In a recent article here on Blogging The Boys, we looked at Wins Over Average (WOA) and which NFL quarterbacks were winning games for their teams. In the course of the ensuing discussion, some readers wondered whether WOA was something that could be applied to head coaches as well.
Traditionally, coaches are evaluated by their winning percentage, and an easy way to calculate a head coach’s Wins Over Average (WOA) would be to subtract the league’s average winning percentage (.500) from the coach’s career
winning percentage, then multiply that difference by their total games coached.
This baseline formula would isolate how many actual victories the coach secured compared to what a perfectly average (.500) coach would have achieved over the exact same number of games.
But that just gives you a souped up version of the winning percentage. What if there was a way to isolate the head coach’s impact more effectively?
Pythagorean Expectation
So instead of using 0.500 as the baseline, we‘ll use the Pythagorean Expectation to create a custom baseline derived entirely from the team’s points scored and points allowed. This isolates the head coach’s impact by comparing a team’s actual wins against how many wins the team’s point differential says they should have had. So instead of Wins Over Average (WOA) we get Wins Above Expectation (WAE), which we’ll use for the rest of this post.
If you’ve been on this blog for a while, you’re familiar with the Pythagorean Formula. If not, here is what it is: The formula was originally developed by the godfather of baseball stats, Bill James, who surmised that a team’s true strength could be measured more accurately by looking at points scored and points allowed, rather than by looking at wins and losses. And here’s why we’re using it:
1. Winning percentage ignores margin of victory
Winning percentage treats every victory and defeat identically, regardless of how the game was played. A 1-point win secured on a lucky, deflected pass at the buzzer looks exactly the same on a record sheet as a 35-point blowout where a team dominated from the opening kickoff. Winning percentage tells you what happened, but it tells you nothing about how it happened.
2. The Pythagorean Formula accounts for roster strength
The Pythagorean expectation shifts the focus from volatile win-loss records to a team’s underlying point differential. In the NFL, points scored and points allowed are far more stable and mathematically predictive of future success than a raw win-loss record. By calculating how many games a team should have won based on their total points, the Pythagorean formula establishes a highly accurate baseline for the true talent level of the roster.
3. It corrects for one-score game randomness
NFL games decided by eight points or fewer are historically highly volatile and prone to random variance (e.g., a bad referee call, an unusual bounce of a fumble, or extreme weather).
- Overachieving: A team like the 2025 Broncos (11-2 in one-score games, 14-3 overall) usually isn’t as elite as their record suggests. Winning percentage crowns them a powerhouse, but the Pythagorean baseline reveals they had the point profile of a 11-win team, showing the coach either squeezed out some extra efficiency, or simply got lucky.
- Underachieving: Conversely, a team like the 2025 Chiefs that finished 6-11 but was 1-9 in one-score games is often much better than their record. The Pythagorean baseline has them as a 9.5-win team, suggesting the coach might have left wins on the table.
4. It gives us a clear “control group” baseline
To truly measure a head coach’s impact, you need a statistical control group—an “average coach” baseline. Simply subtracting .500 from a coach’s winning percentage assumes every coach inherits an exactly average roster. This punishes good coaches on bad teams and rewards bad coaches who inherit good rosters. The Pythagorean expectation creates a customized baseline for that specific roster. It acts as a simulation, asking: “How many games would a perfectly average, replacement-level coach win if you handed them this exact point-differential profile?”
When you subtract Pythagorean expected wins from actual wins to find Wins Above Expectations (WAE), you isolate the exact margins where a coach makes his money. If a coach consistently beats their Pythagorean expectation, it is concrete statistical evidence of high-value coaching impact: elite late-game clock management, superior high-leverage play-calling, and structural discipline in one-score situations.
NFL head coach tiers
With all that out of the way, here’s a look at 16 currently active NFL head coaches (marked in bold) with at least three years of experience with their current team, along with a selection of 23 former head coaches. We’ll look at the 10 active head coaches with one or two years of experience a little further down this post.
