DVOA is a situation-adjusted efficiency stat. It measures how much better or worse a team or unit plays than a league-average team would in the same spots. In plain English: did the yards and scores come in easy situations or hard ones? Then grade the performance on that basis.
What is DVOA in football?
DVOA stands for Defense-adjusted Value Over Average. It is the Football Outsiders / FTN Football system that grades every play against a baseline for down, distance, field position, time, and opponent strength. A 7-yard gain on 3rd-and-2 is not the same as a 7-yard gain on 3rd-and-12. DVOA scores success relative to what an average offense would do in that exact spot, then adjusts for the defense faced.
Raw box-score stats love volume and ignore context. A team can stack yards against soft prevent looks late, or pad rushing totals when the game is already decided. DVOA tries to strip that noise out. Success on early downs in tight games against good defenses grades higher than empty yards in garbage time. That is why people who care about success rate and true drive value often land next to DVOA when they want a cleaner team read.
You will usually see DVOA as a percentage. Positive means better than average. Negative means worse. A +15% offensive DVOA roughly means the unit was about 15% more valuable per play than a league-average offense in comparable situations. League average sits near 0%. The scale is relative. It is not a counting stat like yards or points.
| Lens | What it counts | What it misses | Prediction use |
|---|---|---|---|
| Total yards | Volume of production | Down, distance, score, opponent | Weak for next-week strength |
| Points per game | Final scoreboard output | Turnovers, field position luck, schedule | Okay ceiling, noisy week to week |
| Success rate | Down-to-down efficiency | Explosive value and full drive worth | Strong for "are they in rhythm?" |
| DVOA | Situation- and defense-adjusted value | Tiny samples early; special teams splits matter | Best single team-efficiency snapshot |
Takeaway: yards tell you what the tape counter saw. DVOA tells you whether those plays were hard wins or easy ones.
What does "situation-adjusted" really mean?
Situation-adjusted means each play gets judged against the expected value of that specific down, distance, field position, and game state, not against a flat yards-per-play average. A 4-yard run on 1st-and-10 is fine. A 4-yard run on 3rd-and-8 is a failure. Same gain, opposite grade.
Picture a drive that starts at your own 5 versus one that starts at the opponent 40 after a short punt. The second drive needs fewer good plays to score. If you only look at points or total yards, both offenses can look similar. Situation adjustment asks whether the offense created value given where it started and what the down-and-distance ladder looked like.
Time and score matter too. Kneel-downs, end-of-half heaves, and prevent-defense checkdowns warp raw averages. DVOA down-weights empty late-game junk and up-weights plays that change win probability in competitive spots. That is the core of "value over average." Average is not a random team on a neutral field forever. Average is what the league typically does in that same situation.
For prediction, this is the part casual box-score reads skip. A backup running back can look great on 25 late carries when the defense is soft. A tight end can post a quiet yardage line while converting two critical 3rd downs. Situation-adjusted thinking forces you to ask when the production happened, not only how much showed up in the final line.
How does the defense adjustment work?
The defense adjustment credits or debits performance based on the quality of the opponent unit faced. Torching a weak defense grades lower than the same output against a strong one. That keeps schedule strength from lying to you for a full season.
If Team A puts up the same raw efficiency against the league's worst pass defense that Team B puts up against a top unit, DVOA will like Team B more. Over 17 games that compounds. A soft early schedule can inflate yards and points for a month. Defense-adjusted numbers usually cool that story sooner.
This is also why DVOA is more useful midseason than after Week 1. Opponent adjustments need enough data to mean something. Early on, the system still grades situations, but the "who you played" piece is noisier. By the time teams have faced a real mix of fronts and coverages, the ranking stabilizes and becomes a better prior for "are they actually good?"
When you project player props, remember DVOA is primarily a team and unit framework. It still helps your player reads indirectly. A receiver's target tree looks different against a bottom-five pass defense than against a top pass rush that collapses clean pockets. Pairing unit DVOA with usage and pressure context beats trusting last week's box score alone. For the pressure side of that chain, pressure rate is the natural companion number.
Why do yards and points mislead more than DVOA?
Yards and points mislead because they ignore difficulty, opponent quality, and empty volume. DVOA grades efficiency in context. A 400-yard passing day can be mostly dump-offs on 2nd-and-long against a defense that sold out to stop the run and protected a lead.
Classic traps show up every season. A team ranks top-five in total offense after feasting on poor defenses and short fields from turnovers. Another ranks mid-pack in yards but keeps converting in the red zone and winning early-down leverage. The second team is often the one DVOA prefers, and the one that holds up better when the schedule hardens.
