Over the past two games we've been quite frustrated with the so-called "Bend Don't Break" mentality of the defense. This is not to speak of the easy first possession TD's for games #2 through #6, which is another problem.

The general perception is that the defense is bending a lot but still breaking. This is not only ineffective as a defense, but affects the offense by limiting possessions and melting clock. Many have adopted "Ted 'Leaky' Roof" as a new slogan.

So, how does this perception stand up to the statistics?

(Stats courtesy of a link Joey Weaver posted earlier, with my additions for Mercer (FCS)).

It's widely accepted that pace of play in GT games has been slow this year, with limited possessions and plays. Due to pace of play, I'll look at points per possession (PPP) and yards per play (YPP). Note that I do not account for non-offensive TDs, aside from explicitly stated exceptions for GT below.

7 weeks into the season, we have enough data about all of our past opponents to make general conclusions about their offense and defense. How many PPP and YPP has their offense averaged through 6-7 games? What about their defense?

We can then see how GT compared in each individual game. Did the defense outperform compared to the opponent's offense typical production?

Note that I make no adjustments for strength of schedule, opponent adjusted defense strength, etc. This is just aggregate data of possessions, points, plays, and yards.

Here are the results:

Adjustments: I added the Miami fumble returns adjustment because those TDs weren't allowed by defense. We don't know if the defense would have stopped those possessions, so the line is dotted while still showing the raw data in solid. I did not adjust for the Clemson "McPick 2" because, well, it was the defense's fault.

For YPP, I would say the defense hasn't really made much of an impact on a per-opponent basis. The best you could say is there are not many breakdowns (big plays). Is 6-7 YPP acceptable to give up? No, but we're not really doing much worse than anyone else. The noted exception is GSU, we really made them dink and dunk.

PPP is more insightful here. For games #1-5, the defense did a good job limiting points if not yards, especially compared to the opposition. In #1-5 especially, I recall the second half of games showed the defense generating stops.

The last two games show a change for the worse. We allowed Pitt 1.0 more PPP than their average, which is a lot. If that was 0.5 instead of 1.0, we win. GSU was better, but there ought to be a talent gap (sorry, Eagles fans) that should limit their PPP below their average.

My guess is the difference has been in second half performances. Games #1-5 showed the defense locking down, games #6-7 have not.

To address the main question of "Bend vs Break", I devised a metric of Points per 100 Yards. The theory is that the defense bends and allows lots of yards but doesn't break to allow points. This does NOT account for field position. The FBS average is 7.13 points / 100 yards, on 367,884 yards of gridiron this season to date.

I also looked at Points per 10 Plays to measure how well we limit explosive plays, which the defense is predicated upon. A bendy defense will allow lots of plays but few points.

Side note: I would also think that "Bend don't Break" should generate more turnovers on a per-possession basis, assuming turnovers are about constant on a per-play basis. I don't have the data, but I doubt we have better than average turnovers per possession. I think 2014 showed an example of that working.

Looking at Points per 10 plays, I would say the defense generally limits explosive plays. In fact, the only two that come to mind are the RB bust against BC and the tipped pass at Pitt. The defense is generally not of the "break" variety, for the most part.

Looking Points per 100 yards, the first 3 games were quintessentially "Bend don't Break", with less than 4 points / 100 yards. This likely means a lot of drives go on for a while but fizzle out. The defense was still bend-y but not too break-y against Clemson and Miami, especially adjusted for each team's offensive efficiency.

Again, we see a trend for the worse in the last two games. The defense basically didn't slow down Pitt or GSU. Yeah, we forced GSU into long possessions for them to score (low YPP), but they still scored (PPP and per 100 yds). That's Bend AND Break. Against Pitt the defense was basically abysmal.

Verdict: Ted Roof's defense has been bending AND breaking over the last two games, not slowing the opponents' scoring. The first 5 games were more or less "Bend don't Break".

Bonus: Since I already have the data, here is the flip side looking at our offense per-game compared to the opponents' season average. No write up here, but ya'll are smart, you can figure it out.

I'd say that other than a slow start against BC and the disaster against Clemson, the offense has been fine in moving the ball and scoring points. That said, hidden in these stats are the two scoop-n-scores that were fumbled away against Miami. That's a big demerit.

To summarize:

• BC: W, Offense got away with one
• Mercer: W, Both units good
• Vandy: W, Both units good
• Clemson: L, Offense cost the game
• Miami: L, Offense cost the game
• Pitt: L, Defense cost the game
• Ga So: W, Defense got away with one