The table is color coded into tiers to improve legibility, and I’ll expand on those tiers after the table.
body .sbnu-legacy-content-table td, body .sbnu-legacy-content-table th, body .sbnu-legacy-content-table { border: 1px solid #000 !important; border-collapse: collapse !important; } body .sbnu-legacy-content-table td, body .sbnu-legacy-content-table th { padding: 0px 6px !important; }| NFL head coaches and WAE | ||||
| Tier | Coach | 2026 team | Seasons | WAE per season |
| Elite Optimizers | Kevin O’Connell | MIN | 4 | +1.67 |
| Nick Sirianni | PHI | 5 | +1.27 | |
| Jon Gruden (LV) | 2018-21 | 3.3 | +1.01 | |
| Matt LaFleur | GB | 7 | +0.97 | |
| DeMeco Ryans | HOU | 3 | +0.94 | |
| Andy Reid | KC | 13 | +0.87 | |
| Sean Payton | DEN | 3 | +0.82 | |
| Mike Tomlin (PIT) | 2007-25 | 19 | +0.67 | |
| Sean McDermot (BUF) | 2017-25 | 9 | +0.66 | |
| Mike Vrabel (TEN) | 2018-23 | 6 | +0.57 | |
| Sean McVay | LAR | 9 | +0.51 | |
| Floor Raisers | Kevin Stefanski (CLE) | ATL | 6 | +0.44 |
| Bill Belichick (NE) | 2000-23 | 24 | +0.34 | |
| Mike McCarthy (GB) | 2006-18 | 12.8 | +0.32 | |
| Mike McDaniel (MIA) | 2022-25 | 4 | +0.31 | |
| Dan Campbell | DET | 5 | +0.25 | |
| Roster Stabilizers | Pete Carroll (SEA) | 2010-23 | 14 | +0.20 |
| Jason Garrett (DAL) | 2010-19 | 9.5 | +0.19 | |
| Tom Coughlin (NYG) | 2004-15 | 12 | +0.14 | |
| Mike McCarthy (DAL) | PIT | 5 | +0.09 | |
| Jeff Fisher (LAR) | 2012-16 | 5 | +0.08 | |
| Ron Rivera (WAS) | 2020-23 | 4 | +0.05 | |
| Mike Zimmer (MIN) | 2014-21 | 8 | +0.03 | |
| Shane Steichen | IND | 3 | 0.00 | |
| Todd Bowles | TB | 4 | -0.04 | |
| Bruce Arians (TB) | 2019-21 | 3 | -0.05 | |
| Brian Daboll (NYG) | 2021-2025 | 3.6 | -0.17 | |
| Andy Reid (PHI) | 1999-2012 | 14 | -0.17 | |
| Efficiency Bleeders | Doug Pederson (PHI) | 2016-20 | 5 | -0.30 |
| Zac Taylor | CIN | 7 | -0.30 | |
| John Harbaugh (BAL) | NYG | 18 | -0.31 | |
| Robert Saleh (NYJ) | TEN | 4 | -0.36 | |
| Kyle Shanahan | SF | 9 | -0.37 | |
| High-Risk Underperformers | Dan Quinn (ATL) | 2015-20 | 6 | -0.58 |
| Jon Gruden (TB) | 2002-08 | 7 | -0.62 | |
| Doug Pederson (JAX) | 2022-24 | 3 | -0.71 | |
| Matt Eberflus (CHI) | 2022-24 | 3.7 | -1.39 | |
| Dennis Allen (NO) | 2022-24 | 3 | -1.49 | |
| Jonathan Gannon (ARI) | 2023-25 | 3 | -1.63 | |
You may have immediately noticed one glaring oddity here, and that oddity also shows the limitation of this approach: Kyle Shanahan is ranked as an efficiency bleeder, just narrowly missing the red tier.
