fumble and run with it" versus "accidentally push the fumble out of bounds/not control the fumble"  that eventually lets the other team keep possession. I think defensive players, in the heat of the moment, do not have the ability to evaluate this risk in real time. My opinion--if you are tall (harder to bend over), moving quickly, or in a congested situation, fall on it. Otherwise, IMO, possession of the ball is more important than any possible return. 

5 I remember being coached…

I remember being coached from the elementary school level to just fall on the ball. There are some edge cases where that strategy is suboptimal, but unless the ball bounced right into your hands, just cover the thing and ensure possession.

\we were similarly cautioned to not attempt to field a punt unless we were the punt returner

4 How common are failed falls?

I think the biggest assumption the article makes is:

assuming Young could have simply fallen on the ball

How many times do we see 3-4 linemen dive at a loose ball, each not quite managing to bring it in? The Saints safety vs the Chiefs appears to be a textbook case of this -- one Saints player, with no significant opponent pressure, just has to simply fall on the ball in the endzone. Which he does, and the ball comes free and flies out of bounds for the safety instead of the touchdown. Then we have this week's Browns and Giants "turns out it's difficult to even sit on a ball that's right under you" recoveries that almost weren't.

Which means the true baseline risk is somewhere between "Young recovers at the 36" and "Philly recovers at the 36". It's probably not an exact split-the-difference, but let's go with 75% GWC as being more or less between the two. The risk is down to 8%. The reward is up to 8%. So, as long as you've got time to pause for a split-second, crouch, and pick up the ball with both hands before an opponent arrives (Young did) -- yeah, go for the yardage. The case against it, as presented thus far, is not compelling.

7 I would agree that players…

I would agree that players can not always just fall on the ball successfully and that needs to be part of the consideration. In this instance it appears Young would have had a very high percentage chance if he chose to fall on the ball as he was somewhat uncontested.   Another consideration is that even after a successful scoop and run, there is an increased risk of being stripped as the defensive players are not used to being ball carriers.  The article is intended to shed some light on what is a common practice by defenders.   This instance served as a good general example to examine the risk/reward in GWC units, and is not meant to be a criticism of Chase Young.

8 Yes, understood, and I wasn…

Yes, understood, and I wasn't trying to criticize the analysis, although my comment may have come across as such.  By "missing" I meant that this data didn't exist, not that it existed and you ignored it.

The point about defenders subsequently fumbling the ball back is a good one.  Intuitively, that matches what I think I've seen in many games.  What the actual % chance is, though, I do not know.

9 Falling on the ball

Agreed, based on what we see all the time, successfully falling on the ball is not trivial.  Years ago I remember hearing an ex-player announcer saying that you need to fall next to, not on, the ball. Then you can gather it in without it squirting away.  With movement tracking technology like we use to mimic individual players in nfl madden or whatever, you’d think someone could do a good study as to the best technique(s) to use in different circumstances for recovering a fumble.