Pro Football Reference Blog

The data itself is very interesting, but I think you are looking at this topic backwards. A bad offense can not afford many turnovers, hence they must try especially hard to avoid them. If this year's Lions turn it over 3 times there is little to no chance they can win the game, and thus it is critical that the bad offense avoid turnovers at all cost. On the other hand, lets look at the 2007 Patriots offense. If they turned it over 3 times, they could still win the game because they would be able to score on almost all of their other possessions. Therefore, there is no reason for them to stress turnover-free football as much as the Lions. Hence, bad offenses HAVE to prevent turnovers, where good offenses do not. The "turnovers per non-scoring possession" rates are a result of the offense being bad or good (for the reasons described above) and NOT the cause of the offense being good or bad, as you seem to imply.

I'll add this--good or bad on offense, a turnover is almost always worse than a punt--because of where the other team gets the ball. Which are most common--INT returns for TD's, fumble returns for TD's, or punt returns for TD's? (Somebody run the data, but I would bet INT's.) Unless it is a non-returned INT on a bomb, a punt gives you more chance to win the game than the turnover. Now, we all like to criticize a coach for the draw play that the defense was waiting on on a 3rd & long, or the 3 yd. checkdown, etc. but they are generally playing the percentages. Of course, the score/time left in the game may say that the coach/team was trying to keep from getting beat by 4 TD's instead of 3.
On the topic of Peterson, I think he ought to focus on ball security. (I have NO INTEREST AT ALL in the Vikes). With RB fumble rates league-wide pretty low, I'd say that he shouldn't fumble more than 3 times a year. As mentioned by Andrew in #1, the Vikes are a mediocre offense who need ball security to maximize points, possessions, and ToP. (IMO as a Saints fan, Reggie Bush suffers from trying to do too much, resulting in lost yardage and fumbles. Pierre Thomas, on the other hand, is ALWAYS fighting for more yds, pushing the pile forward, and has 1 fumble on almost 300 career touches (rush, rec., and returns).)

Where are the stock market analogies?
"Low risk, low reward; high risk, high reward"
"Buy low, sell high"
And my least favorite, the oft-untrue "It takes money to make money"
???
Risk/Reward: The less willing you are to expose the ball flor potential danger (in a calculated risk sense), the less freely you will gain as many yards as if you were a bit riskier on offense.
Buy low/sell high: Create "cheap" takeaways on defense, and make the other team earn the turnovers that you do commit.
Takes money to make money: You need the ball to score. If you have the ball, you are in shape to potentially score. You might turn it over, but I'm guessing that successful teams don't really worry about turnovers all that much. Similarly, successful investors try to make the smartest "scores" they can in the market, not really worrying about the occasional bust. Good offenses know they will turn the ball over now and then, but they are in it to score, not to worry about turnovers.

Elite RBs holding onto the ball is a relatively recent trend in the NFL. Marshall Faulk, during his 99-01 MVP-level run, fumbled a total of 5 times. But before that, especially in the 70s and 80s when guys carried the ball with one hand, fumbles were incredibly common.

Eric Dickerson fumbled 27 times during his first two seasons, which no one -- not even Adrian Peterson -- could get away with today. Walter Payton fumbled 11 times in 14 games during his 1977 MVP season.

I have little doubt that runners decrease their productivity in other areas by getting to the airport too early. It would be interesting to see if there's an optimal number of fumbles that has any statistical significance. I suspect this would be very difficult to test, since fumble totals have changed so much in the past decade or so. It's unusual now for more than one RB to top six fumbles in a season.

I don't know how useful this really is, but I rate RBs every year, and here are the fumble totals for my top-rated back every year during the 16-game schedule. A clear difference is apparent between the 80s and 90s, then again between the 90s and the 00s.

78 W.Payton - 5
79 E.Campbell - 8
80 E.Campbell - 4
81 B.Sims - 9
82* M.Allen - 3
83 E.Dickerson - 13
84 E.Dickerson - 14
85 M.Allen - 3
86 E.Dickerson - 12
87* H.Walker - 4
88 E.Dickerson - 5
89 B.Sanders - 10
90 B.Sanders - 4
91 T.Thomas - 5
92 E.Smith - 4
93 E.Smith - 4
94 B.Sanders - 0
95 E.Smith - 7
96 T.Davis - 5
97 B.Sanders - 2
98 T.Davis - 1
99 M.Faulk - 2
00 M.Faulk - 0
01 M.Faulk - 3
02 P.Holmes - 1
03 L.Tomlinson - 2
04 T.Barber - 5
05 T.Barber - 1
06 L.Tomlinson - 2
07 L.Tomlinson - 0
08 D.Williams - 0 (through Week 16)

It is apparent that young players fumble much more than veterans.

I buy the heart of the article, but not the AP example. For RBs ball security is largely an innate skill. It's learnable (Tiki Barber comes to mind), but backs usually are who they have been fumble wise. There are DeAngelo Williams' and Joseph Addai's (fumble less than 1 per 200 carries) and there are your Peterson's (who has fumbled more than 1 per 50 carries).

Hmm, I've also wondered if trying too hard results in more injuries. Trying to eke out that extra two yards while stumbling and off balance seems like prime injury time too.

Perhaps this can still be explained without resorting to "trying too hard" though. Good teams are more likely to be on parts of the field where an interception/fumble is less costly, like that weird outside-field-goal-range-but-too-close-to-get-much-out-of-a-punt area. Bad teams will more likely be spending time on their own side of the field, where a turnover is more likely to give the opponents an easy score.

Take an obvious try-too-hard situation, 3rd and long. If I have the Colts, I'm not all that uncomfortable about letting Manning throw it. He may throw the occasional pick, but he's also got a good chance of converting. If I have the Lions... Well, even if the odds of an interception are equal, the odds of converting certainly aren't. And with the Lions, I'm probably at my own 20 yard line still, where with the Colts, I may be around midfield. etc. etc.

This argument is a bit like the one that emerges from those insufferable articles bemoaning Adam Dunn's strikeouts -- "if only he put the ball in play more often, think of how many more hits he'd have!" Well, no. Adam Dunn putting the ball in play more often would require him changing the hitting style that gets him 40+ HR... He may hit for a higher average with the new approach, but is it worth the incumbent loss of power that comes with a new, contact-centric approach at the plate?

The same goes for offenses and turnovers. Ask Peterson or the Greatest Show on Turf Rams to be more careful with the ball, and they could do it... But it would require a fundamental change in the way they approach the game on every play, and would likely be a tradeoff that isn't worth making. If we could cherry-pick bad outcomes (strikeouts, turnovers) ex post facto and say "let's have a do-over where you play more conservatively," then you could say somebody like Peterson would be even better if he cut down on his mistakes. But AD takes the same approach to a run when he busts a long gain as a run when he fumbles, so there'd be a cost involved in being more conservative. Unfortunately, this concept of trade-offs is something a lot of fans don't get.