Special thanks to the Purple Antics Podcast for joining me to discuss James Madison’s roster on this week’s podcast:

James Madison completed a 12-1 season by defeating Troy in the championship game to win the Sun Belt conference in Coach Chesney’s second year, after taking over for Curt Cignetti when the latter left for Indiana at the end of the 2023 season. Chesney had been a nearly 20-year veteran of Division-II/III football when he took over FCS Holy Cross in 2018 and led the Crusaders to their most successful six-year span since the glory years in the 1980s under Rick Carter and Mark Duffner. UCLA hired Chesney two weeks ago to lead the Bruins for the 2026 season, but Chesney and his staff have pledged to stay with the James Madison throughout their playoff run.

The Dukes were undefeated in conference play and 3-1 in their non-conference slate, beating FCS Weber State, in-state rival Liberty, and a wandering Washington State (incidentally, the Cougs played five teams from south of the Mason-Dixon line this season, four on the road). JMU’s only loss was to Louisville from the ACC, though the Dukes led at halftime and the Cardinals only pulled ahead late on a sack-fumble in the endzone in the 4^th quarter.

I acquired and charted the film on all 13 games of JMU’s season to date, though as standard practice the statistical production from garbage time and the FCS game are excluded from the analysis in this article. There are two noteworthy factors in evaluating the sweep of the Dukes’ 2025 season for the reader to keep in mind.

First is that unlike many teams which play a diverse range of opponent schemes and qualities, JMU’s slate featured almost entirely copycats — seven of their conference games plus one of their non-cons, Liberty — with extraordinarily similar playstyles and advanced statistical rankings on both sides of the ball. This background of routine was then punctuated by a couple of opponents who played quite differently and/or turned in dramatically higher marks.

This chart gives the F+ advanced statistics ranks for each of JMU’s FBS opponents at the end of the regular season – overall ranking, offense, and defense:

The second factor is how certain key personnel changes affected JMU’s production. The Dukes played relatively stable football in 2025, and while almost every position unit had at least one notable change-up at some point during the season due to injury or the staff making a midseason choice, most didn’t produce any measurable effect on performance or output in the statistical regression once opponent quality was properly controlled. But there are three major exceptions where there’s a clear change in the data correlated with the player’s absence from the lineup.

Those are: a) after the week 2 game at Louisville, JMU stopped using their two-QB system and stuck with one starting QB only, relegating the other QB to an occasional backup; b) midway through the week 12 game vs Appalachian St, their tallest (and I will argue, most “clutch”) wide receiver, #13 WR Ellis, was injured and has been absent or effectively so ever since; and c) by far their biggest defensive tackle (by perhaps 50 lbs), #0 DT I. Bush, missed time at a few different points, most recently the second half of the week 14 game at Coastal Carolina and the entire conference championship against Troy.

While the QB change has been for the better, being without the effective use of some of the highest leverage players on each side of the ball (actually, Ellis has been on the field, but only as a decoy … in fact on the rare occasion when he was thrown the ball a Troy defender stripped it from him for a turnover) at the same time in my opinion made the conference championship game much closer than it otherwise might have been. From the statistical profile, Troy is barely distinguishable from the rest of the Sun Belt teams that JMU mopped up when they were healthy, and the Trojans themselves experienced bad injury luck themselves when their starting QB got hurt in the second half.

In addition to making the Louisville and Troy games less helpful as single data points for understanding JMU, the significant upshot is that the Dukes have the potential to improve significantly if they get one or both of those key players back to full strength on Saturday. On the podcast we discussed reports that both will be suited up and playing, though whether they’re at 100% remains to be seen.

Offense

Everything about James Madison’s offense is centered on their run game – designed QB runs figure prominently in the offense (and were an important part of the two-QB system backstory, though the full-time starter has since assumed the entire rushing playbook), the Dukes run 55% of the time overall, on 60% of 1^st downs, and 75% of 2^nd & shorts, 2^nd & mediums, and 3^rd & shorts. They even run on half of 2^nd & longs, higher than the FBS average, despite passing being significantly more efficient for them in that situation. It’s not until 3^rd & medium or longer that they flip over and become predominantly a passing team.

