There’s no question that Indiana’s defense struggled in 2015. It was a big reason why the Hoosiers only won 6 games and resulted in several coaching changes. In terms of yards allowed per game, IU improved from 120th in 2013 to 93rd in 2014, the first year under former DC Brian Knorr. Despite showing solid improvement, Indiana regressed in 2015 ranking 120th out of 127 FBS schools.
In general on this site, we like to break things down either into a per-play or percentage basis. While not perfect, it tends to remove other factors. For example, IU faced 148 more plays last season over the prior year. Facing more plays inherently increases the yardage allowed. One reason they faced more plays was because Indiana (thankfully) played in a bowl game[ref]https://en.wikipedia.org/wiki/Bowl_game[/ref] which gave the Hoosiers a 13th game. However, even on a per game basis, IU went from defending 74 plays per game in 2014 to 80 plays per game last year. All else equal, by just defending 6 additional plays per game one would expect the yards allowed to increase. Even if IU maintained their 5.85 yards per play from 2014[ref]They didn’t. Indiana fell to 6.38 yards per play last year.[/ref] on these 6 additional plays you would expect an increase in 35 yards allowed per game. Therefore, we need to look at the evidence on either a per-play or percentage basis.
The method to the madness
Taking the 2013, 2014, and 2015 play-by-play data, I sorted each year from biggest loss forced to largest gain allowed. In 2015, that was a loss of 10 to a gain of 85. I then graphed the average yards per gain at each percentile of total plays. This helps eliminate any noise caused by a difference in the total number of plays between the seasons. For example, at the 10th percentile point, 2015 includes 104 plays[ref]Yes, there is slight rounding[/ref] based on the 1038 total plays that IU defended. At that same point, 2014 includes 89 plays based on the 890 plays defended.
The lines containing the prior 2 seasons follow the same shape and it is difficult to make a conclusion.
What is slightly more illustrating is graphing the increase (or decrease) in the average yards per play. That can better tell us if the improvement or regression was from a higher or lower frequency of tackles for loss, short gains, intermediate gains, or big plays. Doing that can help us better identify the position group most impactful[ref]Positively or negatively[/ref] on the change.
Before we get to last year, let’s revisit the 2014 improvement when the average yards per play dropped 0.89 yards from 6.74 to 5.85. Anytime the graph is above the X-axis it means that Indiana allowed a higher average yards per play at that specific percentile than 2013. When the graph is below the X-axis, Indiana was improved[ref]Meaning a lower yards per play[/ref] over 2013.
The 2014 Hoosiers were nearly even with the 2013 defense until the 30th percentile. At that point the improvement really began. Indiana was able to shave off nearly 0.6 yards per play for the next 60% of plays until the 90th percentile. The 90th percentile in 2014 was essentially gains up to 15 yards. That improvement came from more frequently turning a 2 yard gain in 2013 into a 1 yard gain in 2014. Turning a 7 yard gain into a 6 yard gain and so on. Fewer missed tackles also helps a great deal. For example, a missed tackle in 2013 turned a 5 yard gain into a 15 yard gain. Having fewer missed tackles in 2014, therefore a higher frequency of 5 yard gains instead of 15 yard gains, helped with this improvement.
Finally, what also helped facilitate the improvement in 2014 was by limiting the frequency of big plays. From the 90th percentile onward, Indiana shaved off an additional 0.3 yards per play. The average yards per play on the last 10% of plays decreased from 32.1 yards per play to 29.1.
As a result of these improvements, IU went from 117th in yards per play in 2013 to 91st under new coordinator Brian Knorr. While 91st was definitely not a reason to celebrate, there was a level of optimism in his second year with an experienced front 7 returning. Perhaps Indiana could crack the top 80, which still isn’t great, but combined with a high-powered offense was sufficient at that moment in Indiana’s program.
As you can see, the graph is entirely above the X-axis meaning a higher yards per play at every percentile.
To start, the Hoosiers forced a lower percentage of negative plays last season. In 2014, 9.4% of all IU plays defended were at a loss compared to 8.9% last year. In addition, those negative plays averaged -3.7 yards per play in 2014 compared to -3.5 last year.
Unlike the 2014 improvement when Indiana shaved off nearly 0.6 yards per play in between the 30 and 80 percentiles, the 2015 version showed little advancement in this area.
At the 50th percentile, the IU defense is maintaining nearly the same pace as 2014. In 2015, the 50th percentile would include 519 plays.[ref]Out of 1038[/ref] Those plays resulted in a total loss of 29 yards. If Indiana had been able to maintain this level they would have finished at 5.88 yards per play to rank 93rd. This result would not have been great, but at worst would have only represented a nominal regression.
However, things gradually increase until the 80th percentile when Indiana’s average yards allowed per play increased to 2.3[ref]Compared to 2.2 in 2014, a difference of 0.1[/ref]. Soon afterward illustrates where things got away from the Hoosiers resulting in the 0.53 yards per play increase. The average gain on plays between the 80th and 100th percentile increased from 20.6 to 22.8. This means of the 20% of plays in both seasons that went for the largest gain, that gain was over 2 yards larger per play last year.
Even worse, the average gain on plays between the 95th and 100th percentile moved from 39.9 to 43.6. Indiana allowed 8 gains of 70+ yards last year compared to 4 in 2014. Despite a 17% jump in the number of plays, IU allowed a 50% increase in the number of 70+ yard plays.
Simply put, IU allowed a higher frequency of big plays and on average those big plays went for more yards, a bad combination.
Next week we will break apart the data into rushing and passing plays to determine which areas Indiana regressed the most and how much they regressed. Or put it differently, how close or far was Indiana from repeating their 2014 performance[ref]Again, 2014 wasn’t a good defensive year, but sadly it is the best season that Indiana has had in the last 3 years[/ref]. We’ll try to look at whether it was 1 play per game, more than one, or possibly even less.