Looking at the numbers, 2021 preseason: Big Plays

That’s right, talkin’ Big Plays today. We’re living large and we’re in charge.

For anyone just arriving to this website (welcome!), here is a quick list of the posts that have preceded this one over the past couple of weeks:

I suppose I should explain what a Big Play is for the purposes of this post. While there are a lot of different statistical methods to group long-gaining plays from scrimmage, I’m defining them here as I have for several years now: any scrimmage play, running or passing, that gains 20 yards or more. Note that this does not include defensive/special teams returns.

Also, even if there is a turnover at the end of the play in question, if the net gain (prior to the possession change) is at least 20 yards, that still counts. You could make a decent argument that it shouldn’t, but I’ve elected to include such plays in the totals for the sake of consistency. (For the record, I only found two such events in the entirety of the SoCon spring slate.)

The Citadel, 2021 Spring Football

The Bulldogs had a negative overall margin as far as big plays are concerned (22 for, 40 against). That was the largest per-game deficiency in the league (-2.25 per contest). 

Here is the breakdown of offensive big plays in conference action (spring games only):

  • Samford: 48 (7 games)
  • Mercer: 39 (8 games)
  • VMI: 33 (7 games)
  • Furman: 30 (7 games)
  • Chattanooga: 24 (4 games)
  • The Citadel: 22 (8 games)
  • East Tennessee State: 21 (6 games)
  • Wofford: 19 (5 games)
  • Western Carolina: 19 (6 games)

Samford had the most plays from scrimmage of 20+ yards, while Chattanooga averaged six big plays per game, which was the second-best rate in the conference. The Citadel’s rate of 2.75 big plays per contest ranked last.

How do those numbers compare nationally? Well, as I mentioned in my previous post, trying to compare FCS stats for F20/S21 is largely a waste of time, for a host of obvious reasons, and even if you wanted to do so, compiling all of those statistics would be very difficult, if not impossible. (I can attest that assembling these types of stats for just the SoCon teams isn’t easy.)

On the other hand, FBS statistics are more manageable when it comes to organizing such numbers. Here is a sampling from that subdivision, in which 127 teams participated in 2020. On a per-game basis, the top 5 in big plays (20+ yards or more) were UCF (8.30 per contest), Ohio State, North Carolina, Florida, and Mississippi.

No one reading this who follows college football is shocked at all by that list. Other teams that fared well in this area included BYU, Clemson, Alabama, and Oklahoma.

The bottom five: Massachusetts (last in FBS at 2.0 per game, with the Minutemen only playing four contests in all), Western Kentucky, FIU, Kansas, and Syracuse. Again, nobody is surprised.

Incidentally, The Citadel’s 2.75 big plays per game exactly matched the average for Army, which had 33 in 12 games. The Black Knights ranked 121st overall in FBS.

Another, arguably more substantive way to look at this is on a per-play basis, rather than per-game. In terms of big plays per offensive snap (lower is better, of course):

  • Chattanooga: 10.50 snaps per big play
  • Samford: 11.40
  • Mercer: 14.13
  • Furman: 16.30
  • VMI: 16.55
  • Wofford: 17.05
  • Western Carolina: 17.53
  • East Tennessee State: 18.71
  • The Citadel: 26.09

Let’s compare those numbers to FBS programs.

The list of top schools is similar to that of the per-game grouping. This time, Ohio State led all squads, with a rate of just 9.50 snaps per big play. BYU was second (9.63).

A somewhat surprising team in this category is Maryland, which finished fifth (10.00). The Terrapins only played five games, but there is no doubt that when Mike Locksley’s offense is rolling, it can produce a lot of long gainers.

The bottom five on a per-play basis matched the bottom five on a per-game rate, except that Wisconsin (!) replaced Syracuse, with the Badgers only recording a 20+ yard play every 26.05 snaps from scrimmage. That was still better than Western Kentucky, which finished last (29.15).

The Hilltoppers decided to do something about it, though. WKU brought in Houston Baptist’s starting quarterback as a transfer, along with his top three receiving targets.

Houston Baptist was one of the FCS teams that played a few games in the fall, but did not compete in the spring. HBU faced three FBS squads, and while losing all three of those games, the Huskies put up 31, 33, and 38 points. In one of those games, against Texas Tech, star quarterback Bailey Zappe had 560 yards of total offense.

Now, Zappe and several of his friends are taking their talents to Bowling Green, Kentucky. It’s a brave new world.