JMU’s desire to run the ball, how good they are at it in short yardage (72% success rate on 2^nd & short, 87% on 3^rd & short), and how ineffective they are once they fall behind the chains (40% success rate on 2^nd & long, 24% on 3^rd & long) paints a very familiar picture to college football fans acquainted with run-first offenses. It seems that a lot of opponents took the typical advice for playing teams which looked like JMU to heart: pull out all the stops to stuff the run, because the offense will be helpless at playing from behind.

But doing so was a huge mistake, because JMU is not a typical run-first team. In fact they’re almost the exact opposite of one – unless they get the ball on a short field from their defense or special teams, virtually all of their scoring has come from highly explosive plays, and they’ve been almost totally incapable of sustaining methodical scoring drives. We had a podcast discussion about OC Kennedy’s playcalling style in this regard – they gave him a lot of credit for trapping and punishing defenses like this by design, but if defenses call the bluff it can leave Kennedy “frazzled” to the frustration of the fanbase.

Of the 118 drives I charted outside garbage time, I excluded five because they were backed up against their own goalposts (these are distortive) and 26 because they were short fields. Of the 87 meaningful full-field drives, 32 scored touchdowns … and virtually all of them, 29, resulted from 40+ yard plays, with more than half of those TD drives lasting five plays or fewer. No meaningful full-field drive with a super-explosive play resulted in a turnover or empty possession, and only one resulted in a field goal.

However, if JMU did not get a super-explosive play, their drive outcomes were below FBS averages in all categories. Only three of these 57 methodical possessions resulted in touchdowns and two in field goals. Eleven resulted in turnovers, while the remaining 41 (71.9%, almost two standard deviations more than median) were empty possessions. These also tended to be over quickly: 32 of the Dukes’ 52 non-scoring possessions (61.6%, also about two standard deviations greater than median) ended in five plays or fewer.

The profound irony is that the single biggest correlation in JMU generating super-explosive plays was defenses taking safeties and linebackers from depth and crowding the line of scrimmage to stuff the run and take away an efficiency conversion – in other words, by trying to stop something that had little chance of hurting them, those defenses almost guaranteed the one thing that would! Here are examples from each of the last seven weeks:

The starter who won the job is #14 QB Barnett, who was the starter in 2024 as well though he missed the bowl game that year with an injury. Chesney had brought in #9 QB Sluka, who had played at Holy Cross before transferring to UNLV then transferring out after some contratemps, and in the first two weeks of the 2025 season — against FCS Weber State and in the loss to Louisville — they would trade off snaps constantly, with a very specific power running playbook for Sluka while Barnett was the passer and zone-read runner. The Dukes had a bye in week 3 and the staff ditched this system in their time off, with Barnett playing full-time and Sluka only coming in as occasional relief.

The most significant change, other than the usual admonition about having two quarterbacks, is that Barnett absorbed Sluka’s QB power playbook, and although it fluctuated for the next four weeks, by week 8 and for the rest of the season Barnett was getting about 20% of meaningful carries on designed runs. Barnett is essentially tied for the second most frequently used ballcarrier in the run game, and on a per-carry basis (excluding QB sacks/scrambles on called passing plays) the most effective with a 60.8% success rate given the down & distance and averaging 6.4 adjusted YPC.

The leading running back by a wide margin is #3 RB Knight, with 160 meaningful carries on the season. His yardage average is excellent, 6.2 adjusted, though his success rate is just a few points above average at 54.3%. The second most used back has been #23 RB Fuller at 72 carries, with a 57% success rate but a much lower 4.6 YPC, though he’s been out for the last three games. #28 RB Malary has pretty similar per-carry numbers to Fuller at 57% and 4.9 YPC despite being built more like a goalline specialist, but the redshirt senior has only played in the last three games, clearly as a response to Fuller’s injury, and has jumped over Pettaway in the order.

Parsing the other three backs is tricky. They’ve had #6 RB Pettaway all season long, though his carry count of two or three each game hasn’t gone up or down – his numbers are very poor at 35.3% success and 3.9 YPC. #5 RB Adeyi had fantastic numbers in a single game against Liberty, then only got a couple carries in one other week and otherwise I haven’t seen him at all – reportedly he’s taking a redshirt this year, but theoretically he could play in the postseason. When I pressed on the podcast for why Adeyi hasn’t taken over for Pettaway in the championship game, the response sounded to me like the staff was playing Pettaway, a senior last year’s starter, his respects out of sentimentality rather than bench him out of brute expediency.