The Shanahan Paradox
Much like Andy Reid’s early era in Philadelphia, Shanahan’s negative career WAE is actually a testament to his elite offensive design, as crazy as that might sound. When the 49ers are fully healthy, they build historically massive point differentials by blowing out opponents (fourth overall in points differential since 2019). Because they dominate so thoroughly in (some of) their victories, the Pythagorean formula locks in extreme, near-perfect win expectations that are nearly impossible to maintain in reality, which suppresses Shanahan’s close-game WAE score.
But a high point differential does not automatically sentence a head coach to a negative or flat WAE score. The 2007 Patriots are the ultimate case study in proving that a team can possess both an historically massive point differential and generate an elite, positive coaching value.
The 2007 Patriots scored 589 points and allowed 274 points, giving them a Pythagorean expectation of 13.76 wins and a massive +2.24 WAE premium. But the 2007 Patriots didn’t just have one or two massive blowouts that inflated their season stats. They outscored their opponents by a staggering 315 points because they dominated almost every single week. In contrast, teams caught in the Shanahan Paradox often compile a massive point differential by winning a handful of games in historic fashion (e.g., winning 42-10 or 45-7), but then dropping three or four highly volatile, tight games by a field goal.
Tier 1: Elite Optimizers (>0.5 WAE)
This tier features some of the best head coaches in the game today, head coaches who successfully manage thin margins and consistently squeeze extra wins out of tight games.
Eagles fans should disregard this entire post. Nick Sirianni, well-known for his grumpiness, persistent offensive stagnation, poor late-game management, and lack of locker room discipline should be fired immediately. Who needs a guy whose teams consistently find ways to win close games?
Not even Bill Belichick cracks this tier over a 24-year timeline because his rosters were not able to carry an unrepeatable situational curve over decades.
Breaking down Belichick’s 24-year tenure in New England into three buckets isolates exactly how much of his coaching efficiency was tethered to stable quarterback play versus schematic point dominance.
- -0.80 WAE : The Non-Brady Years (2000, 2008 due to injury, and 2020–2023). While Belichick field-marshaled a legendary 11–5 campaign out of Matt Cassel in 2008 and dragged Mac Jones to the playoffs in 2021, his non-Brady teams routinely left minor fractions of wins on the table due to late-game offensive stagnation. The math suggests that his base defensive schemes kept the point margins close, but without elite quarterback execution, he lost more close, one-score variance games than the formula anticipated.
- +0.88 WAE : The Brady “Growth” Era (2001–2006). This is the absolute apex of Bill Belichick’s structural coaching value, comfortably tracking at a Tier 1 pace. During New England’s initial three-championship run, the Patriots were not yet statistical juggernauts; they did not blow opponents out regularly, resulting in modest point differentials that only expected an average of 11 wins per season. Because Belichick’s situational defense, historic special teams execution, and error-free game management consistently maximized thin margins, his teams beat their baselines year after year.
- +0.64 WAE : The Brady “Blowout” Dynasty (2007–2019). Once the offense added weapons like Randy Moss and shifted into a pass-heavy spread, the Patriots began registering massive multi-score blowouts. This introduces the “Shanahan Paradox” we observed earlier. Because the 2007–2019 Patriots regularly compiled massive positive point margins, the Pythagorean formula assigned them near-perfect win expectations (often projecting 13 or 14 wins). And even though Belichick’s WAE average flattened slightly, he remained a Tier 1 head coach.
Tier 2: The Floor Raisers (0.2-0.5 WAE)
These are high-level tacticians who regularly tilt close-game variance in their team’s favor. We’ll go into the Green Bay McCarthy versus Dallas McCarthy debate a bit later in this post.
Tier 3: Roster Stabilizers (-0.2 – +0.2 WAE)
The vast majority of competent, long-term NFL coaches land here. They operate as reliable stewards who win the exact amount of games their rosters deserve. This tier may feel a bit boring, because there doesn’t appear to be much happening in either a positive or negative direction, but the total WAE numbers here often hide some serious volatility. Take Jason Garret, he of the 8-8 finishes.
