Turnover luck

Here’s how Bill Connelly defines it:

The concepts are pretty simple: Over time, you’re going to recover about 50 percent of all fumbles, but in a given year, you might recover 70 percent, or you might recover 30. The same goes with passes defensed; on average, you can expect to intercept about 22 percent of the passes you defense. (Passes defensed = interceptions + break-ups.) This is a bit mushier a concept … but over time a particularly butter-fingered year will be balanced by a sticky one.

The way I measure turnovers luck is pretty simple: how far were you from 50% with your fumble recoveries? How far were you from ~21-22% in your interceptions-to-breakups, offensively and defensively? By looking at broader numbers, we can basically create an expected turnover margin (called Adj. TO Margin below) and compare it to the real margin. The difference becomes your Turnovers Luck for the season. On average, a turnover is worth about five points in terms of field position value, so I take the difference between turnovers and expected turnovers, multiply it by five, divide it by games played, and voila: Turnovers Luck Per Game.

Georgia finishes as one of his top teams from last season, which shouldn’t come as a surprise to anyone who’s familiar with Georgia’s 2014 turnover margin.  Similarly, there’s a huge swing from the 2013 luck tally. What you wonder about in the Dawgs’ case is how much of that luck came from changes in the defensive approach to turnovers and the emphasis on offense in avoiding turnovers at the quarterback position and how much was randomly generated.  Sample size can be a bitch, you know.

40 Comments

Filed under Georgia Football, Stats Geek!

40 responses to “Turnover luck

  1. True, but man, did we have some back luck against Tech this year. Those two fumbles inside the 1 was the reason we lost that game. Lady luck pulled the rug from under us.

    Besides that, I think we were pretty good. I think the defensive approach to not giving up a big play and forcing the opponent to run more plays to score also was a part of it. The more plays your opponents run, the more they have the chance to turn the ball over.

    Like

    • Yours truly,
      Captain obvious

      Like

      • Mayor

        Dear Captain Oblivious: No. Sh!t happens in football games. Turnovers are part of any game. Tech had a critical turnover in the 4th quarter when the Tech QB dropped the ball and Georgia recovered. Mason then took the team on a great drive down the field and scored what should have been the winning TD with 18 seconds left in the game. The reason why Georgia lost that game is the HC called a “pooch” kick-off that allowed Tech the opportunity to kick a game-tying FG and then win the game in OT. I just don’t get how you can go back to the beginning of a game and point at one or two things as the cause, when what happened at the end clearly caused that loss. The team played well enough to win, warts and all–the coach then blew the game with a bad decision. Face it.

        Like

        • Fair enough. Those turnovers were a factor tho, which was my point. I’m not gonna defend the pooch kick thats for sure.

          Like

          • Cojones

            There is a forgotten play of the big gain that put them in position to tie with a great kick from a kid who deserves all the credit. Those two plays snatched defeat from the jaws of victory, not the squib kick. Obfuscation of facts shouldn’t advance any argument anymore than using that tactic against Bobo and Richt didn’t work.

            Any time you guys throw out the squib kick, you deserve this reply and you will continue to get it.

            Like

            • Mayor

              All credit to the Tech QB and kicker but the pooch kick-off opened the door to those 2 plays. Without the pooch even if the Tech QB runs his big gainer they would still be too far out to kick the FG.

              Like

            • Macallanlover

              Perhaps some will, but don’t count on one traqck being able to learn a new trick, or broaden his mind to anything beyond his one, myopic thought. There were a hundred+ plays, and thousands of options in that game, but to this kook it was one option not utilized, which could have turned out worse, or the same. This football is too complex for him.

              Like

              • Mayor

                What we have here is the definitive example of the “doesn’t have anything substantive to say so attack the other poster” school of (lack of ) thought epitomized by the in-house blog drunkard. He hasn’t had an original thought in years–so just make up an insult. There were a hundred+ plays and thousands of options in that game and he doesn’t understand any of them, particularly the ones that win or lose the game.

                Like

                • Macallanlover

                  But you do…sadly laughable. Frankly no one knows what the uncalled plays/decisions would have done to the outcome, but good to know that you have that nailed down. I am sure everyone is comfortable with your glimpses into the unknown, or have the definitive answer on what should have been done. Guess we can just cancel all future games and read your book.