The impediments to JMU’s methodical run game are twofold. First, they can only give 20% of carries to their most effective runner while a third of carries are going to guys not named Barnett or Knight whose numbers are pretty mediocre – 54% success and 4.5 YPC cumulatively for the backups. Second, all these numbers are juiced by the super-explosives and hide the below-average rate of chunk yardage rushing (10+ yard runs) which are necessary to keep methodical drives going.

Narrowing down to runs of 25 yards or fewer (which is to say, ignoring only the 14 longest runs across the entire team during the whole season) doesn’t budge the success rates at all, but it brings down the global YPC by -1.26 to 4.50 and the explosive rush rate by five percentage points to 9.3%, while individually it brings Barnett and Knight’s YPC down more than two full yards and everyone else at least half a yard.

This is again to say – in the efficiency run game, the Dukes’ bark is worse than their bite. The biggest mistake a defense can make is overreact to it by allocating their resources to stop a relatively modest threat up front, and leave the back end exposed where the real damage has been done on the plays they break through.

Here’s a representative sample of successful designed runs:

(Reminder – you can use the button in the lower right corner to control playback speed)

1. :00 – This midline read of the backer on counter action is the most common outside run, the basic play of Barnett’s playbook before he absorbed Sluka’s (if he keeps he follows the pullers, in QB power the RB is the lead blocker), and JMU’s most likely play to generate a 5-7 yard gain. Knight already has the acceleration and cut to get around the end and the read had frozen the backer, he gets more this time because of a poor angle by the safety.
2. :08 – Here’s the inside read version, not a great hit by the the slicing H-back but more than enough for Fuller to run through since the end is frozen on Barnett and the tackles are slanting the wrong way, so the left side of the line climbs easily to the backers and. Note in particular how much the LT is hammering the backer with clean contact.
3. :16 – This is late in the season with Fuller out and Malary stepping in for him, in his element as a short-yardage back running over people. The blocking is an absolute mess as the entire left side gets cleaned out (the LT is too high and the LG and C get ploughed under, preventing the backside pull), but Malary keeps his head up and maintains his balance rolling over bodies for the endline.
4. :31 – The defense falls for the zone read, incredibly, and Barnett’s keep is without hesitation. The WRs used extensively as run game blockers and the LT’s uncalled takedown are common.

And unsuccessful runs:

1. :00 – Here it’s the H-back as lead blocker, the RB motioning to empty just before the snap is a dead giveaway. It’s 1^st & 15 here and the offense has no numbers advantage, no box eliminator, and no punisher for the defense playing cover-0. The end simply wrong-arms the pulling RG, the backer takes on the H-back, and the unoccupied safety ends the play.
2. :06 – Three-down fronts and lighter boxes had statistically no dip in rush defense performance in the regression, controlling for field position effects. Keeping personnel back gave more time to read the play, which could usually be inferred from the formation anyway. Here the defense correctly flows from the tailback and H-back’s alignment to the boundary and overwhelm the offense before the blocks can set up. Note the backside control on the QB from the safety while they still have three over two to the field.
3. :13 – This is Pettaway, still getting his weekly couple of carries in week 14. He’s meant to go to the right A-gap, and the LT’s zone assignment is to get inside leverage and turn the backer away, but his feet are too lazy on the play and he lunges late. So now the LT rides the backer all the way into the line, and Pettaway decides to cut opposite of the play design to try and take advantage, but the MIKE is right there since the center now can’t get to him (I’m not sure he would have gotten off the combo with the LG and up to him if this went down properly but at least his momentum was taking him the right way), and Pettaway can’t escape the tackle.
4. :25 – Setting their tackling issues aside, Troy had JMU’s read-option game solved: 3-4 against 12-pers with the overhang backer/STUD controlling the QB keep on the backside, and flowing to the obvious run strength of the formation. Note the footwork on the left side, there’s some real comedy.

The starters at the center and tackle spots have been set all year and with the exception of some minor injury-related absences we’ve seen them in every meaningful snap: #79 LT Simmons who was a Cignetti recruit and #76 RT McMurtrie who came over with Chesney from Holy Cross last year, and #60 C Greenburg who started out at the FCS level then went to North Carolina before transferring to JMU this year.