Garrett spent nine and a half seasons in charge of the Cowboys (2010–2019), first as the interim guy, then as the full-time head coach, and compiled an 85–67 regular-season record. Garrett finished his Dallas career with a seemingly steady +0.19 WAE per year, and his career a defined by two franchise quarterbacks and the ups and downs that came with that.
1. The Tony Romo Era (2011–2015)
Garrett’s initial years were spent maximizing the arm talent of Tony Romo, a partnership characterized by high-volume passing offenses and extreme point-differential volatility.
- The 8–8 Stagnation (2011–2013, -0.05 WAE): Romo’s offense scored at a top-tier clip, but historically porous defenses regularly surrendered equal chunks of points.
- The 2014 Peak (+1.42 WAE): Backed by Romo’s league-leading 113.2 passer rating and a dominant offensive line, Dallas went 12–4. Because they won tight, high-leverage games instead of logging useless blowout points, the formula projected 10.58 wins, netting Garrett a massive +1.42 WAE premium.
- The 2015 Disaster (-1.21 WAE): When Romo broke his collarbone, Garrett’s inability to adapt his scheme to backup quarterbacks Brandon Weeden and Matt Cassel was exposed. The roster’s scoring floor caved in, and the team left over a full win on the table relative to their baseline, finishing 4–12.
2. The Dak Prescott Hand-Off (2016–2019)
The arrival of rookie Dak Prescott in 2016 completely transformed Garrett’s coaching profile, turning the Cowboys into a run-heavy, ball-control team built around Ezekiel Elliott.
- The 2016 rookie surge (+2.11 WAE): Forced to pivot after another preseason Romo injury, Garrett customized the offense for Prescott. The rookie responded with a 23-to-4 touchdown-to-interception ratio. By protecting the football and mastering late-game clock sequencing, Dallas cruised to 13–3 against a 10.89 expected win baseline, securing Garrett’s career-high +2.11 single-season WAE.
- The roster stabilizer (2017–2018, +0.98 WAE): As Prescott matured, Garrett’s teams continued to overperform thin scoring profiles to steal wins, logging a +0.39 WAE in 2017 and a highly efficient +1.57 WAE in 2018 by routinely dominating close, one-score finishes.
- The 2019 Final Collapse (-2.74 WAE): Garrett’s tenure ended on a historically severe instance of the “Shanahan Paradox.” Prescott threw for a career-high 4,902 yards, and the team piled up a massive +113 point differential by blowing out bad opponents. However, because they completely stalled out in close games against good teams, they finished a mediocre 8–8 and a massive -2.74 WAE underperformance ultimately cost Garrett his job.
Tier 4: Efficiency Bleeders (-0.5 – -0.2 WAE)
This tier features coaches that build massive point margins in blowouts but struggle to close tight, high-leverage games.
Tier 5: High-risk underperformers (< -0.5 WAE)
Coaches who consistently leave wins on the table, either by dropping a high volume of winnable, one-score games or by absorbing severe blowout losses that tank their point efficiency.
- Matt Eberflus and Jonathan Gannon were frequently plagued by historic late-game defensive collapses, blown multi-score leads in the fourth quarter, and highly conservative situational execution that actively converted winnable point profiles into real losses.
- Dennis Allen and Doug Pederson were erratic game managers with questionable fourth-down decision-making, and recurring red-zone inefficiencies that stripped points off the board and stalled out close games.
- Dan Quinn and Jon Gruden are historical examples of this tier during their specific stints. Despite possessing highly explosive roster talent that built massive point advantages, their teams notoriously suffered from extreme second-half stagnation and legendary game-management failures .
