Louisville football by the numbers: Turnovers and what could have been

On November 20th, 2020, with three games left to play CC’s own deputy editor, Keith Wynne, wrote a sobering and brutally honest column titled, “Louisville Doesn’t Value The Football and They Won’t Reach Their Potential Until They Do.” Almost two months later as we look back on whatever the hell the 2020 college football season was, the question remains the same for Louisville: “what if we didn’t turn the ball over so much?”

Louisville went on to finish the season with a -1.09 turnover margin; good enough for 119th in the country. For those who aren’t football junkies reading this (Hi, Mom), the turnover margin is the number of takeaways caused by the team’s defense minus the number of giveaways caused by the team’s offense. So, if you’re losing the turnover battle on average by a full point or more, that’s real bad.

Now obviously turnovers aren’t the only factor when it comes to winning and losing football games, but when your team follows up a 8-5 season with a 4-7 clunker that includes four losses by seven-or-less points, then yeah, it matters a hell of a lot.

In fact, SBNation’s Football Study Hall did a fascinating study a few years back where they took the turnover margin data from the 2009 to 2012 college football seasons and made a model that showed how turnover margin correlated to winning percentage. Each dot represents a team in that four-year period, and the per-game turnover margin is on the X-axis with the winning percentage on the Y-axis:

“The R squared (coefficient of determination) for the trendline shown in the chart is 37%, meaning that turnovers “explain” 37% of the win total data in this model.” Even in just a four-year sample, to find that per-game turnover margins directly correlate to the winning percentage of more than a third of college football teams is staggering. It just further proves the point of how important it is to win the turnover battle, and how losing those battles usually leads to losses.

Football Study Hall then took their model and converted it to a formula which gives us how many wins a team with a turnover margin of “x” should achieve given the historical data. The formula’s output was then adjusted from a winning percentage to number of wins in a 12-game schedule:

Even though UofL played just 11 games this past season, it’s not hard to see how the data lines up here. And hey, if you’re an optimist, you could even say that Louisville making it to four wins with a -1.09 turnover margin is a slight overachievement.

But Football Study Hall explains the best way to look at this data is as follows: “Let’s say your team is talented enough to win around six games. For every 0.5 in turnover margin that they are able to achieve, you can expect around 1.25 extra wins.” The same works in reverse.

I think we’d all agree that based on the 8-5 bowl-winning season we saw in 2019, the amount of talent we had returning in 2020, and how well the team played when they weren’t giving the ball to the other team, that this team was more than talented enough to win six games. So if you take that and apply the theory in reverse where for every 0.5 in turnover margin you lose, you can expect about 1.25 extra losses you end up with….yep, you guessed it, a four-win season.

TL;DR: UofL obviously was better than their 4-7 record and your “what-ifs” are valid.

The biggest thing that upsets me about the turnovers, other than their direct correlation to our losses, is how it marred the perception of how much the defense continues to prove and what a fantastic job Bryan Brown has done.

22%! Surrendering 1.98 points per drive is obviously not where the defense wants to be, but it’s a massive improvement from where they were last year, and light years ahead of where they were in 2018. And to find out that 22% of those points surrendered came right after the offense turned the ball over makes that 1.98 ppd look even better. This unit and Brown’s staff deserve a hell of a lot more praise for their improvement, and the offense putting them in bad situations shouldn’t be ignored.

Where do we go from here? Well, the bad news is that turnovers are almost entirely luck-based. There isn’t a magic bullet that causes teams to create more turnovers on defense and turn the ball over less on offense. And analytics can’t tell you who is more likely to turnover the ball more than others from year-to-year because it’s virtually never consistent. That’s why they call it turnover luck.

Many people, including myself, are with Keith in believing that having a turnover “bug” is largely a mental issue and that it can be fixed. I think there’s a lot of truth to that, especially when Keith talks about how the culture of this program is being built around accountability and teamwork, and how that should be the cure for the issues Louisville had this season and in the past.

But I’m also a believer in that statistics show that bad turnover luck doesn’t last, and that a team suffering through turnover woes like Louisville will eventually see their turnover margin progress to the mean. Now, Keith has a good rebuttal to that point since Louisville has averaged 24 turnovers a season since joining the ACC in 2014. But where I see a potential for the bad “luck” to end is that in 2018 with a brand new staff UofL saw that number dip to 18.

So, I’m hopeful, as Keith is, that the further we get away from Bobby’s lingering stench, and the more this staff implements their values, especially accountability, that the ball will literally start bouncing Louisville way more. And hopefully that leads to a more accurate representation of how strong our teams are, which should in turn mean more wins.