Simmons is the biggest of the linemen by far and uses his size liberally, there’s the most oscillation on my tally sheet with him since he has the raw power to do the most but technically there’s a lot of wild play. McMurtrie is almost completely the other way around, I like his technical refinement but he could really stand to add about 10 lbs of muscle and get stronger through his base. I haven’t seen anyone play backup center in meaningful time (and no one on the podcast knew who it’d be either) but the tackles have been spelled in a few moments by one of the guards or backup underclassmen #58 LT Hart and #71 RT Rawls.

I’ve seen four guys rotate through at the two guard spots. I think at first it might have been injury related but a few weeks into the season I think the staff decided they just liked having all of them get playing time. By the end of the year the two I’ve seen the most are junior #52 OG Sweazie, mostly on the left, and sophomore #55 OG Robell, mostly on the right, followed by true freshman #68 OG Wilson who’s rotated at both spots. Redshirt senior #72 OG Toner began the year as the starting right guard but I believe he got hurt, and when he worked his way back in they had gone to the rotation system, and at this point I’ve seen him the least.

The tight end on the field the most is sophomore #89 TE Phifer, used almost exclusively as a blocker with fairly high marks on my tally sheet. There’s something of a giveaway when they switch to #15 TE Dippre or #18 TE Kyle in 11-pers, or bring one or both of them in for 12-pers, because their blocking grades are much, much lower but they’re more extensively used as passing targets. Dippre is the more effective of them, in fact his are some of the highest per-target receiving figures on the team at 66.5% success and 8.6 adjusted YPC with the size and vertical to haul in some of Barnett’s high passes and bail the team out of some dicey situations (Kyle’s numbers are more average, he doesn’t look to have quite the same ability on tape to me).

There are four main wideouts in the lineup, all of whom transferred in this year from an FCS school, or Maryland, or in one case both. In my opinion, Ellis — whom I mentioned above as being hurt midway through week 12 — has been far and away the most valuable, and I think it’s telling that he still leads the unit in meaningful targets despite not really playing for the last three and a half games. Ellis has a 65.2% per-target success rate and 9.8 adjusted YPT which are elite numbers, and he has the size and physicality (when healthy) to go up and get balls that I don’t think is replicated in the WR corps.

Furthermore, it was clear from the frequency distribution analysis that the staff knew it, because Ellis’ targets spiked in “clutch” situations like 3^rd & long and redzone scoring opportunities (I actually think Dippre is a good candidate for this as well, but there’s no such spike in his data … perhaps the staff needs more analysts). In the last couple weeks of the season, the staff put Ellis on the field, but he wasn’t really a viable target, more of a decoy to occupy defenses who was rarely if ever targeted. On the podcast we talked about an interview Ellis recently gave in which he said he planned on playing during the playoffs, though it’s uncertain whether this means at full strength, as the same just-a-decoy role as before, or something in between.

At the other outside spot is #11 WR DiGennaro, with a 51.3% success rate and 8.6 adjusted YPT. In the slot is #0 WR Sanchez, with 48.7% success and 7.9 YPT. The swing man at 6’0” is #10 WR Wisloski at 44.8% success and 8.8 YPT, he’s gotten the fewest targets but had a big game in week 7 and then gotten more playing time lately with Ellis’ injury. All three of them are fast, dangerous receivers with multiple super-explosive plays pulling up their yardage averages, but their mediocre-to-poor success rates speak to a combination of their own catch radii and Barnett’s tendency to miss high – the QB has a strong arm but his refinement is such that he really needs his receiver to be wide open, or to be big enough to physically win a contested catch.

If I might speculate, I think all three of DiGennaro, Sanchez, and Wisloski’s numbers would skyrocket on a Mike Leach-style Air Raid where an accurate QB was zipping them quick short passes before the DB can get there and letting them run after the catch (there was some irony when I suggested this on the podcast – years ago when JMU was an FCS school they were in fact an Air Raid team). The Dukes’ current system doesn’t quite suit their strengths but if they get the ball in space they can certainly generate big yards.

Another issue in the passing game is the truly massive volume of targets going to the running back Knight … in fact during the championship game he edged out Ellis as the team’s overall leading target. About a third of these are last-ditch passes like checkdowns and dumpoffs, another 15% are short designed downfield passes like a hitch out of the backfield or a wheel to the boundary, but most are screens, usually the same type of smoke screen.