One of the things that I wondered about as I looked over these numbers: does the triple option offense tend to inhibit big play production? By and large, the FBS triple option teams did not fare well in this category. I’ve included them in this list of “other teams of interest” in the big plays per-snap statistic, along with all teams facing SoCon opposition this fall, plus a random team or two:

  • Florida: 9.66 big plays per snap (3rd in FBS)
  • North Carolina: 9.76 (4th)
  • Alabama: 10.37 (8th)
  • Oklahoma: 10.85 (12th)
  • Clemson: 11.12 (14th)
  • North Carolina State: 11.75 (19th)
  • Coastal Carolina: 12.58 (29th)
  • Kent State: 13.04 (36th)
  • South Carolina: 15.09 (64th)
  • Georgia Southern: 15.96 (73rd)
  • East Carolina: 16.38 (79th)
  • Navy: 17.34 (93rd)
  • Air Force: 17.76 (98th)
  • Kentucky: 17.87 (100th)
  • Vanderbilt: 23.22 (120th)
  • Army: 25.50 (122nd)

Of course, one reason why triple option teams don’t have as many longer plays from scrimmage is because they are primarily (and sometimes almost exclusively) running teams, and the majority of 20+ yard plays are via the air — for example, last season in FBS, 68.2% of such plays were passes.

That percentage was almost reversed in the triple option universe. Last year, Georgia Southern, Navy, Army, and Air Force combined for 140 plays of 20+ yards, with 61.4% of them coming on the ground. There was room for outliers — Navy had almost as many big passing plays (17) as it did rushing (18) — but when you run the ball 83.1% of the time (the combined percentage for the four-team tandem), that is the end result.

What it means, though, is that those triple option teams still had a higher percentage of pass attempts result in big plays than their rushes. This wasn’t true for The Citadel, which actually had fewer snaps per rushing big plays (25.89) than pass plays (27.33). Intuitively, that needs to change this fall for the Bulldogs’ offense to be more successful.

You might have noticed that I referenced both rushing and passing big plays in those last few paragraphs. I did break down big plays in SoCon action by run and pass; while I’m not going to list them out separately in this post, they can be found on this handy-dandy spreadsheet:

Big Plays, SoCon spring 2021

Not only does it list big plays by run and pass, it is color-coded! And not completely impossible to understand!

Okay, now to talk about defending big plays.

Big plays allowed in SoCon action, spring 2021:

  • Chattanooga: 7 (4 games)
  • East Tennessee State: 17 (6 games)
  • Western Carolina: 23 (6 games)
  • Wofford: 28 (5 games)
  • VMI: 29 (7 games)
  • Furman: 31 (7 games)
  • The Citadel: 40 (8 games)
  • Mercer: 40 (8 games)
  • Samford: 40 (7 games)

Chattanooga was by some distance the best league team at preventing big plays, albeit in a smaller sample size. Samford allowed 5.71 per game, the most on average, with Wofford just behind at 5.60 per contest.

Air Force allowed the fewest big plays in FBS on a per-game basis, at 1.67 (the Falcons played six contests in 2020). Marshall, Iowa, Louisiana-Lafayette, and BYU rounded out the top five; influential backers of the Thundering Herd were so impressed that head coach Doc Holliday was fired.

The worst team in FBS in this category was LSU, which gave up 79 such plays in 10 games (7.90 average). Did the Tigers do any better when considering big plays allowed on a per-play basis? Uh, no. LSU was last by that metric as well, as it allowed a big play every 8.58 snaps.

In term of pass defense specifically, the Tigers allowed the most plays of 20 yards or more…and the most of 30 yards or more…and the most of 40 yards or more…and the most of 50 yards or more. LSU was also one of seven FBS schools to give up a pass play of more than 90 yards.

Former defensive coordinator Bo Pelini’s three-year contract was guaranteed, and he had two years remaining on his deal at $2.3 million per annum when he was fired last December. LSU had to pony up $4 million in severance pay by January 31, 2021. What a country.

Here is that list of FBS teams from earlier, this time ranked on defending big plays on a per-snap basis (to avoid confusion, just remember that for defenses, a larger number is better when it comes to this statistic):

  • Air Force: 36.00 snaps per big play allowed (1st in FBS)
  • Coastal Carolina: 18.33 (20th)
  • Alabama: 17.48 (26th)
  • Kentucky: 17.36 (28th)
  • Clemson: 16.98 (34th)
  • Florida: 15.98 (50th)
  • Navy: 15.71 (56th)
  • North Carolina: 15.64 (58th)
  • North Carolina State: 15.15 (62nd)
  • Army: 15.02 (66th)
  • Georgia Southern: 14.79 (68th)
  • Oklahoma: 14.57 (70th)
  • South Carolina: 12.74 (95th)
  • East Carolina: 12.71 (96th)
  • Kent State: 12.43 (104th)
  • Vanderbilt: 10.22 (123rd)