                  Like

                • Mayor

                  For Pete’s sake, there were 18 seconds left in the game and Georgia was leading by 3 and kicking off. Any deep kick and the game is over as long as the kickoff isn’t run back for a TD–and Tech hadn’t scared a return all game. Just kick it through the endzone. If the kicker can’t kick it though the endzone, kick it as far as possible and the returner uses up most of the 18 seconds on the return. But pooch it? Up by 3? Every TV football commentator on every sports program in the nation that night said Georgia should have kicked it deep. Not to mention that CMR admitted after the game that he should have kicked it deep. But not you–you always know more than everybody else. This just proves even more that you don’t know sh!t about football. You’re a legend in your own mind.

                  Like

                • Macallanlover

                  I feel for you man.

                  Like

        • Cosmic Dawg

          I thought the pooch kick was a horrible call, and I thought the fraidy cat prevent defense that followed it was an equally bad (if less obvious) bit of coaching.

          However, just because Dumb Event A happens in the first quarter and equally Dumb Event B happens at the end of the game doesn’t mean B was worse – the points don’t count for more at the end.

          That game was so obviously a coaches-players-everybody loss, I’d be surprised if the guy doing the laundry on Friday didn’t screw something up, too.

          Like

      • Russ

        Yeah, but hard to overlook the 14 point swing when we took the ball from their QB and ran 95 yards because of no whistle. Breaks were pretty even in that game, even if they were ALL huge. We just happened to gack last.

        Like

  2. Macallanlover

    Bill is a man of statistics but I feel he is right on this, the amount of just pure luck/rub of the green/fortuitous bounces, etc. plays a role. No one can prove that, and doubt they ever really “even out” over time, but Ol’ Lady Luck plays a role. I am not discounting that being in the proper position and using the best techniques can increase your chances of turnover opportunities, they do. I noticed a difference in the way the UGA defensive secondary attacked the ball when it reached a receiver’s hands as early as the Clemson game. Our dramatic increase in TOs this year certainly can be attributed to both solid coaching and good fortune. UGA is due a few more seasons of breaks coming our way.

    Like

    • Brandon

      Recovering lose fumbles/interceptions on batted balls = luck, some. Taking care of the football and not throwing INT’s which is half of this “margin” has absolutely nothing to do with luck

      Like

      • Good point, though I wouldn’t say “no” luck at all is involved. Mason did have a few this year (and Ramsey had one in the bowl game) that hit defenders square in the hands, and the guy just dropped it. Keep in mind this article isn’t really looking at how many or how few fumbles/passes defensed happened…….moreso it’s looking at when they DID happen, what percent turned into turnovers?

        Like

      • Mayor

        +1. The problem with this whole “luck” thing is that it ignores effort, skill, coaching and coachability. If there is a fumble and 2 players (1 from each team) are right next to the ball, and one goes after the ball aggressively while the other makes a half-assed effort, whom do you think will end up with the ball? Hustle beats luck every time.

        Like

        • 81Dog

          exactly, Mayor. Luck or chance is what determines the flip of a coin. While there is some luck, or perhaps, random chance, involved in the generation and recovery of turnovers, there is also a fair amount of it that isn’t random. Treating it as random across the board doesn’t seem like a sound statistical approach, does it? It would certainly be hard to factor in the non-random elements for some kind of sabremetric approach, but that’s why places like the Elias Sports Bureau exist, I guess.

          Like

        • “Hustle beats luck every time.”

          I get the sentiment, but an absolute statement like that simply isn’t true. Sometimes….heck, A LOT of times, the ball hits the ground and just bounces directly to one player or the other. Or a pass gets deflected and either deflects to a defender or it doesn’t. While hustle/effort/technique etc may go a long way towards creating the fumble or deflected pass, it often times has nothing to do with where the ball ends up.

          Sure, there are 50/50 chances that occur, but I’d say at least as many or more have nothing to do with hustle once the ball is on the ground or deflected in the air. How often do you catch yourself thinking “man if he had hustled more, he would have gotten that fumble”? I rarely find myself thinking something like that, pretty much everybody gets after a loose ball pretty hard.

          Like

          • Cojones

            The last part of your last sentence is spot-on. I don’t believe there are “half-ass” attitudes on fumbles anywhere nowadays.