While these are a significant contributor to the “crowded line, super-explosive” phenomenon I’ve discussed, the vast majority gain minimal yardage at best or lose yardage at worst, and are a major drain on the passing game’s efficiency. Knight’s per-target success rate is underwater at 48% with 6.4 adjusted YPT, which for a team’s leading pass target are simply not figures that are conducive to sustaining drives.

Here’s a representative sample of successful passes:

1. :00 – Here’s Ellis, just a complete package of acceleration, smooth route running with timing, vertical, hands, and body control in the back of the endzone. Nobody else on the team — or any of the teams I watched during this project — could do this.
2. :14 – Dippre on the scramble drill. This throw is … inadvisable, and Dippre makes a heroic play as well as securing the catch against some hard hits.
3. :33 – This is a typical depth of target for DeGennaro, and also typically there’s not a ton of separation and Barnett places the ball kind of awkwardly on the receiver, which is why he’s at pretty close to a 50/50 success rate. But when he snags it, it’s a pretty nice play.
4. :51 – This time it’s Wisloski in the slot, blowing past the woolgathering nickel. The defense is sending everybody at Barnett and he’s late with the ball while he’s figuring out his footwork so it’s underthrown. Wisloski has to fight off the nickel he already beat once but there’s no safety help and his body control is impressive.

And unsuccessful passing plays:

1. :00 – Louisville recorded six sacks in this game and five hits on the QB as he was releasing, and unfortunately the offensive line had trouble avoiding injuries with far more swapouts this week than any other, so it was a challenge to find clips with the starters. Here McMurtrie is just getting worked solo by the speed rusher on the right, meanwhile the left is mishandling a three-man stunt – Sweazie is ready to accept the handoff of the nose from Greenburg so the center can take the looper, but Greenburg has turned his body and isn’t getting any help from Knight or Robell. Simmons doesn’t have his base set and the stunter flips around him. Facing everybody getting through, Barnett starts to run, and the center just plays through Sweazie for the sack.
2. :30 – While Barnett has the physical traits of a great athlete in the designed run game, from watching tape I don’t see the instincts of a scrambler when called passing plays break down. The play here is a clockwise spin out to the field, not a leg to the boundary. I don’t really understand how this protection was supposed to work.
3. :47 – This diamond formation is always a box check – if they have numbers it’s a QB run, if not it’s a screen. For almost every team forcing the run is the smart defensive play but with JMU it’s actually the other way around, Barnett is a very reliable runner and the perimeter blocking is … less so.
4. :53 – It’s plays like these that have made watching JMU’s offense feel a little upside down – while they can manipulate defenses into giving up big bombs with relative ease, they’re constantly struggling with efficiency plays like these quick hitches which many other teams feature because their success rates are near automatic, if for modest yardage. Here Barnett’s throw is hurried and off target for no real reason, Sanchez tries to catch it outside the frame of his body but can’t bring it in.

Defense

JMU employs a 4-2-5 on most downs, pulling the nickel and switching to a 4-3 against some heavy offensive configurations but still tending to stay in the 5-DB look against about a third of 12-personnel offensive sets.

The three main things which set the Dukes’ defense apart compared to almost every other Sun Belt defense — as well as Liberty and Wazzu — that I watched their offense play were: first, they have a full complement of effective and reliable linemen so they could freely rotate through four tackles and four ends without fatigue or second-line dropoff problems as well as field a racecar package on 3^rd downs; second, aggressive play immdiately attacking the run from at least one linebacker or safety on virtually every snap except very obvious passing situations; and third, the confidence to play man coverage with little in the way of safety help (usually a single high safety and he wasn’t even that high, averaging 10.5 yards deep at the snap and keeping his eyes in the backfield).

The combination of this approach, JMU’s talent, and the offensive style that they predominantly faced resulted in a very successful season for the Dukes’ defense, and I would say that even more than their offense, the defense was responsible for their 12-1 season. According to the charting metrics I use, two of the three biggest threats they faced during the year — on either side of the ball — were the offenses of Texas State and Old Dominion, and I would argue that the multiple-interception performances from JMU’s defense in each of those games saved the day (Louisville’s defense was the third, Wazzu’s defense would have been a bigger threat but at the point in the season they played the Cougs were down several key personnel).

The essential thing to JMU’s approach is winning 1^st downs, which they do at a championship caliber regardless of play selection – 59.8% defensive success rate against the run and 62.2% against the pass on 1^st & 10. It’s a short line to the end of the possession from there, as they win 2^nd & long against the run and pass at nearly identical rates, and 3^rd & long at a top-notch 73.9%.