Special teams and turnovers also scramble plain scoring stats. A defense can look elite on points allowed because the offense never puts it in long fields, or look bad because every opponent drive starts on a short field. DVOA splits offense, defense, and special teams so you can see which unit is actually carrying the rating. That split matters when you predict next week's side of the ball, not just "good team vs bad team."
Historical leaderboards on sites that publish the metric, including long-running writeups tied to the Football Outsiders tradition now carried forward in FTN's DVOA work at FTN Fantasy's DVOA hub and reference archives discussed across Pro-Football-Reference era pages, keep showing the same pattern: high DVOA teams sustain wins better than high-yardage teams with weak situational profiles. You do not need the proprietary play-by-play file to use the idea. You need the habit of asking whether the production was expensive or cheap to earn.
How should you use DVOA when predicting next week?
Use DVOA as a strength-of-performance prior for team units, then layer matchup, injuries, and role changes before you lock a player or team expectation. It is a base rate, not a final answer.
Before you lean on the number:
- Check sample size and schedule: is this an early-season mirage or a 8-plus game profile against mixed competition?
- Split the unit: confirm whether offense, defense, or special teams is driving the headline DVOA you are citing.
- Translate to the prediction: name the next-game implication (pass-game efficiency, red-zone drives, early-down stickiness) instead of stopping at "they're ranked 6th."
Practical examples help. If a team ranks well in offensive DVOA but lives on explosives and sits average on early downs, week-to-week passing yards can swing hard. If defensive DVOA is strong against the pass but opponent quarterbacks still get clean launches, look at pass rush and secondary injuries before fading a receiver. If special teams DVOA is carrying a mediocre offense, do not expect the same scoring environment when the return game cools.
Player-level reads still need usage. DVOA will not tell you a back is about to jump from 12 to 22 touches after an injury. It will tell you whether the offense inventing those touches is actually efficient in the situations that create lasting fantasy and prediction value. Combine unit quality with snap share, route participation, and goal-line roles. On scoring chances specifically, red zone efficiency often explains why two similar yardage offenses produce different touchdown totals.
Weighted DVOA and late-season variants matter as the year goes on because recent games get more voice. If you are predicting December football off September dominance, check whether the weighted profile still agrees with the full-season number. Big gaps mean the team has changed, the schedule flipped, or both.
Where does DVOA still fall short?
DVOA still falls short in tiny samples, opaque single-play luck spikes, and anything that needs pure player isolation rather than unit context. It is a team efficiency lens first.
Week 1 and Week 2 rankings bounce. One exploded return, two fluky turnovers, or a monsoon game can shove a unit around the table. Treat early DVOA as a clue, not a verdict. By midseason the signal usually outweighs the noise for team strength.
It also will not replace quarterback charting traits like accuracy over expected on a throw-by-throw basis. Companion metrics such as CPOE answer a different question: did this passer hit the throws the design and separation called for? DVOA answers whether the offense as a whole created value after situation and opponent adjustments. You want both habits, not a holy war between them.
Finally, keep some predictive humility. A great DVOA team can still drop a game on the road in the rain with three starters out. The stat improves your prior. It does not eliminate variance. Your job as a fan who predicts is to start from a better baseline than "they had 380 yards last week," then adjust for the real world of injuries, scripting, and game flow.
How does DVOA connect to reading games better?
DVOA connects to better game reading by training you to value hard yards in hard spots over pretty box scores. That is the same skill you use when you forecast what happens next. Once you internalize situation value, you watch drives differently.
You start noticing whether an offense is winning early downs or surviving on 3rd-and-long heroes. You notice whether a defense is actually stiff or just eating soft schedules and short fields. You stop overrating garbage-time pad stats when you build expectations for the next slate.
That mindset is the whole point of stats literacy on GAGE: learn the number, then use it to call the next performance with clearer eyes. If you want a place to practice those reads against real slates and score how sharp your takes are, Download GAGE and put the framework to work on live games.
Is DVOA the same as EPA?
No. EPA (Expected Points Added) estimates how much a play changes expected points. DVOA builds a defense-adjusted value-over-average percentage from situation-based success and opponent adjustments. They often point the same direction, but they are not identical systems or identical scales.
What is a good DVOA number?
Near 0% is average. Roughly +10% to +15% is strong. Sustained figures around +20% or better are elite territory for a unit over a meaningful sample. Always read the number next to games played and schedule, not as a lone badge.
Can you use DVOA for player props?
Yes, as a unit-quality prior. It helps you judge whether the offense or defense creating the opportunity stream is truly efficient. Then you still need snaps, targets, red-zone role, and matchup detail to land on a player line. DVOA sets the weather. Usage sets the forecast for that skill player.