            Like

            • Mayor

              Spoken like a man who didn’t watch many Georgia games in 2010. By the end of that season several guys were mailing it in. That’s why there was a purge in 2011. Remember “energy vampires?” Remember the Liberty Bowl where damn near the entire team acted like they didn’t want to be there? There was plenty of “half-ass” effort on that team and while I haven’t done an NCAA-wide study on the subject (and will not as I have better things to do) I guarantee you that there were plenty of players and teams with attitude problems last season that would skew the “luck” analysis by lack of effort.

              Like

              • Cojones

                I’m sure that all of those events/ people were considered in the article. Lifting certain parts out and cherry picking a year of player attitude doesn’t help your argument.

                Like

      • Biggus Rickus

        Well, no, it isn’t. The number of turnovers by your offense still involves falling on loose balls and luck on potential interceptions. There were several dropped interceptions by Louisville in the bowl game, for example. On the season Georgia fumbled 21 times and only lost 7. That’s good luck. The year before they were right around average, losing 10 out of 22. Georgia’s opponents have been right around average the last two years, losing 8 of 18 in ’13 and 13 of 25 in ’14. Now, they did cause more fumbles this year, which is good. How much to credit anything in particular for the increased number of fumbles caused is hard to determine.

        Like

        • Biggus Dickus

          So the interceptions dropped by the Louisville DBs were “luck?” Hitting the ball carrier so hard he coughs up the ball is “luck” also, eh? I guess winning games is ‘luck,” too. Kentucky has been awfully unlucky ever since Bear Bryant left.

          Like

          • Biggus Rickus

            Yes, winning football games is luck. That’s the logical end of the argument that luck plays a role in turnovers. Jesus.

            Like

      • Macallanlover

        True to an extent Brandon, but the luck of the bounce can make the ball comes back to you, go out of bounds, or go to your player rather than yours….even if you fumbled it. I don’t minimize better ball security, and defensive technique/positioning but you will always have an element of luck. And you see it every week.

        Like

  3. Not saying coaching plays no role at all, but look at the discrepancy between us and Bama from this past year. I do think the Tide has lost a bit of their luster, but with THAT MUCH of a discrepancy, makes me think luck is more at play than coaching. That oblong ball bounces in crazy ways.

    Now if we look at things like raw number of passes defensed, I think coaching has more of an impact on those numbers than the percentage that turn into interceptions.

    Like

    • Cojones

      ‘ll never forget Jones knocking the ball forward before the gator player crossed the goal line – and right to a UGA player in the end zone for recovery. There were a number of plays/fumbles like that.

      Like

  4. Rick

    I think bill makes a good case for his metric being 100% luck. By looking only at passes defended and balls on the ground, he is controlling for any differences in scheme that might be causing more of those.

    UGA had an unlucky season overall because all but one of the losses were so close. But when it comes to TOs specifically, they were very fortunate.

    Like

  5. DawgFaithful

    Yeah. That luck thing is bullshit. I’m not sure if it even exists. Momentum is real but not sure about luck. If you have premium talent and you play sound defense, you’ll get turnovers. It comes down to coaching too. You have to teach them to force turnovers. What coach doesn’t do that though? Sometimes there’s what appears to be luck. Sometimes the QB makes a stupid throw but most of the time turnovers are forced by good defense.

    Like

    • Forcing turnovers, yes. But fumble recovery is somewhat random, depending on where it happens (best chance of successful recovery comes behind line of scrimmage).

      Like

      • Cojones

        That was shown statistically to be true from an article you posted last year. You may need to put it up again.

        Like

      • Mayor

        You make a good point about the “where” a fumble occurs, Senator. If the QB has dropped back to pass, gets hit and fumbles, likely the D will recover because the rushers are next to him when the ball comes out. Same with a receiver catching a pass then fumbling–he’ll be surrounded by DBs so the D will likely recover. These are not lucky recoveries. They are recoveries that you would expect to happen because of circumstance. This is why ball security is so important.

        Like

    • D.N. Nation

      A few year ago, a Saint Saban-led Alabama team was around last in the nation in fumble recovery.

      Like

  6. Daniel Simpson Day

    Turnover Luck is inversely proportional to Pre-snap Hand Waving.

    Like

    • Cojones

      Those are on-field stretching exercises. Or signals to the opposing QB which mean “Throw it this way. I don’t know what I’m doing because I didn’t get the communication”.

      Like