The defensive numbers break down, in the alternative. First of all, opposing offenses threw the ball so infrequently in 2^nd & short / medium that I don’t have enough of a sample for quality statistical evaluation (as I said, it’s a copycat league). Those offenses strongly preferred rushing on 2^nd down with 6 or fewer to go, and JMU’s defensive effectiveness fell dramatically to only a 40% success rate, or 30% if it’s within 3 yards. I can say from watching tape that I don’t believe this is because JMU backed out to play the pass in these situations – they maintained their aggressive run defense posture, it’s just that they tend to always allow about 2-4 yards, and in shorter yardage situations this is enough for an offensive win.

On 3^rd downs they switch to their racecar package (in most cases regardless of the distance, though there are a few exceptions like the goalline), flexing one of their bigger DEs inside to play 3-tech and putting the two fastest DEs on the edges. This has worked as intended on 3^rd & long as mentioned above, but the effects are strange and unexpected otherwise – the other thing the racecar package is good at is stopping 3^rd & short rushing conversions, at 53.8% (much better than their 2^nd & short rate, when they have heavier personnel in), meanwhile their pass defense in shorter yardage collapses, to 49% on 3^rd & medium and 37.5% on 3^rd & short.

The four primary rotational defensive tackles are #0 DT I. Bush, #90 DT Lockett, #93 DT McMullin, and #33 DT Taddeo. All are redshirt seniors from the 2021 cycle except McMullin who’s a sophomore from 2023. Lockett, McMullin, and Taddeo are built pretty similarly, around 6’1” and 290 lbs, but Bush is instantly recognizable at the same height but close to 350 lbs and impressively wide.

During the times he’s been out, the response has mostly been to increase the rotations of the other three, and give some snaps to #88 DT K. Bush (I don’t believe there’s any relation), who’s a bit bigger at 6’3” and 310 lbs but just doesn’t fit the same human mountain dimensions and it clearly affects the front’s run fits. Surprisingly, Coach Chesney gave an interview on the same day we recorded the podcast in which he said that the starter Bush will suit up and play as a full-go – usually the staff doesn’t give out any reports like that, which actually made me a little suspicious that it might be disinformation. At any rate, his size alone is enough to affect the play so even if he’s not at 100% his mere presence is worth considering.

The two speed rushers are #9 DE X. Holmes and #8 DE A. Thomas, both about 6’2” and 240 lbs. Both have pretty good havoc rates on my tally sheet but due to their size and length they really need to get around the edge, or convert speed to power and catch an OT or TE before he’s anchored, because they tend to get stoned in a straight fight with a blocker who’s set. The end who plays with the most leverage is the redshirt freshman #15 DE West; he has by far the highest havoc rate and can really give tackles trouble with his length at 6’4”.

The most interesting is #97 DE Gobaira, a former bluechip transfer from Notre Dame – he’s the one who flexes inside on racecar packages. Gobaira is listed at 6’5” and 255 lbs, but I have a hard time believing that from watching tape, he doesn’t look that long and I would say he’s about 15 lbs heavier, so it doesn’t surprise me he’s playing the heavy end or tweener end/tackle spot.

At linebacker, the senior who’s played nearly every meaningful snap is #5 LB Hendrick – this is his second full year as starter, after playing his freshman and sophomore years as a frequently used backup behind Aiden Fisher and Jailin Walker before they transferred to Indiana. Hendrick’s grades are solid if not spectacular on my tally sheet.

The other spot(s) have been in more flux. Up through week 12, I would describe the other starter as junior #3 LB Weathersby, and #56 LB Spinogatti, who came over last year from Holy Cross, as the SAM who only played in the 4-3 configuration, with the backup if someone needed a break (though this was rare) being freshman #46 LB Ezeogu. Weathersby is perhaps the most hot-and-cold player on my tally sheet – very aggressive, great burst, hits with more pop than just about anybody I watched in this entire project, but also off his assignment quite a bit with more of a tendency to misread the play.

Weathersby followed up a great game in week 11 with an almost no-show game in week 12, and after that I think the staff flipped the jobs to have Spinogatti start and Weathersby play as the SAM, and Ezeogu rotated in a little more often. Spinogatti is less of a wild man than Weathersby is and fewer plays get by him for being out of position, but the athletic ceiling is obvious and he can be a liability in coverage.