The spring 2021 SoCon list of big plays allowed, per snap:

  • Chattanooga, 36.14
  • East Tennessee State, 23.59
  • Western Carolina, 22.00
  • VMI, 16.93
  • Samford, 14.30
  • Furman, 14.19
  • Mercer, 13.83
  • Wofford, 11.68
  • The Citadel, 11.65

Clearly, the Bulldogs have a lot of room for improvement when it comes to both producing and preventing big plays. I went back and looked at the numbers for the 2016 season, just to compare:

  • 2016 offense — 26 big plays (15 runs, 11 passes), 22.19 snaps per big play
  • 2016 defense — 28 big plays allowed (9 runs, 19 passes), 16.46 snaps per big play

I think it is interesting that the 2016 championship team did not actually have an exceptionally large number of offensive big plays. It must be pointed out, though, that its big plays were often BIG.

Eight of the 26 long gainers in 2016 were over 50 yards in length, including three that were 70+. In contrast, the Bulldogs had just two plays from scrimmage this spring that went for 50 yards or more.

Defensively, 16 snaps per big play allowed would seem like a respectable target to try to reach for the Bulldogs this fall. That would be about 1½ fewer such plays per game. In the spring, The Citadel had all kinds of defensive problems on the opening possession; eliminating most of those issues would be a very good start (pun intended).

More statistical reviews and thoughts to come, as the season creeps even closer…

Looking at the numbers, 2021 preseason: Havoc Rate

This is the first in an occasional preseason series (at least, I hope it is occasional) in which I take a closer look at a few of The Citadel’s spring football 2021 statistics.

What improvements can the Bulldogs make? What are their most significant deficiencies from a statistical perspective? What are their strengths? How do they compare to other SoCon teams in various categories?

I’ll also highlight various FBS teams as points of comparison, in part because for some of these statistics, FCS numbers are very hard to come by. Besides, let’s face it — the F20/S21 season for FCS on a national level was a slow-motion trainwreck anyway.

In a previous post, I introduced the spreadsheet from which I’ll be working. Here it is again:

The Citadel, 2021 Spring Football

What is havoc rate? Well, it is a statistic that was essentially created by Bill Connelly (now of ESPN) in 2015. In recent years, it has gained a lot of credence, particularly in the coaching community, as this 2019 article about Georgia football suggests:

On the first day of spring practice, Georgia coach Kirby Smart said he wanted to improve the team’s havoc rate. The term has been tossed around for months now by players and coaches in Athens…

Havoc rate comes from the total number of tackles for loss, passes defensed (interceptions and breakups) and forced fumbles divided by the total number of plays.

“We feel that we should have 20 percent of the plays, two of every 10 should be a ball disruption, a turnover, a PBU, a tackle for a loss, so we’re charting that,” Smart said this spring…

…Smart said Georgia studied the top 10 teams in havoc rate….“We’re trying to do some of the things they do and we’re trying to put guys in position to do that,” Smart said…

…Safety J.R. Reed said the first of two preseason scrimmages produced the most forced turnovers in the last two years.

That came after a spring in which a player a day was asked to stand up and give a definition of havoc in the defensive team meeting room.

“Everybody in that room, from the highest SAT/ACT to the lowest has got to stand up and give us what havoc rate is,” Smart said. “If they understand what it is, they know we’re trying to cause it.”

Basically, this stat is about how often a defense creates disruptive and/or negative plays. The national average for havoc rate tends to be around 16%, while the top teams in the category will exceed 20%.

In eight league games this spring, The Citadel had a defensive havoc rate of 14.38%. How did that compare in the SoCon? I’m glad you asked. Here is each conference team’s rate (league games only):

  • East Tennessee State: 17.21%
  • Chattanooga: 16.60%
  • Furman: 15.91%
  • VMI: 15.89%
  • Mercer: 15.01%
  • Samford: 14.86%
  • The Citadel: 14.38%
  • Wofford: 11.93%
  • Western Carolina: 8.89%

Keep in mind the disparity in games played this spring. Chattanooga only played four contests, while Wofford suited up for five. Western Carolina and ETSU played six; Furman, VMI, and Samford played seven; and The Citadel and Mercer each completed the full slate of conference matchups, with eight.