Rookie and sophomore NFL head coaches
This entire analysis is operating on relatively small sample sizes, and we see that this can lead to some extreme outliers for head coaches with one or two years of experience. Which is why they’re grouped here separately.
body .sbnu-legacy-content-table td, body .sbnu-legacy-content-table th, body .sbnu-legacy-content-table { border: 1px solid #000 !important; border-collapse: collapse !important; } body .sbnu-legacy-content-table td, body .sbnu-legacy-content-table th { padding: 0px 6px !important; }| Active head coaches with 1 or 2 years in current role | |||||
| Coach | Team | Seasons | WAE per season | ||
| Ben Johnson | CHI | 1 | 1.89 | ||
| Mike Vrabel | NE | 1 | 1.54 | ||
| Liam Coen | JAX | 1 | 1.21 | ||
| Mike Macdonald | SEA | 2 | 1.13 | ||
| Dave Canales | CAR | 2 | 1.06 | ||
| Jim Harbaugh | LAC | 2 | 0.70 | ||
| Dan Quinn | WAS | 2 | 0.10 | ||
| Brian Schottenheimer | DAL | 1 | -0.18 | ||
| Kellen Moore | NO | 1 | -0.29 | ||
| Aaron Glenn | NYJ | 1 | -0.86 | ||
Nothing much to say here except for two statistical points: History suggests the kind of WAE overperformance we are seeing with the “blue” guys here is not sustainable. And in the unlikely case that it is, then we’ll have to redefine the tiers; can’t have 13 out of 32 active head coaches in an “Elite” tier.
Dallas Cowboys head coaches
Before looking at the rest of the Cowboys coaches, we’ll first look at Mike McCarthy’s uncomfortably long tenure in Dallas to explain the difference between what the Cowboys thought they were buying and what they ended up getting.
body .sbnu-legacy-content-table td, body .sbnu-legacy-content-table th, body .sbnu-legacy-content-table { border: 1px solid #000 !important; border-collapse: collapse !important; } body .sbnu-legacy-content-table td, body .sbnu-legacy-content-table th { padding: 0px 6px !important; }| Mike McCarthy Coaching tenure in Dallas | |||||
| Year | W/L | W/L% | Points For | Points Against | Wins over Average |
| 2024 | 7-10 | 0.412 | 350 | 468 | 1.32 |
| 2023 | 12-5 | 0.706 | 509 | 315 | -0.87 |
| 2022 | 12-5 | 0.706 | 467 | 342 | 0.50 |
| 2021 | 12-5 | 0.706 | 530 | 358 | -0.19 |
| 2020 | 6-10 | 0.375 | 395 | 473 | -0.32 |
| Total | 49-35 | 0.583 | 2,251 | 1,956 | 0.43 |
| Source: ProFootballReference.com | |||||
Over five seasons with the Cowboys, Mike McCarthy finished with +0.43 Pythagorean Wins Above Expectations (WAE), or +0.09 per season. His numbers are closely aligned with the team’s talent profile and he largely performed near the expected win totals. While the 2021–2023 “12-5 blowout era” saw slightly lower performance than expected due to massive point differentials, the 2024 season marked his best analytical showing with a +1.32 overperformance, establishing him as a stabilizer who rarely over- or underachieved his roster’s capability. But if we exclude 2024, McCarthy is at a very pedestrian -0.22 WAE, so he was cruising along as an efficiency bleeder for four years, delivered a catastrophic playoff loss in 2023, and the Cowboys still didn’t see fit to fire him.
How does that compare to his tenure in Green Bay (2008-2018), you might wonder? In Green Bay, McCarthy proved to be a highly effective “floor-raiser” and high-leverage optimizer. Over 13 seasons, he steered the Packers to a solid +0.312 wins above expectation per season, and his teams outperformed their expected wins in nine out of 13 seasons. Because he routinely squeezed extra victories out of relatively lean point structures (such as turning a modest +45 point differential in 2015 into 10 actual wins with +0.77 WAE) he demonstrated strong situational execution, keeping the franchise consistently above its mathematical baseline. But like Belichick, McCarthy’s story in Green Bay cannot be told without understanding the quarterback situation.
- +1.64 WAE : The Favre Transition (2006-2007). Before Aaron Rodgers ever started a game, McCarthy operated at a Tier 1 pace. Taking over a fractured 4–12 roster, McCarthy maximized thin scoring margins with Brett Favre en route to 8-8 and 13-3 finishes. The suggests early McCarthy was an exceptional floor-raiser who actively manufactured wins before starting to lean on Aaron Rodgers.