Here’s a representative sample of successful rush defenses:

1. :00 – Look at Weathersby creeping up before the ball is snapped and immediately firing into the designed run lane, redirecting the back. Bush (the big one) cannot be controlled by the RT so the back goes even wider, where West is unblocked as the read man since this was an option play, and he wraps up without allowing anything more while the rest of the defense arrives.
2. :09 – This was just before Bush exited the game with a ding, and there’s a night and day different in how the front handles the interior lanes with him in. He’s clogged up everything between the right and left guards, so they just need Weathersby in the middle while Hendrick can run around to one side and the strong safety contains the other.
3. :19 – It’s 2^nd & 10, and note how most of the defense is focused on the run game. The nickel and safety perform an exchange while the other safety spins up a few yards to handle the motion man, and the corner over the X remains in press, but otherwise everyone’s locked into the backfield including the other corner who’s basically playing defensive end.
4. :28 – Here’s how the front looks with the other Bush in playing 1-tech now that big starter is out, and Spinogatti starting at linebacker. Still press man from the corners, single high safety locked into the backfield, and the safety and nickel are within 4 yards of the line of scrimmage. But there’s not an immediate attack of the line, it’s a little more conservative, and the back gets three yards or so while the defense rallies to the stop.

And unsuccessful rush defenses:

1. :00 – Very aggressive goalline attack out of the 4-3, with the backers and strong safety pushing in from the backside and Spinogatti shooting the gap playside. But he whiffs and now there’s no one behind him for 10 yards.
2. :13 – Here’s the same offense vs the 4-2-5 as in the second clip above but with Bush out of the lineup, and the response is to use multiple linebackers to plug the interior gaps and have the strong safety fire down immediately. They’re shoved aside and the opening at the third level creates a bigger run opportunity.
3. :26 – The offense is in 12-pers, note Spinogatti as the MIKE, Weathersby as the SAM, and the strong safety up on the line of scrimmage over outside TEs. It’s a midline read with counter action by the left side (like the very first clip in this article … copycat league) with the safety and SAM getting proper leverage on the TE blocks, and Hendrick triggers the other way on the QB keep, but Spinogatti is far too late – he needs to be flowing to the play sooner and faster.
4. :46 – Unlike the Sun Belt offenses, Wazzu’s zone blocking scheme wasn’t boom-or-bust but rather produced constant methodical gains against JMU. Not much to diagram here to a West coast football fan, the Cougs are simply getting a hat on a hat across the board, controlling with proper leverage, with the LT and RG up to the 2^nd level to turn the backers away from the run.

It is a challenge to evaluate the defensive backfield in a vacuum, entirely on their own merits, because there is a very clear strategic choice by the coaching staff to aggressively leverage attacks at the line of scrimmage to stifle the run and generate pass pressure, constantly leaving the coverage on islands. This not only includes pulling linebackers out of throwing lanes and playing the corners and nickel in man coverage, but eliminating most forms of safety help by frequently having the strong safety in the box or even on the line of scrimmage, while the free safety does not backpedal deep off the snap but rather keeps his eyes in the backfield in case it’s a draw or scramble.

This strategy is essentially a bet that opposing offenses don’t have the wherewithal — protection, receiver skill, QB arm talent — to take advantage of the minimal defensive resources JMU puts in the backfield and so it’s better to force the issue up front.

On the podcast we discussed for quite a while the defensive strategy, the essential gamble that it’s making and the context for why it’s paid off in the Sun Belt, and the prospects for seeing something different like a switch to zone (which they did in the second half against Texas State, the most prolific passing team they faced). The conclusion that we reached is that JMU is likely going to go with their main approach during the year to begin the game and see where it gets them, and perhaps reassess later in the game.

We also agreed that the defensive backs are doing their jobs pretty well all things considered, and that it’s the schematic choice that puts them in tough spots where they’re going to inevitably get burned sometimes, or have to play physically and get some flags on a certain number of reps – in other words, the DBs are the ones who look bad if the pass rush doesn’t get home, but it’s not fair to conclude that it’s because the DBs are in and of themselves bad, because they’re really not, and actually grade out pretty well in the context of the scheme.