UTC’s defensive havoc rate would have been 19.67% if you took out the results from its game versus Mercer, when the Mocs fielded what amounted to a “B” team. On the other hand, that number would have been for just three games anyway; we’re talking about a lot of statistical variance in this instance.

Last season, Pittsburgh led all FBS teams in defensive havoc rate at 22.75%, just ahead of Clemson. Other squads with rates greater than 20%: Utah, San Diego State, Notre Dame, and TCU. Colorado and Oklahoma just missed hitting that mark.

One thing I’ll try to do in this series is list (when applicable) the category statistics for other teams of interest in FBS, with a particular focus on those which will be facing SoCon opposition this fall. Each league team will play one FBS foe in 2021.

I’ve already mentioned Oklahoma, the FBS opponent for Western Carolina this year (good luck, Kerwin Bell). The Sooners had a defensive havoc rate of 19.78%, which was 8th-best in FBS. Others of note:

  • Army, 18.76%, 19th nationally (9-3 last season, with two wins over SoCon squads)
  • Alabama, 18.59%, 20th (SoCon opponent in 2021: Mercer)
  • North Carolina State, 17.71%, 32nd (SoCon opponent in 2021: Furman)
  • Coastal Carolina, 17.53%, 36th (SoCon opponent in 2021: The Citadel)
  • Kent State, 17.24%, 43rd (SoCon opponent in 2021: VMI)
  • Florida, 15.58%, 71st (SoCon opponent in 2021: Samford)
  • North Carolina, 15.08%, 79th (SoCon opponent in 2021: Wofford)
  • Kentucky, 13.35%, 102nd (SoCon opponent in 2021: Chattanooga)
  • Navy, 12.89%, 108th (3-7 last season)
  • Vanderbilt, 11.44%, 118th (SoCon opponent in 2021: ETSU)
  • South Carolina, 10.90%, 125th (2-8 last season)
  • Air Force, 9.44%, 126th (3-3 last season)
  • Akron, 7.95%, 127th and last (1-5 last season)

From The Citadel’s perspective, I think a reasonable goal in 2021 would be to increase its defensive havoc rate to at least 16%. That might not sound like a major step forward from 14.38%, but if the Bulldogs were to have a DHR of 16%, it would be an increase of almost exactly one more disruptive/negative play per game.

That one play — a forced fumble, a big tackle for loss, an interception — could well be the difference between a win or a loss. After all, just think about how many close games The Citadel has played in the conference in recent years.

One specific area of potential improvement for The Citadel in this regard could be sack rate. The Bulldogs had a defensive sack rate of 4.17%, which if applied to FBS statistics would have ranked in the bottom 15% nationally (with the same percentage as that of Michigan).

It is also possible to calculate Havoc Rate Against (as I call it), or how disrupted offenses are by negative plays. I don’t have FBS numbers for this (finding forced fumble and TFL stats for offenses can be difficult), but I did put together a chart for SoCon spring play, similar to the one for the defenses. Here is the HRA breakdown for league games in 2021 (remember, the lower the percentage in this category, the better):

  • VMI: 11.17%
  • The Citadel: 11.67%
  • Chattanooga: 14.29%
  • Western Carolina: 14.71%
  • Wofford: 15.12%
  • Samford 15.54%
  • Mercer: 15.79%
  • ETSU: 16.28%
  • Furman: 16.36%

While the Bulldogs are second in this grouping, I tend to believe that 11.67% is not necessarily an outstanding outcome, given the nature of the triple option attack.

For example, a tackle for loss should be an unusual outcome for a play in The Citadel’s offense. However, the Bulldogs suffered a larger-than-expected number of tackles for loss in spring 2021, almost entirely due to an abysmal sack percentage against (17.07%).

This can be attributed in large part to the fact that most of The Citadel’s pass plays occurred when the Bulldogs were trailing (64.3% of the sacks came in the 4th quarter).

Another team that had issues with negative plays on offense was Furman. If not for a completely dominant performance in its opener versus Western Carolina, FU’s havoc rate against would have been even higher; if you take out the Paladins’ numbers against WCU, their HRA jumps to 18.84%.

Of course, that game still counted. Removing it from the remaining six games Furman played tends to unfairly skew things in the opposite direction.

Conversely, 11.17% is surely a very impressive result for VMI’s pass-happy offense. 

Incidentally, the league average for havoc rate (which obviously applies both defensively and offensively) was 14.42%. The median was slightly above 15% on both sides of the ball.

In my next post, I’ll discuss another statistical category, one that can dovetail with havoc rate.

Sometimes they are referred to as long gainers, but here at The Sports Arsenal, we call them BIG PLAYS.

You’ve been warned…