- +0.21 WAE : The Rodgers Era (2008–2017). Once Rodgers assumed the starting role and reached his physical prime, McCarthy’s profile dropped into Tier 2. Unlike Belichick, who experienced an efficiency boost during his prime quarterback pairing, McCarthy regressed as a head coach, but still could be considered a “floor-raiser”. Notably, when Rodgers missed large chunks of the season in 2013 and 2017, McCarthy still managed a positive WAE with Matt Flynn (+0.75 WAE) and Brett Hundley (+0.70 WAE). While the raw win totals plummeted without Rodgers, McCarthy’s situational efficiency actually stayed positive. He squeezed extra wins out of highly compromised, low-scoring rosters, a pattern he would repeat in 2022 and 2024 with Dak Prescott out injured.
- -1.35 WAE : The Rodgers breakup. That was a bad breakup in many ways, and without that final season, McCarthy would have had a +0.45 WAE for his tenure in Green Bay, borderline Tier 1.
Ultimately, Green Bay McCarthy was a coach who actively manufactured extra wins through sharp high-leverage execution. Dallas McCarthy was a conservative manager who kept a steady hand on the wheel, delivering exactly what the roster was able to achieve – and nothing more.
Now on to the Cowboys head coaches, in chronological order.
body .sbnu-legacy-content-table td, body .sbnu-legacy-content-table th, body .sbnu-legacy-content-table { border: 1px solid #000 !important; border-collapse: collapse !important; } body .sbnu-legacy-content-table td, body .sbnu-legacy-content-table th { padding: 0px 6px !important; }| Dallas Cowboys head coaches and WAE | |||||
| Coach | Years | Seasons | WAE per season | W/L | W/L% |
| Schottenheimer | 2025 | 1 | -0.18 | 7-9-1 | 0.441 |
| McCarthy | 2020-2024 | 5 | +0.09 | 49-35 | 0.583 |
| Garrett | 2010-2019 | 9.5 | +0.19 | 85-67 | 0.559 |
| Phillips | 2007-2010 | 3.5 | +0.48 | 56-34 | 0.607 |
| Parcells | 2003-2006 | 4 | +0.40 | 34-30 | 0.531 |
| Campo | 2000-2002 | 3 | -0.19 | 15-33 | 0.313 |
| Gailey | 1998-1999 | 2 | -1.59 | 18-14 | 0.563 |
| Switzer | 1994-1997 | 4 | -0.21 | 40-24 | 0.625 |
| Johnson | 1989-1993 | 5 | +0.36 | 44-46 | 0.550 |
| Landry | 1986-1988 | 3 | -1.10 | 17-30-0 | 0.361 |
| Landry | 1966-1985 | 20 | +0.34 | 208-79-2 | 0.723 |
| Landry | 1960-1965 | 5 | -0.74 | 25-53-4 | 0.329 |
| Source: ProFootballReference.com | |||||
Too much yellow, not enough green, and no blue. That pretty much sums up the WAE for the Cowboys head coaches.
A few things stand out for me though.
- I was surprised to see Wade Phillips up so high. And if we were to discard that 2010 disaster, Phillips would be a Tier 1 baller with +0.75 WAE. It has become clear to me during this exercise that you can’t separate the coach from the QB, and Phillips was playing with prime Tony Romo, so his numbers shouldn’t really be that much of a surprise in the end.
- But that makes Parcells’ numbers even more impressive, considering he had to field the likes of Quincy Carter, Vinny Testaverde, or Drew Bledsoe.
- It also makes Garrett’s and McCarthy’s numbers look even weaker: They had Dak Prescott in his prime and couldn’t do much with it.
- I don’t know much about the Cowboys history under Tom Landry, but the split I used here is bracketed by his first playoff appearance in 1966 and his last in 1986.
- And where are we with Schottenheimer? Yes, he was hampered by an historically bad defense, but he finished almost exactly in line with his roster’s statistical baseline. No more, no less.
Which coaches stand out to you, both positively and negatively?