So while coverage has been pretty good and they’ve got a decent interception rate (3.046% of passes attempted, 45^th nationally), it’s very clear from the nearly 40-percentage point dropoff between 3^rd & long and 3^rd & short pass defense success rate that the secondary’s role is hanging on and if the rush doesn’t get home quickly they get cooked. The one defect in the Dukes’ fundamental metrics from charting is explosive pass defense, and situationally the biggest damage comes on 3^rd & short when the pass rush is most neutered. Some examples:

The same five defensive backs play almost every meaningful snap, with only light rotation at free safety and a handful of backup snaps at corner (and of course, the nickel is off the field in the 4-3 configuration). The two starting corners are #30 CB Eaglin, a Cignetti recruit from 2022, and #2 CB Culp who’s on his fourth school since 2020 going from the FCS to Troy to Charlotte and finally JMU this year. I’ve seen only a little bit of #1 CB Rodgers, who came over from Div-II this year, though he has been trusted with some critical situations like the final possession in a one-score game in the championship.

At nickel I’ve only ever seen #32 DB Barksdale, a junior from the 2023 cycle. The roster lists #14 DB Flowe as the only other nickel who’s seen the field, though I don’t have him on my tally sheet during meaningful play.

At strong safety is senior #7 DB J. Thomas, and I’ve never seen him leave the field, which is a bit surprising because I think he has the most physically demanding job of the DBs as he’s constantly charging the offensive backfield and has quite a few snaps where he lines up across from a TE and takes on blocks. The free safety is sophomore #13 DB T. Brown, though as mentioned earlier he does not play as deep as many free safeties do and is usually assigned to the QB. The secondary backup I’ve seen most often is at this spot, senior #21 DB Harris-Lopez who transferred in from Holy Cross.

Here’s a representative sample of successfully defended passing plays:

1. :00 – Against Texas State’s prolific passing offense, JMU played the most zone coverage that they had all year. It tended to be more effective for them overall, if unfamiliar. The QB can’t find anything against the layered coverage to the field, and the E-T stunt is threatening him from the right so he dumps it down to the back. Culp does his assignment properly by hanging back in the lane then triggering for the tackle.
2. :07 – Here’s Weathersby on the blitz, the offense has no answer for him and the man coverage is holding up against all the WRs and TE. The RB is wide open on the wheel for a big gain (Hendrick is late recovering from the bluff) but the QB doesn’t have time to set his feet and Weathersby has him panicking.
3. :20 – Pretty obvious creeper from the strong safety but Thomas still blows past the LT. Note the end bailing out into coverage on the opposite side and the confusion between Barksdale and Culp on the switch.
4. :34 – Here’s the racecar package with Gobaira playing fist down, West and Holmes on either side of him, and this late in the season with the Bush injury and the LB switch, Spinogatti and Ezeogu on-ball. The coverage backs into zone, a better throw would be to the No.2 with a big gap between Spinogatti and Barksdale since the nickel has to be underneath the No.1’s sideline route, but the stunt pressures the QB into an early throw and Brown gets an easy PBU.

And unsuccessfully defended passing plays:

1. :00 – This blitz brings Thomas from seven yards deep and it’s way too long developing, the QB has time to do a complete field read and reset his feet. The offense is unbalanced to the field so bailing the coverage to the boundary isn’t helpful, the TE has no one assigned to him in the flat. Eaglin overruns the in-out move and gets flipped around. Two wide open options to choose from and time to make the choice, 30 yard gain on 1^st & 15.
2. :16 – Texas State had come off of several plays hammering JMU playing man in the 1^st half, and the Dukes switched to disguised zone to start the 2^nd. This didn’t go much better, Thomas and Weathersby are out of their lanes and Eaglin is late over the top.
3. :39 – JMU’s secondary was flagged for DPI or holding on 5.41% of meaningful downfield pass attempts (the model indicates 4.72% for a team playing this much man is expected, so a bit higher but not wildly so). Here Barksdale doesn’t want the inside receiver to park at the line to gain and doesn’t start backpedaling until they make contact, and then rides him with his upfield and downfield arm restricting.
4. :57 – This is what the split safety man coverage looks like, on 1^st down and very long yardage after an offensive foul. Note that all the help is still to the inside — for the nickel on the TE who does release, maybe the RB or other TE if they leak out after blocking, or the QB on a scramble — while the corners in man are still completely on their own. The pass rush is picked up, the WR gets a step on Culp, and there’s no answer for the well placed back shoulder fade.